2019-04-30 08:27:42 +08:00
|
|
|
# frozen_string_literal: true
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
require "html_to_markdown"
|
|
|
|
|
|
2022-07-28 10:27:38 +08:00
|
|
|
RSpec.describe HtmlToMarkdown do
|
2017-05-04 04:53:47 +08:00
|
|
|
def html_to_markdown(html, opts = {})
|
|
|
|
|
HtmlToMarkdown.new(html, opts).to_markdown
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
2017-05-04 00:04:31 +08:00
|
|
|
it "remove whitespaces" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
html = <<-HTML
|
2017-05-04 00:04:31 +08:00
|
|
|
<div dir="auto">Hello,
|
|
|
|
|
<div dir="auto"><br></div>
|
|
|
|
|
<div dir="auto"> This is the 1st paragraph. </div>
|
|
|
|
|
<div dir="auto"><br></div>
|
|
|
|
|
<div dir="auto">
|
|
|
|
|
This is another paragraph
|
|
|
|
|
</div>
|
|
|
|
|
</div>
|
|
|
|
|
HTML
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq(
|
|
|
|
|
"Hello,\n\nThis is the 1st paragraph.\n\nThis is another paragraph",
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
html = <<~HTML
|
|
|
|
|
<body text="#000000" bgcolor="#FFFFFF">
|
|
|
|
|
<p>Let me see if it happens by answering your message through
|
|
|
|
|
Thunderbird.</p>
|
|
|
|
|
<p>Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1
|
|
|
|
|
Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1
|
|
|
|
|
Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1
|
|
|
|
|
Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1
|
|
|
|
|
Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1
|
|
|
|
|
Long sentence 1
|
|
|
|
|
</p>
|
|
|
|
|
</body>
|
|
|
|
|
HTML
|
|
|
|
|
|
|
|
|
|
markdown = <<~MD
|
2021-09-17 16:41:34 +08:00
|
|
|
Let me see if it happens by answering your message through Thunderbird.
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
|
2021-09-17 16:41:34 +08:00
|
|
|
Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1 Long sentence 1
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
MD
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq(markdown.strip)
|
|
|
|
|
|
|
|
|
|
html = <<~HTML
|
|
|
|
|
<p> This post
|
|
|
|
|
has lots<br> of
|
|
|
|
|
space
|
|
|
|
|
</p>
|
|
|
|
|
<pre> This space was left untouched !</pre>
|
|
|
|
|
HTML
|
|
|
|
|
|
|
|
|
|
markdown = <<~MD
|
|
|
|
|
This post has lots
|
|
|
|
|
of space
|
|
|
|
|
|
|
|
|
|
```
|
|
|
|
|
This space was left untouched !
|
|
|
|
|
```
|
|
|
|
|
MD
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq(markdown.strip)
|
2017-05-04 00:04:31 +08:00
|
|
|
end
|
|
|
|
|
|
2024-06-09 18:51:23 +08:00
|
|
|
it "removes tags that aren't allowed" do
|
|
|
|
|
html = <<~HTML
|
|
|
|
|
<custom>Text withing custom <span>tag</span></custom>
|
|
|
|
|
<div>Text within allowed tag</div>
|
|
|
|
|
HTML
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq("Text within allowed tag")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "allows additional tags that can be consumed by subclasses" do
|
|
|
|
|
class ExtendedHtmlToMarkdown < HtmlToMarkdown
|
|
|
|
|
def to_markdown
|
|
|
|
|
yield @doc
|
|
|
|
|
super
|
|
|
|
|
end
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
html = <<~HTML
|
|
|
|
|
<custom-image image-id="42">Image text</custom-image>
|
|
|
|
|
<div>Text within allowed tag</div>
|
|
|
|
|
HTML
|
|
|
|
|
|
|
|
|
|
md =
|
|
|
|
|
ExtendedHtmlToMarkdown
|
|
|
|
|
.new(html)
|
|
|
|
|
.to_markdown { |doc| expect(doc.css("custom-image")).to be_empty }
|
|
|
|
|
expect(md).to eq("Text within allowed tag")
|
|
|
|
|
|
|
|
|
|
md =
|
|
|
|
|
ExtendedHtmlToMarkdown
|
|
|
|
|
.new(html, { additional_allowed_tags: ["custom-image"] })
|
|
|
|
|
.to_markdown do |doc|
|
|
|
|
|
doc.css("custom-image").each { |img| img.replace("Image #{img["image-id"]}") }
|
|
|
|
|
end
|
|
|
|
|
expect(md).to eq("Image 42\nText within allowed tag")
|
|
|
|
|
end
|
|
|
|
|
|
2020-06-04 05:26:40 +08:00
|
|
|
it "doesn't error on non-inline elements like (aside, section)" do
|
|
|
|
|
html = <<~HTML
|
|
|
|
|
<aside class="quote no-group">
|
|
|
|
|
<blockquote>
|
2021-09-17 16:41:34 +08:00
|
|
|
<p>Hello,<br>is it me you're looking for?</p>
|
2020-06-04 05:26:40 +08:00
|
|
|
</blockquote>
|
2021-09-17 16:41:34 +08:00
|
|
|
<br>
|
2020-06-04 05:26:40 +08:00
|
|
|
</aside>
|
|
|
|
|
HTML
|
|
|
|
|
|
|
|
|
|
markdown = <<~MD
|
2021-09-17 16:41:34 +08:00
|
|
|
> Hello,
|
|
|
|
|
> is it me you're looking for?
