Commit 1414966c authored by Alex Theg's avatar Alex Theg
Browse files

add documentation

parent 91141c59
# Editoria Typescript
Editoria Typescript transforms HTML into a format required for the Coko Foundation's <a href="">Wax</a> WYSIWYG word processor for <a href="">Editoria</a>. While Wax has been built specifically for book editing and publication, it is by no means its only application, and it could be repurposed. Other similar chains could be implemented to target another format.
Editoria Typescript translates the document structure, inline and class formatting, endnotes and footnotes into a subset of near-HTML, while eliminating HTML attributes not used by Wax.
## Contents
* [Pipeline](#pipe)
* [p-split-around-br.xsl](#br)
* [editoria-basic.xsl](#basic)
* [editoria-reduce.xsl](#reduce)
## <a name="pipe"></a>Pipeline
Editoria Typescript should be run in the following order:
1. `p-split-around-br.xsl`
2. `editoria-basic.xsl`
3. `<span class="s1">editoria-reduce.xsl</span>`
### <a name="br"></a>`p-split-around-br.xsl`
It is possible to specify line breaks within paragraphs in Word (`<w:br/>`, which are extracted as XHTML `<br class="br" />` tags).
As Wax does not support `<br>`s, this step simply divides paragraphs on breaks, removing the break and creating two separate `<p>` elements instead.
<p style="font-family: Times New Roman; text-indent: 36pt">
Kṛṣṇadevarāya discusses this practice in the following verse:
<br class="br"/>
Make trustworthy Brahmins
<p>Kṛṣṇadevarāya discusses this practice in the following verse:</p>
<p>Make trustworthy Brahmins</p>
### <a name="basic"></a>`editoria-basic.xsl`
XSweet's initial extraction divides the contents of the HTML document into sections: `<div class-"docx-content">``<div class-"docx-endnotes">`, and `<div class-"docx-footnotes">` . This step rearranges the content:
* `<div class="docx-content">` becomes `<container id="main">`
* Notes are reformatted and moved into a `<div id="notes">`
Notes and their `id`s are also rewritten, from:
<div class="docx-endnotes">
<div class="docx-endnote" id="en1">
<p class="EndnoteText">
<span class="EndnoteReference">
<span class="endnoteRef">1</span>
</span> endnote</p>
<div class="docx-footnotes">
<div class="docx-footnote" id="fn1">
<p class="FootnoteText">
<span class="FootnoteReference">
<span class="footnoteRef">a</span>
</span> footnote</p>
<div id="notes">
<note-container id="container-en1">
<p class="EndnoteText"> endnote</p>
<note-container id="container-fn1">
<p class="FootnoteText"> footnote</p>
These are then properly linked and nicely displayed in Wax. Endnotes and footnotes are combined into one sequential list:
<img class="aligncenter size-large wp-image-618" src="" alt="" />
`editoria-basic.xsl` writes some properties from CSS `style` attributes inline:
* `font-style: italic` is written to inline elements wrapped in an `<em>` tag
* `font-weight: bold` is written inline as `<strong>` tags
* `text-decoration: underline` is written inline as `<i>` tags
The following inline formatting tag mapping then occur:
* `<b>`s are converted to `<strong>`
* `<u>` is converted to `<i>`
* `<i>` is then converted to `<em>`
Note that we have made the decision convert underlining to italics. Wax does not currently support underlining.
### <a name="reduce"></a>`editoria-reduce.xsl`
* All `class` and `style` information is dropped. Bye bye `class`, bye-bye `style`!
