• Stars
    star
    2,126
  • Rank 21,706 (Top 0.5 %)
  • Language
    Haskell
  • License
    GNU General Publi...
  • Created about 16 years ago
  • Updated 10 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

A wiki using HAppS, pandoc, and git

Gitit

Gitit is a wiki program written in Haskell. It uses Happstack for the web server and pandoc for markup processing. Pages and uploaded files are stored in a git, darcs, or mercurial repository and may be modified either by using the VCS's command-line tools or through the wiki's web interface. By default, pandoc's extended version of markdown is used as a markup language, but reStructuredText, LaTeX, HTML, DocBook, or Emacs Org-mode markup can also be used. Gitit can be configured to display TeX math (using texmath) and highlighted source code (using highlighting-kate).

Other features include

  • plugins: dynamically loaded page transformations written in Haskell (see "Network.Gitit.Interface")

  • categories

  • TeX math

  • syntax highlighting of source code files and code snippets (using highlighting-kate)

  • caching

  • Atom feeds (site-wide and per-page)

  • a library, "Network.Gitit", that makes it simple to include a gitit wiki in any happstack application

Getting started

Compiling and installing gitit

The most reliable way to install gitit from source is to get the stack tool. Then clone the gitit repository and use stack to install:

git clone https://github.com/jgm/gitit
cd gitit
stack install

Alternatively, instead of using stack, you can get the Haskell Platform and do the following:

cabal update
cabal install gitit

This will install the latest released version of gitit. To install a version of gitit checked out from the repository, change to the gitit directory and type:

cabal install

The cabal tool will automatically install all of the required haskell libraries. If all goes well, by the end of this process, the latest release of gitit will be installed in your local .cabal directory. You can check this by trying:

gitit --version

If that doesn't work, check to see that gitit is in your local cabal-install executable directory (usually ~/.cabal/bin). And make sure ~/.cabal/bin is in your system path.

Running gitit

To run gitit, you'll need git in your system path. (Or darcs or hg, if you're using darcs or mercurial to store the wiki data.)

Gitit assumes that the page files (stored in the git repository) are encoded as UTF-8. Even page names may be UTF-8 if the file system supports this. So you should make sure that you are using a UTF-8 locale when running gitit. (To check this, type locale.)

Switch to the directory where you want to run gitit. This should be a directory where you have write access, since three directories, static, templates, and wikidata, and two files, gitit-users and gitit.log, will be created here. To start gitit, just type:

gitit

If all goes well, gitit will do the following:

  1. Create a git repository, wikidata, and add a default front page.
  2. Create a static directory containing files to be treated as static files by gitit.
  3. Create a templates directory containing HStringTemplate templates for wiki pages.
  4. Start a web server on port 5001.

Check that it worked: open a web browser and go to http://localhost:5001.

You can control the port that gitit runs on using the -p option: gitit -p 4000 will start gitit on port 4000. Additional runtime options are described by gitit -h.

Using gitit

Wiki links and formatting

For instructions on editing pages and creating links, see the "Help" page.

Gitit interprets links with empty URLs as wikilinks. Thus, in markdown pages, [Front Page]() creates an internal wikilink to the page Front Page. In reStructuredText pages, `Front Page <>`_ has the same effect.

If you want to link to a directory listing for a subdirectory, use a trailing slash: [foo/bar/]() creates a link to the directory for foo/bar.

Page metadata

Pages may optionally begin with a metadata block. Here is an example:

---
format: latex+lhs
categories: haskell math
toc: no
title: Haskell and
  Category Theory
...

\section{Why Category Theory?}

The metadata block consists of a list of key-value pairs, each on a separate line. If needed, the value can be continued on one or more additional line, which must begin with a space. (This is illustrated by the "title" example above.) The metadata block must begin with a line --- and end with a line ... optionally followed by one or more blank lines. (The metadata block is a valid YAML document, though not all YAML documents will be valid metadata blocks.)

Currently the following keys are supported:

format : Overrides the default page type as specified in the configuration file. Possible values are markdown, rst, latex, html, markdown+lhs, rst+lhs, latex+lhs. (Capitalization is ignored, so you can also use LaTeX, HTML, etc.) The +lhs variants indicate that the page is to be interpreted as literate Haskell. If this field is missing, the default page type will be used.

categories : A space or comma separated list of categories to which the page belongs.

toc : Overrides default setting for table-of-contents in the configuration file. Values can be yes, no, true, or false (capitalization is ignored).

title : By default the displayed page title is the page name. This metadata element overrides that default.

Highlighted source code

If gitit was compiled against a version of pandoc that has highlighting support (see above), you can get highlighted source code by using delimited code blocks:

~~~ {.haskell .numberLines}
qsort []     = []
qsort (x:xs) = qsort (filter (< x) xs) ++ [x] ++
               qsort (filter (>= x) xs) 
~~~

To see what languages your pandoc was compiled to highlight:

pandoc -v

Configuring and customizing gitit

Configuration options

Use the option -f [filename] to specify a configuration file:

gitit -f my.conf

The configuration can be split between several files:

gitit -f my.conf -f additional.conf

One use case is to keep sensible part of the configuration outside of a SCM (oauth client secret for example).

If this option is not used, gitit will use a default configuration. To get a copy of the default configuration file, which you can customize, just type:

gitit --print-default-config > my.conf

The default configuration file is documented with comments throughout.