|
2020-06-04 05:26:40 +08:00
|
|
|
MD
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq(markdown.strip)
|
|
|
|
|
end
|
|
|
|
|
|
2017-05-04 01:34:03 +08:00
|
|
|
it "skips hidden tags" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(html_to_markdown("<p>Hello <span hidden>cruel </span>World!</p>")).to eq("Hello World!")
|
2017-05-04 01:34:03 +08:00
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "converts <strong>" do
|
|
|
|
|
expect(html_to_markdown("<strong>Strong</strong>")).to eq("**Strong**")
|
|
|
|
|
expect(html_to_markdown("<strong>Str*ng</strong>")).to eq("__Str*ng__")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "converts <b>" do
|
|
|
|
|
expect(html_to_markdown("<b>Bold</b>")).to eq("**Bold**")
|
|
|
|
|
expect(html_to_markdown("<b>B*ld</b>")).to eq("__B*ld__")
|
2018-12-18 01:38:11 +08:00
|
|
|
|
|
|
|
|
html = <<~HTML
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
Before
|
2018-12-18 01:38:11 +08:00
|
|
|
<p><b>Bold
|
|
|
|
|
<br>
|
|
|
|
|
<br>
|
|
|
|
|
</b>
|
|
|
|
|
</p>
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
After
|
2018-12-18 01:38:11 +08:00
|
|
|
HTML
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(html_to_markdown(html)).to eq("Before\n\n**Bold**\n\nAfter")
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "converts <em>" do
|
|
|
|
|
expect(html_to_markdown("<em>Emphasis</em>")).to eq("*Emphasis*")
|
|
|
|
|
expect(html_to_markdown("<em>Emph*sis</em>")).to eq("_Emph*sis_")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "converts <i>" do
|
|
|
|
|
expect(html_to_markdown("<i>Italic</i>")).to eq("*Italic*")
|
|
|
|
|
expect(html_to_markdown("<i>It*lic</i>")).to eq("_It*lic_")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "converts <a>" do
|
|
|
|
|
expect(html_to_markdown(%Q{<a href="https://www.discourse.org">Discourse</a>})).to eq(
|
|
|
|
|
"[Discourse](https://www.discourse.org)",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
it "supports SiteSetting.allowed_href_schemes" do
|
|
|
|
|
SiteSetting.allowed_href_schemes = "tel|steam"
|
|
|
|
|
expect(html_to_markdown(%Q{<a href="steam://store/48000">LIMBO</a>})).to eq(
|
|
|
|
|
"[LIMBO](steam://store/48000)",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
2017-05-03 22:42:37 +08:00
|
|
|
it "removes empty & invalid <a>" do
|
|
|
|
|
expect(html_to_markdown("<a>Discourse</a>")).to eq("Discourse")
|
|
|
|
|
expect(html_to_markdown(%Q{<a href="">Discourse</a>})).to eq("Discourse")
|
|
|
|
|
expect(html_to_markdown(%Q{<a href="foo.bar">Discourse</a>})).to eq("Discourse")
|
|
|
|
|
end
|
|
|
|
|
|
2024-11-06 06:27:49 +08:00
|
|
|
HTML_WITH_IMG = %Q{<img src="https://www.discourse.org/logo.svg" alt="Discourse Logo">}
|
|
|
|
|
HTML_WITH_CID_IMG = %Q{<img src="cid:ii_1525434659ddb4cb" title="Discourse Logo">}
|
2017-04-29 04:14:46 +08:00
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "converts <img>" do
|
2017-04-29 04:14:46 +08:00
|
|
|
expect(html_to_markdown(HTML_WITH_IMG)).to eq(
|
|
|
|
|
"",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "keeps <img> with 'keep_img_tags'" do
|
2017-05-04 04:53:47 +08:00
|
|
|
expect(html_to_markdown(HTML_WITH_IMG, keep_img_tags: true)).to eq(HTML_WITH_IMG)
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
2024-01-31 08:23:09 +08:00
|
|
|
it "removes newlines from img alt text" do
|
|
|
|
|
html_with_alt_newlines =
|
|
|
|
|
%Q{<img src="https://www.discourse.org/logo.svg" alt="Discourse\n\nLogo">}
|
|
|
|
|
expect(html_to_markdown(html_with_alt_newlines)).to eq(
|
|
|
|
|
"",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
2017-05-04 00:29:25 +08:00
|
|
|
it "removes empty & invalid <img>" do
|
|
|
|
|
expect(html_to_markdown("<img>")).to eq("")
|
|
|
|
|
expect(html_to_markdown(%Q{<img src="">})).to eq("")
|
|
|
|
|
expect(html_to_markdown(%Q{<img src="foo.bar">})).to eq("")
|
|
|
|
|
end
|
|
|
|
|
|
2021-05-21 09:43:47 +08:00
|
|
|
it "keeps <img> with src='cid:' with 'keep_cid_imgs'" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(html_to_markdown(HTML_WITH_CID_IMG, keep_cid_imgs: true)).to eq(HTML_WITH_CID_IMG)
|
2017-05-04 04:53:47 +08:00
|
|
|
end
|
|
|
|
|
|
2024-01-31 08:23:09 +08:00
|
|
|
it "removes newlines from img alt text with cid images" do
|
|
|
|
|
html_with_cid_alt_newlines = %Q{<img src="cid:ii_1525434659ddb4cb" title="Discourse\n\nLogo">}