* `<p class="EndnoteText"> endnote</p>` becomes `<p> endnote</p>`
* Other tag attributes (e.g. `id`) are passed through
* `<sub>` and `<sup>` tags are passed through
* Inline markup on whitespace only (spaces, tabs) is removed, e.g. `<b> <b>`
* `tabs` are removed: `<span class="tab">`
* Paragraphs or headings with only whitespace or no content at all are removed, e.g. `<p></p>`, `<p> </p>`, `<h1></h1>`
* Internal-to-Word bookmarks are removed - [see this example](
* `<head><style>` tag is removed
# HTMLevator
HTMLevator is a series of enhancement utilities for improving HTML documents. These tools can be used on HTML files made from `.docx`s with XSweet, or on other, arbitrary HTML files. HTMLevator features include:
* Semantic structure inferring, adding headings and sections to flat HTML files
* Copyediting cleanups to normalize text
* Tools to add customized transformations in simplified syntax
## Contents
* [Heading promotion](#heading)
* [Formatting-based analysis](#format-based")
* [Outline-level heading promotion](#outline-based")
* [Custom configuration](#custom-hp")
* [Plain text URL linking](#url)
* [Plain text list tagging](#list-detection)
* [Copyediting cleanup](#ucp)
* [ucp-text-macros.xsl](#text-cleanup")
* [ucp-mappings.xsl](#ucp-mappings")
* [Custom transformations (experimental)](#custom-trans)
* [Section inferrer (experimental)](#sections)
## <a name="heading"></a>Heading promotion
HTMLevator includes a feature that attempts to infer which elements are headings, transforming them  from `<p>`s into headings: `<h1>` through `<h6>`. This is more art than science, as the input is generally not semantically tagged and structured. It is sometimes trivial to infer headers but it is also frequently quite difficult or impossible to do so unassisted or programmatically. As such, heading promotion will not catch all headings all the time, and it will work better on some documents than on others.
There are 3 heading promotion strategies built into XSweet:
1. Format-based analysis (default)
2. Outline-level heading promotion
3. Custom configuration for named Word styles or specific text
The `header-promote/header-promotion-CHOOSE.xsl` sheet will try to pick the best approach to use for a given document:
* If no custom configurations are supplied, `header-promotion-CHOOSE.xsl` checks to see whether outline levels appear to have been used. If outline level data exists, it is used as the basis for heading promotion
* If outline levels have not been used, format-based analysis is used to infer and promote headings
Alternatively, you can specify the header promotion method to use by passing it as a runtime parameter with `header-promotion-CHOOSE.xsl`:
* `method=ranked-format`
* `method=outline-level`
* `method=my-styles.xml`
### <a name="format-based"></a>Format-based analysis
As a rule, authors indicate headings with visual formatting far more commonly than by applying named MS Word styles. It's not possible to have a discrete list of what kind of formatting indicates a heading, as it changes from file to file and is highly contextual. Instead, each individual document and its formatting must be analyzed as a whole before making guesses about headings. Format-based heading promotion does just this.
This approach works well for some documents and poorly for others. One size does not fit all, and the approach is simply to optimize for what works well with the greatest number of documents. Table of contents and reference files often contain many short paragraphs, leading to erroneous heading promotion.
The `header-promote/digest-paragraphs.xsl` sheet performs this file analysis. It makes a representation of every `<p>` in the document with relevant formatting properties:
* `font-size`
* `font-style`
* `font-weight`
* `text-decoration`
* `color`
* `text-align`
Next, it sorts paragraphs into groups that share identical formatting, one group for each distinct combination of properties. These groups are candidates for promotion from `<p>` to `<h1-6>`. HTMLevator considers:
* How many paragraphs are formatted the same way
* The average paragraph length in each format group
* How often paragraphs in one format group appear in continuous runs (and thus probably aren't headings)
* Whether paragraphs are all caps
Decisions about what to consider headings are made as follows:
* Anything that is right-aligned is not considered for heading promotion
* The most common type of paragraph in the document (i.e. the combination of paragraph properties that occurs the most) is not considered for heading promotion
* Promote a paragraph group to headings if:
* The average run of consecutive paragraphs styled the same way is 4 or fewer (long runs of `<p>`s with the same styling suggest the paragraphs aren't headings), AND
* The font size specified is not the smallest font size found in the document, AND
* The average length of paragraphs with the given set of properties is not more than 120 characters
* Promote a paragraph group if it is:
* Centered, AND
* Less than 200 characters in average length, AND
* The average consecutive paragraph run is less than 2
* Promote a paragraph group if it never ends in a period
After HTMLevator has identified paragraph groups to mark as headings, it must guess the outline level. It does so based on the following attributes, in these order:
1. Font size (bigger = higher heading level)
2. Italics
3. Bold
4. Underline
5. Always caps
Generally speaking, HTMLevator's heading detection does a better job detecting headings than it does at guessing the heading's level.
