BB Code Enhancements - vLaTeX - Math parsing with LaTeX

vLaTeX allows your users to include mathematical equations written in LaTeX into their posts, rendered in high quality graphic files (either png or gif). Such ability is instrumental in scientific or educational oriented forums. When I started in vb two years ago, there ware a number of solutions available, but none of them suited my needs completely, so I developed my own integration between (a modified version of) Benjamin Zeiss' LatexRender and vBulletin. With the new vBulletin 4.0, I decided to improve some of the features of this product, and to release it with the hope that it turns out to be useful to any of you.


What is LaTeX?

LaTeX is a set of macros for the TeX typesetting system. It is the defacto standard of communication within the scientific and mathematical community. It works pretty much like a programing language: you write a text source file, and the latex executable compiles it in a DeVice Independent file (DVI), which can be later converted to the desired format (postscript, pdf, etc.).

This mod (thanks to a modified version of the LatexRender class) does all this work in the background. It takes the formula code, inserts it in a source file, compiles it and converts the output in a high quality image file (png or gif, depends on your settings) to show in the post.


Requirements

This mod relies on your server having available the following external programs. They are pretty standard in most Linux distributions, so hosts can install them easily (if they want to, some don't). Please, check your server meets this requirements (or ask your host to install the programs) before trying to use vLaTeX:

- latex: the LaTeX compiler. Check the TeX live distribution (http://www.tug.org/texlive/).

- dvips: converts the DVI file to PS. Usually included in all latex distributions.

- convert: as it name suggests, converts between a large number of graphic files. Belongs to the ImageMagick package. Needs ghostscript.

- identify: among other things, measures the size of image files. Also belongs to ImageMagick

- Ghostscript: PostScript and PDF language interpreter and previewer. Is called by convert in order to handle ps files.


Features

Creates a new BBcode, usually [TEX] (thought the tag name is configurable via the admin cp), which contents are processed thought LaTeX to generate image files. There are four calling modes:

- [TEX]E= mc^2[/TEX] in-line mode. The image is included in the middle of the text, without line breaks. It tries to adjust the baseline of the formula by adjusting the vertical-align CSS property; the results aren't always perfect, but it does its best.

- [TEX=null]E = m c^2[/TEX] displayed mode: The formula is displayed (usually centered) between two paragraphs.

- [TEX=*]E = m c^2[/TEX] auto-numbering mode: The formula is displayed (usually centered), and numbered.
NOTE: The auto-numbering depends on the CSS counter system, so some old browsers (most notably ie 7 and older) will show these equations untagged, like [TEX=null]. Don't blame me, blame non-standard compilant browsers.

- [TEX=tag]E = m c^2[/TEX] displayed with tag: The formula is displayed (usually centered) between two paragraphs, and tagged with the specific tag.

In the last two modes, the tags are displayed at the right (left if you use RTL) between brackets, in a (configurable) color.

The formulas are compiled by creating a small tex file and running the native LaTeX program throughout it. You can choose to wrap the formula with the standard dollar ($ ... $), the align or gather amsTeX environments. The standard dollar is recommended since the baseline hack works better with it, but the gather environment allows to create multiline equations using the \\ separator.

This mod also introduces a second BBcode, [EQREF]tag[/EQREF], which outputs (tag), with the same style of the equation tags. It's useful to let users make reference their formulas.

A double click on the image of a generated image will open a pop-up displaying the LaTeX source code used to generate the formula.

Menu with Common used LaTeX commands can be shown below the advanced mode editor. The admin can customize the commands of the menu thought the admin control panel.


Admin features

Image files are cached, so every unique formula is only processed once.

Fully configurable via Style Variables and templates. It is also fully phrased.

It doesn't change any vb default template, so upgrading should be easy.

Includes a list of blacklisted latex commands for security (configurable in the admin cp).

The admin can customize the preamble of the latex file in the admin cp. This is useful to define new commands that your community uses often, or to include packages.

Error messages can be customized in the phrase system.