|
|
|
|
|
expect(html_to_markdown(html_with_cid_alt_newlines, keep_cid_imgs: true)).to eq(
|
|
|
|
|
%Q{<img src="cid:ii_1525434659ddb4cb" title="Discourse Logo">},
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
2017-05-04 01:40:34 +08:00
|
|
|
it "skips hidden <img>" do
|
|
|
|
|
expect(html_to_markdown(%Q{<img src="https://www.discourse.org/logo.svg" width=0>})).to eq("")
|
|
|
|
|
expect(html_to_markdown(%Q{<img src="https://www.discourse.org/logo.svg" height="0">})).to eq(
|
|
|
|
|
"",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
it "supports width/height on <img>" do
|
|
|
|
|
expect(html_to_markdown(%Q{<img src="https://www.discourse.org/logo.svg" height=100>})).to eq(
|
|
|
|
|
"",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown(%Q{<img src="https://www.discourse.org/logo.svg" width=200>})).to eq(
|
|
|
|
|
"",
|
|
|
|
|
)
|
|
|
|
|
expect(
|
|
|
|
|
html_to_markdown(%Q{<img src="https://www.discourse.org/logo.svg" height=100 width=200>}),
|
|
|
|
|
).to eq("")
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
(1..6).each do |n|
|
|
|
|
|
it "converts <h#{n}>" do
|
|
|
|
|
expect(html_to_markdown("<h#{n}>Header #{n}</h#{n}>")).to eq("#" * n + " Header #{n}")
|
|
|
|
|
end
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "converts <br>" do
|
|
|
|
|
expect(html_to_markdown("Before<br>Inside<br>After")).to eq("Before\nInside\nAfter")
|
|
|
|
|
end
|
|
|
|
|
|
2019-04-14 17:14:54 +08:00
|
|
|
it "skips <br> inside <p> if next character is \n" do
|
|
|
|
|
expect(html_to_markdown("<p>Before<br>\nInside<br>After</p>")).to eq("Before\nInside\nAfter")
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "converts <hr>" do
|
|
|
|
|
expect(html_to_markdown("Before<hr>Inside<hr>After")).to eq(
|
|
|
|
|
"Before\n\n---\n\nInside\n\n---\n\nAfter",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "converts <tt>" do
|
|
|
|
|
expect(html_to_markdown("<tt>Teletype</tt>")).to eq("`Teletype`")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "converts <code>" do
|
|
|
|
|
expect(html_to_markdown("<code>Code</code>")).to eq("`Code`")
|
|
|
|
|
end
|
|
|
|
|
|
2024-06-09 20:17:32 +08:00
|
|
|
describe "when HTML is used within Markdown" do
|
|
|
|
|
HtmlToMarkdown::ALLOWED.each do |tag|
|
|
|
|
|
it "keeps mandatory HTML entities in text of <#{tag}>" do
|
|
|
|
|
expect(html_to_markdown("<#{tag}>Less than: <</#{tag}>")).to eq(
|
|
|
|
|
"<#{tag}>Less than: <</#{tag}>",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("<#{tag}>Greater than: >")).to eq(
|
|
|
|
|
"<#{tag}>Greater than: ></#{tag}>",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("<#{tag}>Ampersand: &")).to eq(
|
|
|
|
|
"<#{tag}>Ampersand: &</#{tag}>",
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown("<#{tag}>Double Quote: "</#{tag}>")).to eq(
|
|
|
|
|
"<#{tag}>Double Quote: \"</#{tag}>",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("<#{tag}>Single Quote: '</#{tag}>")).to eq(
|
|
|
|
|
"<#{tag}>Single Quote: '</#{tag}>",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("<#{tag}>Copyright Symbol: ©</#{tag}>")).to eq(
|
|
|
|
|
"<#{tag}>Copyright Symbol: ©</#{tag}>",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("<#{tag}>Euro Symbol: €</#{tag}>")).to eq(
|
|
|
|
|
"<#{tag}>Euro Symbol: €</#{tag}>",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
end
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "supports <ins>" do
|
|
|
|
|
expect(html_to_markdown("This is an <ins>insertion</ins>")).to eq(
|
|
|
|
|
"This is an <ins>insertion</ins>",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <del>" do
|
|
|
|
|
expect(html_to_markdown("This is a <del>deletion</del>")).to eq("This is a <del>deletion</del>")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <sub>" do
|
|
|
|
|
expect(html_to_markdown("H<sub>2</sub>O")).to eq("H<sub>2</sub>O")
|
|
|
|
|
end
|
|
|
|
|
|
2021-02-16 00:17:30 +08:00
|
|
|
it "supports <mark>" do
|
|
|
|
|
expect(html_to_markdown("<mark>This is highlighted!</mark>")).to eq(
|
|
|
|
|
"<mark>This is highlighted!</mark>",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "supports <sup>" do
|
|
|
|
|
expect(html_to_markdown("<sup>Super Script!</sup>")).to eq("<sup>Super Script!</sup>")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <small>" do