#### XSLT sequence
This is the default heading promotion method, run if outline level data is not present. You can also run this `method=ranked-format`
1. First, `header-promote/digest-paragraphs.xsl` makes the paragraph groupings, and guesses what formats should be headings (and what level those headings should be).
2. The `header-promote/make-header-escalator-xslt.xsl` sheet uses the `digest-paragraphs.xsl` output as its input, which it uses to produce a bespoke `XSL` sheet.
3. Running this sheet on the original HTML file implements the heading promotion, replacing the `<p>`s thought to be headings with `<h1-6>`.
### <a name="outline-based"></a>Outline-level heading promotion
An outline level can be specified on a paragraph in Word (which often comes from a named Word style. Some writers use this outlining functionality in Word, either deliberately, or implicitly through careful use of named styles. In these instances, outline levels are often a reliable indicator of headings and heading levels.
When outline levels are specified in Word's XML (e.g. `<w:outlineLvl w:val="0"/>`), they are extracted by XSweet as an `-xsweet-outline-level` property on the `<p>`.
When this property is present at least twice in the HTML document, the `header-promote/header-promotion-CHOOSE.xsl` sheet will elect to use outline levels to promote headings.
### <a name="custom-hp"></a>Custom configuration
To create a custom configuration:
1. Create a custom mapping (`my-styles.xml` or what have you). See the example provided in `config-mockup.xml` for syntax.
2. Run the `header-promotion-CHOOSE.xsl` sheet, passing the custom mapping `.xml` sheet as a runtime parameter (`method=my-styles.xml`)
3. The `make-header-mapper-xslt.xsl` will generate and apply custom XSL sheet based on your XML file
## <a name="url"></a>Plain text URL linking
This sheet searches for plain-text URLs and automatically links them. It can recognize links with the following TLDs:
* .com
* .org
* .net
* .gov
* .mil
* .edu
* .io
* .foundation
* country TLDs
XSweet looks for a top level domain preceded by preceded by one or more strings that contain only letters, numbers, underscores and dashes (no spaces or other punctuation). These strings can be separated by periods (".") Note that this rule will capture a `www.` if it is present.
XSweet will recognizes and include in the link the protocol, if it has been specified (`http://`, `https://`, `ftp:`). If the protocol has not been specified, the link's `href` will be appended with `http://`.
This sheet will also capture query strings on links.
## <a name="list-detection"></a>Plain text list tagging
This module will recognize plain text that looks like a numbered lists and mark the corresponding list (as an `<ol>`) and list items (`<li>`s).
`DETECT-ITEMIZE-LISTS.xsl` runs from within it 3 separate sheets in sequence:
1. `detect-numbered-lists.xsl`, which detects lists and bookends them with `<xsw:list xmlns:xsw="" level="0">`
2. `itemize-detected-lists.xsl`, which converts the `<xsw:list>` tags to `<ol>`, and wraps each paragraph in `<li>`s
3. `scrub-literal-numbering-lists.xsl`, which removes from each list item the leading whitespace, literal text numbering, the period, and the whitespace after it
Lists must match the following pattern to be detected and marked as a numbered list:
* Each list item paragraph may start with any amount of white space (including none), followed by
* a string of one or more numerals, followed by
* a period, followed by
* one or more white space characters.
* Further, at least two or more consecutive paragraphs must meet these criteria to be marked as a list
List items that meet this criteria are scrubbed of their literal numbering (and following white space) in favor of automatically generated `<ol>` numbering.
Note that this feature creates a flat list (one level), rather than nested lists based on indentation.
This module can be run before or after the `PROMOTE-lists.xsl` feature in XSweet Core. To use it, you can modify the `` file of the XSweet_runner_scripts to include this step before the `final-rinse.xsl` step.
See also the documentation for <a href="">marked list handling</a>.