Memory usage of ImageMagick commands (convert and identify) can be limited in the admin control panel.


Installation / Upgrade

1.- Unzip the mod files.

2.- Upload the vLaTeX directory to your forum root. The generated folder structure should look like this:

vLaTeX/
vLaTeX/index.htm
vLaTeX/class_vlatex.php
vLaTeX/functions_vlatex.php
vLaTeX/vlatex.js
vLaTeX/pics/
vLaTeX/pics/index.htm
vLaTeX/temp/
vLaTeX/temp/index.htm

3.- Change the permissions of the pics and temp folders so the web server can write on them.

4.- Import the product XML in your admin control panel. If you're upgrading, set "Allow Overwrite" to yes.

5.- Go to Admin CP > Settings > Options > vLaTeX - Math Parsing

6.- Make sure the path to latex, dvips, convert and identify executables is correct. Those settings default to the most common values in most linux distributions, but every host is a different world.
If you have shell access, you can check the location of the executables with the following commands:
which latex
which dvips
which convert
which identify

7.- Make sure the path to the Images and Temporal directory are correct. You can change them if you don't like the default value, but the directories must exist and the server must have permissions to write on them.
For extra safety, you could move the temp folder out of the public accessible directory three, thought it should not be needed, since files are deleted in matter of milliseconds.

8.- Set the rest of options as you wish.

9.- Test the product in some posts.

10.- If you wish to add a button in the advanced editor tool-bar for this mod, you can add it using the admin cp section Custom BB Codes (write anything you want in Replacement, this mod will overwrite it). You can use any tex icon, for instance http://www.iconfinder.com/icondetails/3780/16/tex_icon


Error messages

The LaTeX parsing can fail for a number of reasons. In such case, an error message is displayed instead of the formula. Here's a list of all possible error messages and their meaning:

- "Image too big {1}, max {2}x{3}", The resulting image is bigger than the allowed dimensions.

- "Blacklisted command", The user attempted to compile a formula with a forbidden command.

- "Can't move image", Image formula could be created, but I could not be moved to the destination folder. Check that the directory exists and has write permissions.

- "Convert failed", Formula could be compiled, but could not be converted to png/gif. Check the convert program path.

- "Compile failed", The formula could not be compiled. Check that the temporal directory has write permissions and that the latex path is correct.

- "Code too long, max. {1} characters", The formula code was too long. The maximum length can be configured in the admincp.

- "Formula not found", The compilation and convert process appeared to not raise any error, but the image was not found on the destination folder. Check for permissions in the destination folder and/or IO errors.

- "Pictures directory unwritable", The destination folder does not exist, or has no write permissions.

- "Temp. dir unwritable", The temporal folder does not exist, or has no write permissions.

- "Unknown error", Any other error condition.


Disclaimer

This mod is in beta phase, if you find bugs, please tell me how to reproduce them so I can improve it. It is distributed as is, with the hope that it is useful for someone, but without any warranty. Nor will I accept any responsibility or liability if it doesn't work as expected, or even if it breaks something.
This mod is mainly developed for personal usage. I will listen to good ideas, but I can't guaranty that I will be implemented requested features.


License

This mod is distributed under the LGPL.(http://www.gnu.org/copyleft/lesser.html). As is, you can freely distribute it or any derivative work provided that: 1) you acknowledge the work of the previous coders (like myself and Zeiss), 2) you release it under the same license (or GPL).



Acknowledgments

This mod (in particular, the class_vlatex.php file) is strongly based on the LatexRender class by Benjamin Zeiss (http://www.mayer.dial.pipex.com/tex.htm).

The baseline hack by the work of Maarten Sneep (http://mactextoolbox.sourceforge.net.../baseline.html).


Compatibility

The last version of this mod (0.5.1) requires vb4.2.0 or higher. It should work on previous CKEditor versions, but it was not tested.

