Merge pull request #161 from barrysteyn/master

latex plugin revamp
This commit is contained in:
Justin Mayer
2014-02-18 09:26:27 -08:00
2 changed files with 346 additions and 74 deletions

View File

@@ -3,73 +3,73 @@ Latex Plugin For Pelican
This plugin allows you to write mathematical equations in your articles using Latex.
It uses the MathJax Latex JavaScript library to render latex that is embedded in
between `$..$` for inline math and `$$..$$` for displayed math. It also allows for
between `$..$` for inline math and `$$..$$` for displayed math. It also allows for
writing equations in by using `\begin{equation}`...`\end{equation}`.
Installation
------------
To enable, ensure that `latex.py` is put somewhere that is accessible.
Then use as follows by adding the following to your settings.py:
PLUGINS = ["latex"]
Be careful: Not loading the plugin is easy to do, and difficult to detect. To
make life easier, find where pelican is installed, and then copy the plugin
there. An easy way to find where pelican is installed is to verbose list the
available themes by typing `pelican-themes -l -v`.
Your site is now capable of rendering latex math using the mathjax JavaScript
library. No alterations to the template file is needed.
Once the pelican folder is found, copy `latex.py` to the `plugins` folder. Then
add to settings.py like this:
PLUGINS = ["pelican.plugins.latex"]
Now all that is left to do is to embed the following to your template file
between the `<head>` parameters (for the NotMyIdea template, this file is base.html)
{% if article and article.latex %}
{{ article.latex }}
{% endif %}
{% if page and page.latex %}
{{ page.latex }}
{% endif %}
### Typogrify
Typogrify will now play nicely with Latex (i.e. typogrify can be enabled
and Latex will be rendered correctly). In order for this to happen,
version 2.07 (or above) of typogrify is required. In fact, this plugin expects
that at least version 2.07 is present and will fail without it.
Usage
-----
Latex will be embedded in every article. If however you want latex only for
selected articles, then in settings.py, add
### Backward Compatibility
This plugin is backward compatible in the sense that it
accompishes what previous versions did without needing any setup in the
metadata or settings files.
LATEX = 'article'
### Settings File
Extra options regarding how mathjax renders latex can be set in the settings
file. These options are in a dictionary variable called `LATEX` in the pelican
settings file.
And in each article, add the metadata key `latex:`. For example, with the above
settings, creating an article that I want to render latex math, I would just
include 'Latex' as part of the metadata without any value:
The dictionary can be set with the following keys:
Date: 1 sep 2012
Status: draft
Latex:
* `wrap`: controls the tags that math is wrapped with inside the resulting
html. For example, setting `wrap` to `'mathjax'` would wrap all math inside
`<mathjax>...</mathjax>` tags. If typogrify is set to True, then math needs
to be wrapped in tags and `wrap` will therefore default to `mathjax` if not
set.
* `align`: controls how displayed math will be aligned. Can be set to either
`left`, `right` or `center` (default is `center`).
* `indent`: if `align` not set to `center`, then this controls the indent
level (default is `0em`).
* `show_menu`: controls whether the mathjax contextual menu is shown.
* `process_escapes`: controls whether mathjax processes escape sequences.
* `preview`: controls the preview message users are seen while mathjax is
loading.
* `color`: controls the color of the mathjax rendered font.
For example, in settings.py, the following would make latex render in blue and
displaymath align to the left:
LATEX = {'color':'blue','align':left}
Latex Examples
--------------
###Inline
Latex between `$`..`$`, for example, `$`x^2`$`, will be rendered inline
Latex between `$`..`$`, for example, `$`x^2`$`, will be rendered inline
with respect to the current html block.
###Displayed Math
Latex between `$$`..`$$`, for example, `$$`x^2`$$`, will be rendered centered in a
Latex between `$$`..`$$`, for example, `$$`x^2`$$`, will be rendered centered in a
new paragraph.
###Equations
Latex between `\begin` and `\end`, for example, `begin{equation}` x^2 `\end{equation}`,
will be rendered centered in a new paragraph with a right justified equation number
at the top of the paragraph. This equation number can be referenced in the document.
To do this, use a `label` inside of the equation format and then refer to that label
using `ref`. For example: `begin{equation}` `\label{eq}` X^2 `\end{equation}`. Now
Latex between `\begin` and `\end`, for example, `begin{equation}` x^2 `\end{equation}`,
will be rendered centered in a new paragraph with a right justified equation number
at the top of the paragraph. This equation number can be referenced in the document.
To do this, use a `label` inside of the equation format and then refer to that label
using `ref`. For example: `begin{equation}` `\label{eq}` X^2 `\end{equation}`. Now
refer to that equation number by `$`\ref{eq}`$`.