## <a name="ucp"></a>Copyediting cleanups
### <a name="text-cleanup"></a>`ucp-cleanup/ucp-text-macros.xsl`
This sheet contains a suite of text cleanups, built specifically for use by the <a href="">University of California Press</a>. It automates many copyediting improvements:
* Hyphens between numerals are converted to en dashes
* Two or more consecutive spaces are converted to a single space
* Any number of spaces before or after em dashes are removed
* Series of periods are converted to ellipses
* Two adjacent hyphens become an em dash
* En dashes surrounded on both sides by spaces are converted to an em dash
* Equal signs are normalized to be surrounded by one space on either side
* Spaces adjacent to tabs are removed
* Spaces at the beginning and end of paragraphs are removed
* Tabs at the end of paragraphs are removed
* Empty paragraphs are removed
* Single and double quotation marks (including backticks) are converted to directional quotation marks
* Hair spaces are inserted between single and double quotation marks
* Punctuation marks are coerced to match the formatting of the previous word; e.g. `<i>extraordinary</i>!` becomes `<i>extraordinary!</i>`. This rule applies to the following punctuation marks:
* "
* '
* :
* ;
* ?
* !
### <a name="ucp-mappings"></a>`ucp-cleanup/ucp-mappings.xsl`
In this step, underlining and bolding is converted to italics, either as inline tags or `style` CSS:
* `<b>`s and `<u>`s are replaced with `<i>`s
* `style="font-weight: bold"` and `style="text-decoration: underline"` become `style="font-style: italic"`
Short and sweet.
## <a name="custom-trans"></a>Custom transformations (experimental)
The files in the `html-tweak` folder can be used to extend XSweet, by defining custom transformations to apply to the text. This can be done on a per-document basis, or to implement generic rules according to your use case.
Use is as follows:
1. Define the custom transformations to be applied in an `.xml` file
2. Run the `APPLY-html-tweaks.xsl` sheet, referencing the above transformations defined in your `xml` file. This:
1. reads the user-defined transformations from your `.xml` file
2. creates a new XSL sheet based on the `.xml` file that will implement the specified transformation (done with the `make-html-tweak-xslt.xsl` sheet)
3. applies the created XSL sheet to the input file
Example use (exact script will depend upon how you are running your XSLT:
`XSLT my-source.html APPLY-html-tweaks.xsl config=my-html-tweaks.xml`
### Tweak definition syntax
The user-specified tweaks work by establishing matches between categories of HTML elements (most commonly but certainly not limited to `<p>`s or `<span>`s), as indicated by:
* CSS property or CSS property-value (on a `style` attribute), or
* Named classes (the `class` attribute)
The syntax to define HTML tweaks uses the following components:
* `where`: a wrapper for a rule
* `match`: conditions on an element for it to match
* `style`: a `style` property name or `property-name: value` combination
* `class`: a class value (name token)
### Example 1
Remove `Default` classes from HTML elements where they appear:
<p class="Default">Here is default class paragraph</p>
<p>Here is default class paragraph</p>
HTML tweak rule:
### Example 2
Remove a specific styling property wherever it's present:
<p style="text-indent:1em; margin-bottom: 1em">Styling includes a property</p>
<p style="text-indent:1em">Styling includes a property</p>
HTML tweak rule:
### Example 3
Remove a `style` property if it has a given value:
<p style="font-family: Helvetica; font-size: 12pt">Remove a property if it has a specific value</p>
<p style="font-size: 12pt">Remove a property if it has a specific value</p>
HTML tweak rule:
<match><style>font-family: Helvetica</style></match>
### Example 4
The following tweak rule will map a specific `class` and `style` to another `class` and `style`:
<style>font-size: 18pt</style>
<style>color: red</style>
For further examples, see the demo files included in the repository:
* `html-tweak-map.xml` defines example transformation definitions
* `html-tweak-demo.xsl` is the resulting XSL sheet made by the `make-html-tweak-xslt.xsl`, which will effect the specified transformation. (This relies on the `html-tweak-lib.xsl` file as a dependency)
## <a name="sections"></a>Section inferrer (experimental)
This utility uses headings (`<h1-6>`) as markers and attempts to add `<section>`s to an HTML file. It is run as a single XSL sheet, `induce-sections/induce-sections.xsl`, which returns the document HTML file unchanged except for the addition of `<section>` tags.