If you wish to use it in a pre-CKEditor vb installation, you can use the previous vLaTeX version (0.4.1). If you want to use it in an older vb4.0.x release, you can download version 0.3 (which is compatible with vb4.0.1 or higher). Take into account that old versions of this hack have less features and might have some issues fixed in later versions.

Support will be given only for the last version of this mod.



History

2009 / 12 / 19 - First (beta) release

2009 / 12 / 20 - version 0.2
Fixed baseline problem of in-line formulas in CMS articles.
Improved error handling.
Included Spanish translation

2010 / 1 / 24 - version 0.3 (requires vb4.0.1)
Fixed extra space after in-line equations.
Won't break down if vb's CSS is stored as files.
New pop-up showing the formula source code upon double-click.

2010 / 7 / 23 - version 0.4 (requires vb4.0.4)
Improved support for CSS stored as files.
Fixed bug of js file not being loaded on certain servers.
Menu with common LaTeX commands under vb's advanced editors.
Equations will no longer be break by preview in blogs or CMS articles.

2010 / 4 / 24 - version 0.4.1 (requires vb4.0.4)
Improved compatibility of the Common Commands quick-menu with IE browser.

2012 / 12 / 07 - version 0.5.1 (requires vb4.2.0)
CKEditor support.
Memory limit for imagemagick commands.
Code pop-up now can overflow postbit boundary.

2012 / 12 / 17 - version 0.5.2 (requires vb4.2.0)
Introduced Formula Wrapper option.

[oxide11]

ok great thing this plugin thanks

[john7911]

Good job , thank you

[Ptah]

Hi. I've installed it on my forum and it works great!

I have a concern though, which becomes a suggestion/request of sorts.

My understanding is that each expression rendered results in a new PNG file in /pics. Makes sense, if so. So with that understanding...

/pics files accumulating when users iteratvely tweak TEX expressions in preview/edit?
Let's say someone is using the preview-edit-preview in a loop whilst developing a post, and tweaking the latex expression each time, hence generating however many intermediate PNG files (per latex expression tweak), only the last of which is really useful (the one they wind up posting). This results in a pile of images sitting around under /pics that are effectively wasted space. I don't think this is too unrealistic of a scenario, being that people may see a problem with their (say, nontrivial) latex in preview, so then tweak it, iteratively, until they see what the want.

Concern of "waste" /pics accumulation
Now my concern. Over time this can become a problem, lots of wasted space.

For instance, anticipating this may be a problem, I have asked the users of my forum to consider using one of the freely available online, interactive LaTeX renderers to "draft" their expressions before posting them on my forum, hoping to head off useless PNG file accumulations.

Extent suggestions/workarounds...
So. I understand that you can purge the post cache and rebuild, but that seems to be long-term unfeasible as the forum ages, and accumulates many posts, right? Just to purge unused /pic files, you have to drop and rebuild the whole post cache? Seems like hitting it with a hammer, unless I've misunderstood?

Suggestion/Question?

I'd consider making the suggestion, "Hey, can we disable the TEX tag from rendering in preview mode?", but that would just encourage people to post, then edit their post. They can iteratively tweak the TEX expressions anyhow, which does not head off the accumulation problem.

So, instead I make the suggestion (or ask for the clarification): is there a way to either prevent or periodically cleanup the long-term accumulation of effectively no-longer-used/intermediately-generated PNG files?

Such as through some "smarts" to drop any net-unused, intermediately generated /pics images during preview-edit-preview loops?

OR

Such as through a cron or somesuch? I'm not sure what the effort would be, imagining something that has to effectively sweep recent posts against the /pics folder, and weed out those which are deemed unused, etc, although I admit I can imagine it being a headache...

Perhaps my understanding is off to begin with?

Any insight here would be great. This mod is a great addition to my forum, it offers a great value to my membership!

[Pod]

Originally Posted by Ptah

Hi. I've installed it on my forum and it works great!

I have a concern though, which becomes a suggestion/request of sorts.

My understanding is that each expression rendered results in a new PNG file in /pics. Makes sense, if so. So with that understanding...