|
|
|
|
|
expect(html_to_markdown("<small>Small</small>")).to eq("<small>Small</small>")
|
2024-06-09 20:17:32 +08:00
|
|
|
expect(html_to_markdown("<mark><small>Small</small></mark>")).to eq(
|
|
|
|
|
"<mark><small>Small</small></mark>",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("<strong><small>Small</small></strong>")).to eq(
|
|
|
|
|
"**<small>Small</small>**",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("<small><strong><small></strong></small>")).to eq(
|
|
|
|
|
"<small>**<small>**</small>",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <big>" do
|
|
|
|
|
expect(html_to_markdown("<big>Big</big>")).to eq("<big>Big</big>")
|
|
|
|
|
expect(html_to_markdown("<big><big></big>")).to eq("<big><big></big>")
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <kbd>" do
|
|
|
|
|
expect(html_to_markdown("<kbd>CTRL</kbd>+<kbd>C</kbd>")).to eq("<kbd>CTRL</kbd>+<kbd>C</kbd>")
|
2024-06-09 20:17:32 +08:00
|
|
|
expect(html_to_markdown("<kbd><</kbd>")).to eq("<kbd><</kbd>")
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <abbr>" do
|
|
|
|
|
expect(
|
|
|
|
|
html_to_markdown(%Q{<abbr title="Civilized Discourse Construction Kit, Inc.">CDCK</abbr>}),
|
|
|
|
|
).to eq(%Q{<abbr title="Civilized Discourse Construction Kit, Inc.">CDCK</abbr>})
|
2024-06-09 20:17:32 +08:00
|
|
|
|
|
|
|
|
expect(
|
|
|
|
|
html_to_markdown(
|
|
|
|
|
%Q{<abbr title=""abbr": The Abbreviation element"><abbr></abbr>},
|
|
|
|
|
),
|
|
|
|
|
).to eq(%Q{<abbr title=""abbr": The Abbreviation element"><abbr></abbr>})
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <s>" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(html_to_markdown("<s>Strike Through</s>")).to eq("~~Strike Through~~")
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <strike>" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(html_to_markdown("<strike>Strike Through</strike>")).to eq("~~Strike Through~~")
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <blockquote>" do
|
|
|
|
|
expect(html_to_markdown("<blockquote>Quote</blockquote>")).to eq("> Quote")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <ul>" do
|
|
|
|
|
expect(html_to_markdown("<ul><li>🍏</li><li>🍐</li><li>🍌</li></ul>")).to eq("- 🍏\n- 🍐\n- 🍌")
|
|
|
|
|
expect(html_to_markdown("<ul>\n<li>🍏</li>\n<li>🍐</li>\n<li>🍌</li>\n</ul>")).to eq(
|
|
|
|
|
"- 🍏\n- 🍐\n- 🍌",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <ol>" do
|
|
|
|
|
expect(html_to_markdown("<ol><li>🍆</li><li>🍅</li><li>🍄</li></ol>")).to eq("1. 🍆\n1. 🍅\n1. 🍄")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <p> inside <li>" do
|
|
|
|
|
expect(html_to_markdown("<ul><li><p>🍏</p></li><li><p>🍐</p></li><li><p>🍌</p></li></ul>")).to eq(
|
|
|
|
|
"- 🍏\n\n- 🍐\n\n- 🍌",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "supports <ul> inside <ul>" do
|
|
|
|
|
expect(html_to_markdown(<<-HTML)).to eq(
|
|
|
|
|
<ul>
|
|
|
|
|
<li>Fruits
|
|
|
|
|
<ul>
|
|
|
|
|
<li>🍏</li>
|
|
|
|
|
<li>🍐</li>
|
|
|
|
|
<li>🍌</li>
|
|
|
|
|
</ul>
|
|
|
|
|
</li>
|
|
|
|
|
<li>Vegetables
|
|
|
|
|
<ul>
|
|
|
|
|
<li>🍆</li>
|
|
|
|
|
<li>🍅</li>
|
|
|
|
|
<li>🍄</li>
|
|
|
|
|
</ul>
|
|
|
|
|
</li>
|
|
|
|
|
</ul>
|
|
|
|
|
HTML
|
|
|
|
|
"- Fruits\n - 🍏\n - 🍐\n - 🍌\n- Vegetables\n - 🍆\n - 🍅\n - 🍄",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
2017-05-17 21:05:11 +08:00
|
|
|
it "supports bare <li>" do
|
|
|
|
|
expect(html_to_markdown("<li>I'm alone</li>")).to eq("- I'm alone")
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "supports <pre>" do
|
|
|
|
|
expect(html_to_markdown("<pre>var foo = 'bar';</pre>")).to eq("```\nvar foo = 'bar';\n```")
|
|
|
|
|
expect(html_to_markdown("<pre><code>var foo = 'bar';</code></pre>")).to eq(
|
|
|
|
|
"```\nvar foo = 'bar';\n```",
|
|
|
|
|
)
|
|
|
|
|
expect(
|
|
|
|
|
html_to_markdown(%Q{<pre><code class="lang-javascript">var foo = 'bar';</code></pre>}),
|
|
|
|
|
).to eq("```javascript\nvar foo = 'bar';\n```")
|
2020-03-19 01:31:10 +08:00
|
|
|
expect(
|
|
|
|
|
html_to_markdown(
|
|
|
|
|
"<pre> function f() {\n console.log('Hello world!');\n }</pre>",
|
2023-01-09 19:18:21 +08:00
|
|
|
),
|
2020-03-19 01:31:10 +08:00
|
|
|
).to eq("```\n function f() {\n console.log('Hello world!');\n }\n```")
|
2024-06-09 20:17:32 +08:00
|
|
|
|
|
|
|
|