Template And Article Examples
-----------------------------
To see an example of this plugin in action, look at
[this article](http://doctrina.org/How-RSA-Works-With-Examples.html). To see how
this plugin works with a template, look at
[this template](https://github.com/barrysteyn/pelican_theme-personal_blog).

View File

@@ -5,51 +5,323 @@ Latex Plugin For Pelican
This plugin allows you to write mathematical equations in your articles using Latex.
It uses the MathJax Latex JavaScript library to render latex that is embedded in
between `$..$` for inline math and `$$..$$` for displayed math. It also allows for
writing equations in by using `\begin{equation}`...`\end{equation}`.
between `$..$` for inline math and `$$..$$` for displayed math. It also allows for
writing equations in by using `\begin{equation}`...`\end{equation}`. No
alteration to a template is required for this plugin to work, just install and
use.
Typogrify Compatibility
-----------------------
This plugin now plays nicely with typogrify, but it requires
typogrify version 2.07 or above.
User Settings
-------------
Users are also able to pass a dictionary of settings in the settings file which
will control how the mathjax library renders thing. This could be very useful
for template builders that want to adjust look and feel of the math.
See README for more details.
"""
from pelican import signals
from pelican import contents
import re
# Reference about dynamic loading of MathJax can be found at http://docs.mathjax.org/en/latest/dynamic.html
# The https cdn address can be found at http://www.mathjax.org/resources/faqs/#problem-https
latexScript = """
<script type= "text/javascript">
# Global Variables
_WRAP_TAG = None # the tag to wrap mathjax in (needed to play nicely with typogrify or for template designers)
_LATEX_REGEX = re.compile(r'(\$\$|\$|\\begin\{(.+?)\}).*?\1|\\end\{\2\}', re.DOTALL | re.IGNORECASE) # used to detect latex
_LATEX_SUMMARY_REGEX = None # used to match latex in summary
_LATEX_PARTIAL_REGEX = None # used to match latex that has been cut off in summary
_MATHJAX_SETTINGS = {} # Settings that can be specified by the user, used to control mathjax script settings
_MATHJAX_SCRIPT="""
<script type= "text/javascript">
if (!document.getElementById('mathjaxscript_pelican')) {{
var s = document.createElement('script');
s.type = 'text/javascript';
s.src = 'https:' == document.location.protocol ? 'https://c328740.ssl.cf1.rackcdn.com/mathjax/latest/MathJax.js' : 'http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML';
s.id = 'mathjaxscript_pelican';
s.type = 'text/javascript'; s.src = 'https:' == document.location.protocol ? 'https://c328740.ssl.cf1.rackcdn.com/mathjax/latest/MathJax.js' : 'http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML';
s[(window.opera ? "innerHTML" : "text")] =
"MathJax.Hub.Config({" +
" config: ['MMLorHTML.js']," +
" jax: ['input/TeX','input/MathML','output/HTML-CSS','output/NativeMML']," +
" TeX: { extensions: ['AMSmath.js','AMSsymbols.js','noErrors.js','noUndefined.js'], equationNumbers: { autoNumber: 'AMS' } }," +
"MathJax.Hub.Config({{" +
" config: ['MMLorHTML.js']," +
" TeX: {{ extensions: ['AMSmath.js','AMSsymbols.js','noErrors.js','noUndefined.js'], equationNumbers: {{ autoNumber: 'AMS' }} }}," +
" jax: ['input/TeX','input/MathML','output/HTML-CSS']," +
" extensions: ['tex2jax.js','mml2jax.js','MathMenu.js','MathZoom.js']," +
" tex2jax: { " +
" displayAlign: '{align}'," +
" displayIndent: '{indent}'," +
" showMathMenu: {show_menu}," +
" tex2jax: {{ " +
" inlineMath: [ [\'$\',\'$\'] ], " +
" displayMath: [ [\'$$\',\'$$\'] ]," +
" processEscapes: true }, " +
" 'HTML-CSS': { " +
" styles: { '.MathJax .mo, .MathJax .mi': {color: 'black ! important'}} " +
" } " +
"}); ";
" processEscapes: {process_escapes}," +
" preview: '{preview}'," +
" }}, " +
" 'HTML-CSS': {{ " +
" styles: {{ '.MathJax_Display, .MathJax .mo, .MathJax .mi, .MathJax .mn': {{color: '{color} ! important'}} }}" +
" }} " +
"}}); ";
(document.body || document.getElementsByTagName('head')[0]).appendChild(s);
</script>
}}
</script>
"""
def addLatex(gen, metadata):
# Python standard library for binary search, namely bisect is cool but I need
# specific business logic to evaluate my search predicate, so I am using my
# own version
def binary_search(match_tuple, ignore_within):
"""Determines if t is within tupleList. Using the fact that tupleList is
ordered, binary search can be performed which is O(logn)
"""
The registered handler for the latex plugin. It will add
the latex script to the article metadata
ignore = False
if ignore_within == []:
return False
lo = 0
hi = len(ignore_within)-1
# Find first value in array where predicate is False
# predicate function: tupleList[mid][0] < t[index]
while lo < hi:
mid = lo + (hi-lo+1)/2
if ignore_within[mid][0] < match_tuple[0]:
lo = mid
else:
hi = mid-1
if lo >= 0 and lo <= len(ignore_within)-1:
ignore = (ignore_within[lo][0] <= match_tuple[0] and ignore_within[lo][1] >= match_tuple[1])
return ignore
def ignore_content(content):
"""Creates a list of match span tuples for which content should be ignored
e.g. <pre> and <code> tags
"""
if 'LATEX' in gen.settings.keys() and gen.settings['LATEX'] == 'article':
if 'latex' in metadata.keys():
metadata['latex'] = latexScript
ignore_within = []
# used to detect all <pre> and <code> tags. NOTE: Alter this regex should
# additional tags need to be ignored
ignore_regex = re.compile(r'<(pre|code).*?>.*?</(\1)>', re.DOTALL | re.IGNORECASE)
for match in ignore_regex.finditer(content):
ignore_within.append(match.span())
return ignore_within
def wrap_latex(content, ignore_within):
"""Wraps latex in user specified tags.