/pics files accumulating when users iteratvely tweak TEX expressions in preview/edit?
Let's say someone is using the preview-edit-preview in a loop whilst developing a post, and tweaking the latex expression each time, hence generating however many intermediate PNG files (per latex expression tweak), only the last of which is really useful (the one they wind up posting). This results in a pile of images sitting around under /pics that are effectively wasted space. I don't think this is too unrealistic of a scenario, being that people may see a problem with their (say, nontrivial) latex in preview, so then tweak it, iteratively, until they see what the want.

Concern of "waste" /pics accumulation
Now my concern. Over time this can become a problem, lots of wasted space.

For instance, anticipating this may be a problem, I have asked the users of my forum to consider using one of the freely available online, interactive LaTeX renderers to "draft" their expressions before posting them on my forum, hoping to head off useless PNG file accumulations.

Extent suggestions/workarounds...
So. I understand that you can purge the post cache and rebuild, but that seems to be long-term unfeasible as the forum ages, and accumulates many posts, right? Just to purge unused /pic files, you have to drop and rebuild the whole post cache? Seems like hitting it with a hammer, unless I've misunderstood?

Suggestion/Question?

I'd consider making the suggestion, "Hey, can we disable the TEX tag from rendering in preview mode?", but that would just encourage people to post, then edit their post. They can iteratively tweak the TEX expressions anyhow, which does not head off the accumulation problem.

So, instead I make the suggestion (or ask for the clarification): is there a way to either prevent or periodically cleanup the long-term accumulation of effectively no-longer-used/intermediately-generated PNG files?

Such as through some "smarts" to drop any net-unused, intermediately generated /pics images during preview-edit-preview loops?

OR

Such as through a cron or somesuch? I'm not sure what the effort would be, imagining something that has to effectively sweep recent posts against the /pics folder, and weed out those which are deemed unused, etc, although I admit I can imagine it being a headache...

Perhaps my understanding is off to begin with?

Any insight here would be great. This mod is a great addition to my forum, it offers a great value to my membership!

What I do myself is to add this line in a crontab:

50 4 * * * find PATH_TO_FORUM/vlatex/pics -atime +35 -name "*_*" | xargs rm -f

This removes every morning (at 4:50) all files from the directory that were not accessed in the last 35 days. You can tune this command as you wish.

It is important that the number of days is slightly larger than the duration of your post cache (it's an option somewhere in vb's admincp), because if the post is cached, then bbcodes are not reprocessed.

If you ever empty the pics folder, then also clear or rebuild the post cache (which can be done in the admincp as well). Doing so, png's will be rebuilt the first time every post is visited by someone (I suggest doing this with the forum closed and visit yourself the posting page, 'cause the large number of formulas in the quick-menu take a while to be processed).

[Ptah]

Pod, thanks for the quick reply. I'll play around with your suggested cron-based (and rather pragmatic) approach and see how that works.

Thanks again.

[oxide11]

After update to latest Vbulletin 4.2.2 and php 5.4 on server this function doesn't work

\begin{align} (formula) \end{align}
\begin{gather} (formula) \end{gather}
\begin{eqnarray} (formula) \end{eqnarray}

Does anyone know what can cause this problem.

[Pod]

Originally Posted by oxide11

After update to latest Vbulletin 4.2.2 and php 5.4 on server this function doesn't work

\begin{align} (formula) \end{align}
\begin{gather} (formula) \end{gather}
\begin{eqnarray} (formula) \end{eqnarray}

Does anyone know what can cause this problem.

I don't think the update should affect the mod. You could try to reinstall the mod.

[tn5421]

You should include mimetex in your install, for people who's servers don't have LaTeX installed.

[Pod]

Originally Posted by tn5421

You should include mimetex in your install, for people who's servers don't have LaTeX installed.

mimeTeX can be easily integrated with VB using the built in bbcode editor, no need for a mod if you're going with mimeTeX

[Daniel]

Any suggestions for something like this? I'm on GoDaddy shared hosting (linux)