* Sections are only added when higher-level headings wrap lower-level ones. Lower-level headings wrapping higher-level ones are not captured as `<section>`s
* Paragraphs and blocks preceding the first header, appear without a section wrapper
(before the first section)
* Paragraphs and all other elements travel with the immediately preceding header
* Files with no headings are unchanged
* Your document must be wrapped in a `<div class="docx-body">` for this sheet to work. It will be wrapped if it has been extracted by XSweet; otherwise you will have to add this element yourself
* If headings skip levels, a note will be added: `<!-- Headers out of regular order: h1, h2, h3, h1, h3-->`
<html xmlns="">
<meta charset="utf-8"></meta>
<div class="docx-body">
<p>h1 para</p>
## h2
<p>h2 para</p>
### h3
<p>h3 para</p>
<p>h3 para</p>
<p>h1 para</p>
### h3
<p>h3 para</p>
<html xmlns="">
<meta charset="utf-8" />
<div class="docx-body">
<!-- Headers out of regular order: h1, h2, h3, h1, h3-->
<p>h1 paragraph</p>
## h2
<p>h2 para</p>
### h3
<p>h3 para</p>
<p>h3 para</p>
<p>h1 paragraph</p>
### h3
<p>h3 para</p>
`mark-sections.xsl` and `nest-sections.xsl` are deprecated; the `induce-sections.xsl` sheet encapsulates the functionality from both.
# XSweet Core
The purpose of XSweet Core is to extract the underlying XML from Word documents (.docx) into valid, clean, and reusable HTML. XSweet Core extracts and saves “important” information from the .docx, such as:
* Font type and size
* Formatting, including bold, italic, underlining, text justification (left, right, center), subscript/superscript, etc.
* Paragraph indentation
* Endnotes and footnotes
* Outline and list levels
After initially extracting selected information, XSweet Core removes the (voluminous) extra noise in the XML, combines similar tags, abstracts away inline or repetitive formatting where possible, and rewrites word-specific feature implementations for HTML (e.g. notes, lists).
From an inscrutable mess comes good, clean, and human readable HTML.
## Contents
* [Pipeline](#pipeline")
* [Initial extraction](#extraction")
* [File structure](#structure")
* [Formatting](#formatting")
* [Word styles](#styles")
* [Links](#links")
* [Tables](#tables")
* [Notes](#notes")
* [Images](#images")
* [Further cleanups](#cleanups")
* [scrub.xsl](#scrub")
* [join-elements.xsl](#join")
* [collapse-paragraphs.xsl](#collapse")
* [Lists](#lists")
* [Math](#math")
* [Final Rinse](#rinse")
* [Additional XSL sheets](#additional")
* [css-abstract.xsl](#css")
* [html-analysis.xsl](#analysis")
* [Serialization](#serialize")
* [html5-serialize.xsl](#html5")
* [xhtml-serialize.xsl](#xhtml")
* [plaintext.xsl](#plaintext")
## <a name="pipeline"></a>Pipeline
Initial extraction is achieved by running the `docx-extract/EXTRACT-docx.xsl` sheet, which in turn runs the following sequence of steps:
1. `docx-extract/docx-html-extract.xsl`
* which also runs `docx-extract/docx-table-extract.xsl`
2. `docx-extract/handle-notes.xsl`
3. `docx-extract/scrub.xsl`
4. `docx-extract/join-elements.xsl`
5. `docx-extract/collapse-paragraphs.xsl`
Running the `EXTRACT-docx.xsl` sheet has exactly the same effect as running these five sheets in sequence, using each step's output as the input for the next.
Next, `list-promote/PROMOTE-lists.xsl` creates HTML representations of lists.
The math sheet, `xsweet_tei_omml2mml.xsl` captures math and equations added in Word.
Finally, run `html-polish/final-rinse.xsl` for additional cleanup to the HTML file.
Note that the outputs of all of these steps technically produce XHTML, not true HTML. This is expedient for chaining transformation steps into a pipeline. You may want to serialize to HTML5 as the final step (see <a href="#html5">html5-serialize.xsl</a>), although it is not recommended to do so until all other desired transformations are finished.