This is needed for typogrify to play nicely with latex but it can also be
styled by template providers
"""
wrap_latex.foundlatex = False
def math_tag_wrap(match):
"""function for use in re.sub"""
# determine if the tags are within <pre> and <code> blocks
ignore = binary_search(match.span(1), ignore_within) and binary_search(match.span(2), ignore_within)
if ignore:
return match.group(0)
else:
wrap_latex.foundlatex = True
return '<%s>%s</%s>' % (_WRAP_TAG, match.group(0), _WRAP_TAG)
return (_LATEX_REGEX.sub(math_tag_wrap, content), wrap_latex.foundlatex)
def process_summary(instance, ignore_within):
"""Summaries need special care. If Latex is cut off, it must be restored.
In addition, the mathjax script must be included if necessary thereby
making it independent to the template
"""
process_summary.altered_summary = False
insert_mathjax_script = False
endtag = '</%s>' % _WRAP_TAG if _WRAP_TAG != None else ''
# use content's _get_summary method to obtain summary
summary = instance._get_summary()
# Determine if there is any math in the summary which are not within the
# ignore_within tags
mathitem = None
for mathitem in _LATEX_SUMMARY_REGEX.finditer(summary):
if binary_search(mathitem.span(), ignore_within):
mathitem = None # In <code> or <pre> tags, so ignore
else:
insert_mathjax_script = True
# Repair the latex if it was cut off mathitem will be the final latex
# code matched that is not within <pre> or <code> tags
if mathitem and mathitem.group(4) == ' ...':
end = r'\end{%s}' % mathitem.group(3) if mathitem.group(3) is not None else mathitem.group(2)
latex_match = re.search('%s.*?%s' % (re.escape(mathitem.group(1)), re.escape(end)), instance._content, re.DOTALL | re.IGNORECASE)
new_summary = summary.replace(mathitem.group(0), latex_match.group(0)+'%s ...' % endtag)
if new_summary != summary:
return new_summary+_MATHJAX_SCRIPT.format(**_MATHJAX_SETTINGS)
def partial_regex(match):
"""function for use in re.sub"""
if binary_search(match.span(), ignore_within):
return match.group(0)
process_summary.altered_summary = True
return match.group(1) + match.group(4)
# check for partial latex tags at end. These must be removed
summary = _LATEX_PARTIAL_REGEX.sub(partial_regex, summary)
if process_summary.altered_summary:
return summary+_MATHJAX_SCRIPT.format(**_MATHJAX_SETTINGS) if insert_mathjax_script else summary
return summary+_MATHJAX_SCRIPT.format(**_MATHJAX_SETTINGS) if insert_mathjax_script else None
def process_settings(settings):
"""Sets user specified MathJax settings (see README for more details)"""
global _MATHJAX_SETTINGS
# NOTE TO FUTURE DEVELOPERS: Look at the README and what is happening in
# this function if any additional changes to the mathjax settings need to
# be incorporated. Also, please inline comment what the variables
# will be used for
# Default settings
_MATHJAX_SETTINGS['align'] = 'center' # controls alignment of of displayed equations (values can be: left, right, center)
_MATHJAX_SETTINGS['indent'] = '0em' # if above is not set to 'center', then this setting acts as an indent
_MATHJAX_SETTINGS['show_menu'] = 'true' # controls whether to attach mathjax contextual menu
_MATHJAX_SETTINGS['process_escapes'] = 'true' # controls whether escapes are processed
_MATHJAX_SETTINGS['preview'] = 'TeX' # controls what user sees as preview
_MATHJAX_SETTINGS['color'] = 'black' # controls color math is rendered in
if not isinstance(settings, dict):
return
# The following mathjax settings can be set via the settings dictionary
# Iterate over dictionary in a way that is compatible with both version 2
# and 3 of python
for key, value in ((key, settings[key]) for key in settings):
if key == 'align' and isinstance(value, str):
if value == 'left' or value == 'right' or value == 'center':
_MATHJAX_SETTINGS[key] = value
else:
_MATHJAX_SETTINGS[key] = 'center'
if key == 'indent':
_MATHJAX_SETTINGS[key] = value
if key == 'show_menu' and isinstance(value, bool):
_MATHJAX_SETTINGS[key] = 'true' if value else 'false'
if key == 'process_escapes' and isinstance(value, bool):
_MATHJAX_SETTINGS[key] = 'true' if value else 'false'
if key == 'preview' and isinstance(value, str):
_MATHJAX_SETTINGS[key] = value
if key == 'color' and isinstance(value, str):
_MATHJAX_SETTINGS[key] = value
def process_content(instance):
"""Processes content, with logic to ensure that typogrify does not clash
with latex.