html = <<~HTML
|
|
|
|
|
<pre data-code-wrap="plaintext"><code class="lang-plaintext">Reported-and-tested-by: A <a@example.com>
|
|
|
|
|
Reviewed-by: B <b@example.com></code></pre>
|
|
|
|
|
HTML
|
|
|
|
|
md = <<~MD
|
|
|
|
|
```plaintext
|
|
|
|
|
Reported-and-tested-by: A <a@example.com>
|
|
|
|
|
Reviewed-by: B <b@example.com>
|
|
|
|
|
```
|
|
|
|
|
MD
|
|
|
|
|
expect(html_to_markdown(html)).to eq(md.strip)
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
2018-02-27 06:28:02 +08:00
|
|
|
it "supports <pre> inside <blockquote>" do
|
|
|
|
|
expect(
|
|
|
|
|
html_to_markdown("<blockquote><pre><code>var foo = 'bar';</code></pre></blockquote>"),
|
|
|
|
|
).to eq("> ```\n> var foo = 'bar';\n> ```")
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "works" do
|
|
|
|
|
expect(
|
|
|
|
|
html_to_markdown(
|
|
|
|
|
"<ul><li><p>A list item with a blockquote:</p><blockquote><p>This is a <strong>blockquote</strong><br>inside a list item.</p></blockquote></li></ul>",
|
2023-01-09 19:18:21 +08:00
|
|
|
),
|
2017-04-25 04:01:41 +08:00
|
|
|
).to eq(
|
|
|
|
|
"- A list item with a blockquote:\n\n > This is a **blockquote**\n > inside a list item.",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-29 04:02:20 +08:00
|
|
|
it "supports html document" do
|
|
|
|
|
expect(html_to_markdown("<html><body>Hello<div>World</div></body></html>")).to eq(
|
|
|
|
|
"Hello\nWorld",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "handles <p>" do
|
|
|
|
|
expect(html_to_markdown("<p>1st paragraph</p><p>2nd paragraph</p>")).to eq(
|
|
|
|
|
"1st paragraph\n\n2nd paragraph",
|
|
|
|
|
)
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(
|
|
|
|
|
html_to_markdown(
|
|
|
|
|
"<body><p>1st paragraph</p>\n <p> 2nd paragraph\n 2nd paragraph</p>\n<p>3rd paragraph</p></body>",
|
2023-01-09 19:18:21 +08:00
|
|
|
),
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
).to eq("1st paragraph\n\n2nd paragraph 2nd paragraph\n\n3rd paragraph")
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "handles <div>" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(html_to_markdown("<div>1st div</div><div>2nd div</div>")).to eq("1st div\n2nd div")
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "swallows <span>" do
|
|
|
|
|
expect(html_to_markdown("<span>Span</span>")).to eq("Span")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "swallows <u>" do
|
|
|
|
|
expect(html_to_markdown("<u>Underline</u>")).to eq("Underline")
|
|
|
|
|
end
|
|
|
|
|
|
2024-06-09 19:32:52 +08:00
|
|
|
it "swallows <center>" do
|
|
|
|
|
expect(html_to_markdown("<center>Centered</center>")).to eq("Centered")
|
|
|
|
|
end
|
|
|
|
|
|
2017-04-25 04:01:41 +08:00
|
|
|
it "removes <script>" do
|
|
|
|
|
expect(html_to_markdown("<script>var foo = 'bar'</script>")).to eq("")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "removes <style>" do
|
|
|
|
|
expect(html_to_markdown("<style>* { margin: 0 }</style>")).to eq("")
|
|
|
|
|
end
|
|
|
|
|
|
2018-12-18 01:39:02 +08:00
|
|
|
it "handles <p> and <div> within <span>" do
|
|
|
|
|
html = "<div>1st paragraph<span><div>2nd paragraph</div><p>3rd paragraph</p></span></div>"
|
|
|
|
|
expect(html_to_markdown(html)).to eq("1st paragraph\n2nd paragraph\n\n3rd paragraph")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "handles <p> and <div> within <font>" do
|
|
|
|
|
html =
|
|
|
|
|
"<font>1st paragraph<br><span>2nd paragraph</span><div>3rd paragraph</div><p>4th paragraph</p></font>"
|
|
|
|
|
expect(html_to_markdown(html)).to eq(
|
|
|
|
|
"1st paragraph\n2nd paragraph\n3rd paragraph\n\n4th paragraph",
|
|
|
|
|
)
|
2017-05-10 00:33:54 +08:00
|
|
|
end
|
|
|
|
|
|
2017-08-03 05:02:59 +08:00
|
|
|
context "with an oddly placed <br>" do
|
|
|
|
|
it "handles <strong>" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(html_to_markdown("Hello <strong><br>Bold</strong> World")).to eq(
|
|
|
|
|
"Hello\n**Bold** World",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("Hello <strong>Bold<br></strong> World")).to eq(
|
|
|
|
|
"Hello **Bold**\nWorld",
|
|
|
|
|
)
|
|
|
|
|
expect(html_to_markdown("Hello <strong>Bold<br>text</strong> World")).to eq(
|
|
|
|
|
"Hello **Bold**\n**text** World",
|
|
|
|
|
)
|
2017-08-03 05:02:59 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "handles <em>" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