## <a name="extraction"></a>Initial extraction
### `docx-html-extract.xsl`
This sheet does the heaving lifting of extracting information from the MS Word XML. Contents are extracted from the body of the `.docx`. Information from headers and footers is _not_ currently extracted.
#### <a name="structure"></a>File structure
The contents of the HTML `body` section are wrapped in the following section containers:
* `<div class="docx-body">`: the main content of the file
* `<div class="docx-endnotes">`: wraps all endnotes
* `<div class="docx-footnotes">`: wraps all footnotes
#### <a name="formatting"></a>Formatting
There is a surprising amount of redundant and conflicting information in Word's XML, so it's important to choose the right level from which to extract properties. Many properties can be specified either on the paragraph level or the text run level.
XSweet extracts the following properties from the paragraph-level (from inside `<w:pPr>` paragraph properties tags):
* Named Word styles
* MS Word outline level
* MS Word list level
Other formatting information specified on the paragraph level is ignored, as it is unreliable and often overridden by properties specified on the text run level (`<w:rPr>`). The following are extracted from the run level:
* Font family
* Font size
* Text alignment
* Font color
* Italicization, bolding, underline, sub- and superscript
Unless formatting in Word is achieved by applied styles (see below), formatting is extracted inline or as element-level CSS (`<p style="[formatting]">`).
Inline elements preserved in HTML:
* Bold `<b>`
* Italic `<i>`
* Underline `<u>`
* Subscript `<sub>`
* Superscript `<sup>`
* Line breaks `<br>`
CSS properties captured include:
* font `font-family`
* font size `font-size`
* `font-weight`:`bold``normal`
* `font-style`: `italic`, `normal`
* `text-decoration: underline`, `none`
* `font-variant`: `normal` or `small-caps`
* text alignment `text-align`: `left`, `right`, `center`
* indentation `text-indent`, `padding-left`, `padding-right`
* margins `margin-top`, `margin-bottom`, `margin-left`, `margin-right`
* color `color`
The following pseudo-properties are also extracted:
* list level `-xsweet-list-level`: used for list extraction
* outline level `-xsweet-outline-level`: used to aid in heading detection
Be aware that XSweet extracts only what information is present in the Word XML. Some information, such as font type and size, etc., is not always explicitly defined in the XML. Word applies formatting from the default `Normal` style when no other style is specified. This can lead to some minor formatting inconsistencies.
##### <a name="styles"></a>Word styles
Word styles are referenced by name inside the Word file's main `document.xml` file, and defined in the `styles.xml` file in the same directory, which contains style formatting and property information. XSweet's initial extraction captures applied Word styles as CSS `class` attributes, and extracts the relevant information from `styles.xml` into a CSS `<style>` tag in the HTML `<head>`:
<html xmlns="">
<meta charset="UTF-8" />
<style type="text/css">
.Title {
/* Normal*/
font-family: Arial Unicode MS;
/* Title*/
margin-bottom: 0pt;
text-align: center;
font-family: Times New Roman;
font-weight: bold;
font-variant: small-caps
<div class="docx-body">
<p class="Title" style="font-family: Times New Roman; font-variant: small-caps; font-weight: bold; margin-bottom: 0pt; text-align: center">Chapter Two</p>
Note that formatting information is often repeated: both inline as `style` and from `class` - there is analytical value to having the CSS inline (e.g. for <a href="">heading promotion</a>).
#### <a name="links"></a>Links
Where they exist, hyperlinks from the Word document are preserved and passed through to the HTML as `<a href>`s. Link text formatting is preserved.
<a href="">
<span style="color: #0000FF; text-decoration: underline">
<span class="Hyperlink"></span>
Internal-to-Word bookmarks are also extracted, although without additional processing (not provided by XSweet), they are merely placeholders and are removed during <a href="">Editoria Typescript</a>.
<a class="bookmarkStart" id="docx-bookmark_0">
<!-- bookmark ='_GoBack'-->
<a href="#docx-bookmark_0">
<!-- bookmark end -->