In addition, mathjax script is inserted at the end of the content thereby
making it independent of the template
"""
if not instance._content:
return
ignore_within = ignore_content(instance._content)
if _WRAP_TAG:
instance._content, latex = wrap_latex(instance._content, ignore_within)
else:
metadata['latex'] = latexScript
latex = True if _LATEX_REGEX.search(instance._content) else False
# The user initially set typogrify to be True, but since it would clash
# with latex, we set it to False. This means that the default reader will
# not call typogrify, so it is called here, where we are able to control
# logic for it ignore latex if necessary
if process_content.typogrify:
# Tell typogrify to ignore the tags that latex has been wrapped in
ignore_tags = [_WRAP_TAG] if _WRAP_TAG else None
# Exact copy of the logic as found in the default reader
from typogrify.filters import typogrify
instance._content = typogrify(instance._content, ignore_tags)
instance.metadata['title'] = typogrify(instance.metadata['title'], ignore_tags)
if latex:
# Mathjax script added to the end of article. Now it does not need to
# be explicitly added to the template
instance._content += _MATHJAX_SCRIPT.format(**_MATHJAX_SETTINGS)
# The summary needs special care because latex math cannot just be cut
# off
summary = process_summary(instance, ignore_within)
if summary != None:
instance._summary = summary
def pelican_init(pelicanobj):
"""Intialializes certain global variables and sets typogogrify setting to
False should it be set to True.
"""
global _WRAP_TAG
global _LATEX_SUMMARY_REGEX
global _LATEX_PARTIAL_REGEX
try:
settings = pelicanobj.settings['LATEX']
except:
settings = None
process_settings(settings)
# Allows mathjax script to be accessed from template should it be needed
pelicanobj.settings['MATHJAXSCRIPT'] = _MATHJAX_SCRIPT.format(**_MATHJAX_SETTINGS)
# If typogrify set to True, then we need to handle it manually so it does
# not conflict with Latex
try:
if pelicanobj.settings['TYPOGRIFY'] == True:
pelicanobj.settings['TYPOGRIFY'] = False
_WRAP_TAG = 'mathjax' # default to wrap mathjax content inside of
process_content.typogrify = True
except KeyError:
pass
# Set _WRAP_TAG to the settings tag if defined. The idea behind this is
# to give template designers control over how math would be rendered
try:
if pelicanobj.settings['LATEX']['wrap']:
_WRAP_TAG = pelicanobj.settings['LATEX']['wrap']
except (KeyError, TypeError):
pass
# regular expressions that depend on _WRAP_TAG are set here
tagstart = r'<%s>' % _WRAP_TAG if not _WRAP_TAG is None else ''
tagend = r'</%s>' % _WRAP_TAG if not _WRAP_TAG is None else ''
latex_summary_regex = r'((\$\$|\$|\\begin\{(.+?)\}).+?)(\2|\\end\{\3\}|\s?\.\.\.)(%s)?' % tagend
latex_partial_regex = r'(.*)(%s)(\\.*?|\$.*?)(\s?\.\.\.)(%s)' % (tagstart, tagend)
_LATEX_SUMMARY_REGEX = re.compile(latex_summary_regex, re.DOTALL | re.IGNORECASE)
_LATEX_PARTIAL_REGEX = re.compile(latex_partial_regex, re.DOTALL | re.IGNORECASE)
def register():
"""
Plugin registration
"""
signals.article_generator_context.connect(addLatex)
signals.page_generator_context.connect(addLatex)
"""Plugin registration"""
signals.initialized.connect(pelican_init)
signals.content_object_init.connect(process_content)