expect(html_to_markdown("Hello <em><br>Italic</em> World")).to eq("Hello\n*Italic* World")
|
|
|
|
|
expect(html_to_markdown("Hello <em>Italic<br></em> World")).to eq("Hello *Italic*\nWorld")
|
|
|
|
|
expect(html_to_markdown("Hello <em>Italic<br>text</em> World")).to eq(
|
|
|
|
|
"Hello *Italic*\n*text* World",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "works" do
|
|
|
|
|
expect(html_to_markdown("<div>A <b> B <i> C <br> D </i> E <br> F </b> G</div>")).to eq(
|
|
|
|
|
"A __B *C*__\n__*D* E__\n**F** G",
|
|
|
|
|
)
|
2017-08-03 05:02:59 +08:00
|
|
|
end
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
context "with an empty tag" do
|
|
|
|
|
it "handles <strong>" do
|
|
|
|
|
expect(html_to_markdown("<strong></strong>")).to eq("")
|
|
|
|
|
expect(html_to_markdown("<strong> </strong>")).to eq("")
|
2017-08-15 04:13:24 +08:00
|
|
|
expect(html_to_markdown("Some<strong> </strong>text")).to eq("Some text")
|
|
|
|
|
expect(html_to_markdown("Some<strong> </strong>text")).to eq("Some text")
|
2017-08-03 05:02:59 +08:00
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "handles <em>" do
|
|
|
|
|
expect(html_to_markdown("<em></em>")).to eq("")
|
|
|
|
|
expect(html_to_markdown("<em> </em>")).to eq("")
|
2017-08-15 04:13:24 +08:00
|
|
|
expect(html_to_markdown("Some<em> </em>text")).to eq("Some text")
|
|
|
|
|
expect(html_to_markdown("Some<em> </em>text")).to eq("Some text")
|
|
|
|
|
end
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
context "with spaces around text" do
|
|
|
|
|
it "handles <strong>" do
|
|
|
|
|
expect(html_to_markdown("<strong> Bold</strong>")).to eq("**Bold**")
|
|
|
|
|
expect(html_to_markdown("<strong> Bold</strong>")).to eq("**Bold**")
|
|
|
|
|
expect(html_to_markdown("<strong>Bold </strong>")).to eq("**Bold**")
|
|
|
|
|
expect(html_to_markdown("<strong>Bold </strong>")).to eq("**Bold**")
|
|
|
|
|
expect(html_to_markdown("Some<strong> bold</strong> text")).to eq("Some **bold** text")
|
|
|
|
|
expect(html_to_markdown("Some<strong> bold</strong> text")).to eq("Some **bold** text")
|
|
|
|
|
expect(html_to_markdown("Some <strong>bold </strong>text")).to eq("Some **bold** text")
|
|
|
|
|
expect(html_to_markdown("Some <strong>bold </strong>text")).to eq("Some **bold** text")
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "handles <em>" do
|
|
|
|
|
expect(html_to_markdown("<em> Italic</em>")).to eq("*Italic*")
|
|
|
|
|
expect(html_to_markdown("<em> Italic</em>")).to eq("*Italic*")
|
|
|
|
|
expect(html_to_markdown("<em>Italic </em>")).to eq("*Italic*")
|
|
|
|
|
expect(html_to_markdown("<em>Italic </em>")).to eq("*Italic*")
|
|
|
|
|
expect(html_to_markdown("Some<em> italic</em> text")).to eq("Some *italic* text")
|
|
|
|
|
expect(html_to_markdown("Some<em> italic</em> text")).to eq("Some *italic* text")
|
|
|
|
|
expect(html_to_markdown("Some <em>italic </em>text")).to eq("Some *italic* text")
|
|
|
|
|
expect(html_to_markdown("Some <em>italic </em>text")).to eq("Some *italic* text")
|
2017-08-03 05:02:59 +08:00
|
|
|
end
|
|
|
|
|
end
|
|
|
|
|
|
2021-05-21 09:43:47 +08:00
|
|
|
it "supports <table>" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
html = <<~HTML
|
|
|
|
|
<table>
|
|
|
|
|
<thead>
|
|
|
|
|
<tr>
|
|
|
|
|
<th>This</th>
|
|
|
|
|
<th>is</th>
|
|
|
|
|
<th>the</th>
|
|
|
|
|
<th><i>headers</i></th>
|
|
|
|
|
</tr>
|
|
|
|
|
</thead>
|
|
|
|
|
<tbody>
|
|
|
|
|
<tr>
|
|
|
|
|
<td>I am</td>
|
|
|
|
|
<td>the</td>
|
|
|
|
|
<td><b>first</b></td>
|
|
|
|
|
<td>row</td>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
|
|
|
|
<td>And this</td>
|
|
|
|
|
<td>is the</td>
|
|
|
|
|
<td>2<sup>nd</sup></td>
|
|
|
|
|
<td>line</td>
|
|
|
|
|
</tr>
|
|
|
|
|
</tbody>
|
2024-06-09 19:26:39 +08:00
|
|
|
<tfoot>
|
|
|
|
|
<tr>
|
|
|
|
|
<td>This</td>
|
|
|
|
|
<td>is</td>
|
|
|
|
|
<td>the</td>
|
|
|
|
|
<td>footer</td>
|
|
|
|
|
</tr>
|
|
|
|
|
</tfoot>
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
</table>
|
|
|
|
|
HTML
|
|
|
|
|
|
|
|
|
|
markdown = <<~MD
|
|
|
|
|
| This | is | the | *headers* |
|
|
|
|
|
| - | - | - | - |
|
|
|
|
|
| I am | the | **first** | row |
|
|
|
|
|
| And this | is the | 2<sup>nd</sup> | line |
|
2024-06-09 19:26:39 +08:00
|
|
|
| This | is | the | footer |
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
MD
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq(markdown.strip)
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown("<table><tr><td>Hello</td><td>World</td></tr></table>")).to eq(
|
|
|
|
|
"| Hello | World |\n| - | - |",
|
|
|
|
|
)
|
|
|
|
|
end
|
|
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
it "keeps HTML for badly formatted <table>" do
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
html = <<~HTML
|
|
|
|
|
<table>
|
|
|
|
|
<tr>
|
|
|
|
|
<th>1</th>
|
|
|
|
|
<th>2</th>
|
|
|
|
|
<th>3</th>
|
|
|
|
|
<th>4</th>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
2024-06-09 19:11:37 +08:00
|
|
|
<td><One></td>
|
|
|
|
|
<td><strong>Two</strong></td>
|
|
|
|
|
<td>Three<script>alert("foo")</script></td>
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
</tr>
|
|
|
|
|
</table>
|
|
|
|
|
HTML
|
|
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
markdown = <<~MD
|
|
|
|
|
<table>
|
|
|
|
|
<tr>
|
|
|
|
|
<th>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
1
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</th>
|
|
|
|
|
<th>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
2
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</th>
|
|
|
|
|
<th>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
3
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</th>
|
|
|
|
|
<th>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
4
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</th>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
|
|
|
|
<td>
|
|
|
|
|
|
|
|
|
|
<One>
|
|
|
|
|
|
|
|
|
|
</td>
|
|
|
|
|
<td>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
**Two**
|
|
|
|
|
|
|
|
|
|
</td>
|
|
|
|
|
<td>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
Three
|
|
|
|
|
|
|
|
|
|
</td>
|
|
|
|
|
</tr>
|
|
|
|
|
</table>
|
|
|
|
|
MD
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq(markdown.strip)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "keeps HTML for <table> with colspan" do
|
|
|
|
|
html = <<~HTML
|
|
|
|
|
<table>
|
|
|
|
|
<tr>
|
|
|
|
|
<th>1</th>
|
|
|
|
|
<th>2</th>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
|
|
|
|
<td colspan="2">One / Two</td>
|
|
|
|
|
</tr>
|
|
|
|
|
</table>
|
|
|
|
|
HTML
|
|
|
|
|
|
|
|
|
|
markdown = <<~MD
|
|
|
|
|
<table>
|
|
|
|
|
<tr>
|
|
|
|
|
<th>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
1
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</th>
|
|
|
|
|
<th>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
2
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</th>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
|
|
|
|
<td colspan="2">
|
|
|
|
|
|
|
|
|
|
One / Two
|
|
|
|
|
|
|
|
|
|
</td>
|
|
|
|
|
</tr>
|
|
|
|
|
</table>
|
|
|
|
|
MD
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq(markdown.strip)
|
|
|
|
|
end
|
|
|
|
|
|
|
|
|
|
it "keeps HTML for <table> with rowspan" do
|
|
|
|
|
html = <<~HTML
|
|
|
|
|
<table>
|
|
|
|
|
<tr>
|
|
|
|
|
<th>1</th>
|
|
|
|
|
<th>2</th>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
|
|
|
|
<td>A</td>
|
|
|
|
|
<td rowspan="2">B</td>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
|
|
|
|
<td>C</td>
|
|
|
|
|
</tr>
|
|
|
|
|
</table>
|
|
|
|
|
HTML
|
|
|
|
|
|
|
|
|
|
markdown = <<~MD
|
|
|
|
|
<table>
|
|
|
|
|
<tr>
|
|
|
|
|
<th>
|
|
|
|
|
|
|
|
|
|
1
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</th>
|
|
|
|
|
<th>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
2
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</th>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
|
|
|
|
<td>
|
|
|
|
|
|
|
|
|
|
A
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</td>
|
|
|
|
|
<td rowspan="2">
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
B
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</td>
|
|
|
|
|
</tr>
|
|
|
|
|
<tr>
|
|
|
|
|
<td>
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
C
|
2024-06-09 19:26:39 +08:00
|
|
|
|
2024-06-09 19:11:37 +08:00
|
|
|
</td>
|
|
|
|
|
</tr>
|
|
|
|
|
</table>
|
|
|
|
|
MD
|
|
|
|
|
|
|
|
|
|
expect(html_to_markdown(html)).to eq(markdown.strip)
|
FIX: server-side HtmlToMarkdown improvements (#9586)
TLDR; this commit vastly improves how whitespaces are handled when converting from HTML to Markdown.
It also adds support for converting HTML <tables> to markdown tables.
The previous 'remove_whitespaces!' method was traversing the whole HTML tree and used a heuristic to remove
leading and trailing whitespaces whenever it was appropriate (ie. mostly before and after HTML block elements)
It was a good idea, but it was very limited and leaded to bad conversion when the html had leading whitespaces on several lines for example.
One such example can be found [here](https://meta.discourse.org/t/86782).
For various reasons, most of the whitespaces in a HTML file is ignored when the page is being displayed in a browser.
The rules that the browsers follow are the [CSS' White Space Processing Rules](https://www.w3.org/TR/css-text-3/#white-space-rules).
They can be quite complicated when you take into account RTL languages and other various tidbits but they boils down to the following:
- Collapse whitespaces down to one space (0x20) inside an inline context (ie. nodes/tags that are being displaying on the same line)
- Remove any leading/trailing whitespaces inside an inline context
One quick & dirty way of getting this 90% solved would be to do 'HTML.gsub!(/[[:space:]]+/, " ")'.
We would also need to hoist <pre> elements in order to not mess with their whitespaces.
Unfortunately, this solution let some whitespaces creep around HTML tags which leads to more '.strip!' calls than I can bear.
I decided to "emulate" the browser's handling of whitespaces and came up with a solution in 4 parts
1. remove_not_allowed!
The HtmlToMarkdown library is recursively "visiting" all the nodes in the HTML in order to convert them to Markdown.
All the nodes that aren't handled by the library (eg. <script>, <style> or any non-textual HTML tags) are "swallowed".
In order to reduce the number of nodes visited, the method 'remove_not_allowed!' will automatically delete all the nodes
that have no "visitor" (eg. a 'visit_<tag>' method) defined.
2. remove_hidden!
Similar purpose as the previous method (eg. reducing number of nodes visited), there's no point trying to convert something that is hidden.
The 'remove_hidden!' method removes any nodes that was hidden using the "hidden" HTML attribute, some CSS or with a width or height equal to 0.
3. hoist_line_breaks!
The 'hoist_line_breaks!' method is there to handle <br> tags. I know those tiny <br> don't do much but they can be quite annoying.
The <br> tags are inline elements but they visually work like a block element (ie. they create a new line).
If you have the following HTML "<i>Foo<br>Bar</i>", it ends up visually similar to "<i>Foo</i><br><i>Bar</i>".
The latter being much more easy to process than the former, so that's what this method is doing.
The "hoist_line_breaks" will hoist <br> tags out of inline tags until their parent is a block element.
4. remove_whitespaces!
The "remove_whitespaces!" is where all the whitespace removal is happening. It's broken down into 4 methods as well
- remove_whitespaces!
- is_inline?
- collapse_spaces!
- remove_trailing_space!
The 'remove_whitespace!' method is recursively walking the HTML tree (skipping <pre> tags).
If a node has any children, they will be chunked into groups of inline elements vs block elements.
For each chunks of inline elements, it will call the "collapse_space!" and "remove_trailing_space!" methods.
For each chunks of block elements, it will call "remote_whitespace!" to keep walking the HTML tree recursively.
The "is_inline?" method determines whether a node is part of a inline context.
A node is inline iif it's a text node or it's an inline tag, but not <br>, and all its children are also inline.
The "collapse_spaces!" method will collapse any kind of (white) space into a single space (" ") character, even accros tags.
For example, if we have " Foo \n<i> Bar </i>\t42", it will return "Foo <i>Bar </i>42".
Finally, the "remove_trailing_space!" method is there to remove any trailing space that might creep in at the end of the inline chunk.
This solution is not 100% bullet-proof.
It does not support RTL languages at all and has some caveats that I felt were not worth the work to get properly fixed.
FIX: better detection of hidden elements when converting HTML to Markdown
FIX: take into account the 'allowed_href_schemes' site setting when converting HTML <a> to Markdown
FIX: added support for 'mailto:' scheme when converting <a> from HTML to Markdown
FIX: added support for <img> dimensions when converting from HTML to Markdown
FIX: added support for <dl>, <dd> and <dt> when converting from HTML to Markdown
FIX: added support for multilines emphases, strongs and strikes when converting from HTML to Markdown
FIX: added support for <acronym> when converting from HTML to Markdown
DEV: remove unused 'sanitize' gem
Wow, did you just read all that?! Congratz, here's a cookie: 🍪.
2020-04-30 18:21:25 +08:00
|
|
|
end
|
2017-04-25 04:01:41 +08:00
|
|
|
end
|