The static directory

On receiving a request, gitit always looks first in the static directory (or in whatever directory is specified for static-dir in the configuration file). If a file corresponding to the request is found there, it is served immediately. If the file is not found in static, gitit next looks in the static subdirectory of gitit's data file ($CABALDIR/share/gitit-x.y.z/data). This is where default css, images, and javascripts are stored. If the file is not found there either, gitit treats the request as a request for a wiki page or wiki command.

So, you can throw anything you want to be served statically (for example, a robots.txt file or favicon.ico) in the static directory. You can override any of gitit's default css, javascript, or image files by putting a file with the same relative path in static. Note that gitit has a default robots.txt file that excludes all URLs beginning with /_.

Note: if you set static-dir to be a subdirectory of repository-path, and then add the files in the static directory to your repository, you can ensure that others who clone your wiki repository get these files as well. It will not be possible to modify these files using the web interface, but they will be modifiable via git.

Using a VCS other than git

By default, gitit will store wiki pages in a git repository in the wikidata directory. If you'd prefer to use darcs instead of git, you need to add the following field to the configuration file:

repository-type: Darcs

If you'd prefer to use mercurial, add:

repository-type: Mercurial

This program may be called "darcsit" instead of "gitit" when a darcs backend is used.

Note: we recommend that you use gitit/darcsit with darcs version 2.3.0 or greater. If you must use an older version of darcs, then you need to compile the filestore library without the (default) maxcount flag, before (re)installing gitit:

cabal install --reinstall filestore -f-maxcount
cabal install --reinstall gitit

Otherwise you will get an error when you attempt to access your repository.

Changing the theme

To change the look of the wiki, you can modify custom.css in static/css.

To change the look of printed pages, copy gitit's default print.css to static/css and modify it.

The logo picture can be changed by copying a new PNG file to static/img/logo.png. The default logo is 138x155 pixels.

To change the footer, modify templates/footer.st.

For more radical changes, you can override any of the default templates in $CABALDIR/share/gitit-x.y.z/data/templates by copying the file into templates, modifying it, and restarting gitit. The page.st template is the master template; it includes the others. Interpolated variables are surrounded by $s, so literal $ must be backslash-escaped.

Adding support for math

To write math on a markdown-formatted wiki page, just enclose it in dollar signs, as in LaTeX:

Here is a formula:  $\frac{1}{\sqrt{c^2}}$

You can write display math by enclosing it in double dollar signs:

$$\frac{1}{\sqrt{c^2}}$$

Gitit can display TeX math in three different ways, depending on the setting of math in the configuration file:

  1. mathjax (default): Math will be rendered using the MathJax javascript.

  2. mathml: Math will be converted to MathML using texmath. This method works with IE+mathplayer, Firefox, and Opera, but not Safari.

  3. raw: Math will be rendered as raw LaTeX codes.

Restricting access

If you want to limit account creation on your wiki, the easiest way to do this is to provide an access-question in your configuration file. (See the commented default configuration file.) Nobody will be able to create an account without knowing the answer to the access question.

Another approach is to use HTTP authentication. (See the config file comments on authentication-method.)

Authentication through github

If you want to authenticate the user from github through oauth2, you need to register your app with github to obtain a OAuth client secret and add the following section to your configuration file:

[Github]
oauthclientid: 01239456789abcdef012
oauthclientsecret: 01239456789abcdef01239456789abcdef012394
oauthcallback: http://mysite/_githubCallback
oauthoauthorizeendpoint: https://github.com/login/oauth/authorize
oauthaccesstokenendpoint: https://github.com/login/oauth/access_token
## Uncomment if you are checking membership against an organization and change
## gitit-testorg to this organization:
# github-org: gitit-testorg

The github authentication uses the scope user:email. This way, gitit gets the email of the user, and the commit can be assigned to the right author if the wikidata repository is pushed to github. Additionally, it uses read:org if you uses the option github-org to check membership against an organization.

To push your repository to gitub after each commit, you can add the file post-commit with the content below in the .git/hooks directory of your wikidata repository.

#!/bin/sh
git push origin master 2>> logit

Plugins

Plugins are small Haskell programs that transform a wiki page after it has been converted from Markdown or another source format. See the example plugins in the plugins directory. To enable a plugin, include the path to the plugin (or its module name) in the plugins field of the configuration file. (If the plugin name starts with Network.Gitit.Plugin., gitit will assume that the plugin is an installed module and will not look for a source file.)

Plugin support is enabled by default. However, plugin support makes the gitit executable considerably larger and more memory-hungry. If you don't need plugins, you may want to compile gitit without plugin support. To do this, unset the plugins Cabal flag:

cabal install --reinstall gitit -f-plugins

Note also that if you compile gitit for executable profiling, attempts to load plugins will result in "internal error: PAP object entered!"

Accessing the wiki through git

All the pages and uploaded files are stored in a git repository. By default, this lives in the wikidata directory (though this can be changed through configuration options). So you can interact with the wiki using git command line tools:

git clone ssh://my.server.edu/path/of/wiki/wikidata
cd wikidata
vim Front\ Page.page  # edit the page
git commit -m "Added message about wiki etiquette" Front\ Page.page
git push

If you now look at the Front Page on the wiki, you should see your changes reflected there. Note that the pages all have the extension .page.

If you are using the darcs or mercurial backend, the commands will be slightly different. See the documentation for your VCS for details.

Performance

Caching

By default, gitit does not cache content. If your wiki receives a lot of traffic or contains pages that are slow to render, you may want to activate caching. To do this, set the configuration option use-cache to yes. By default, rendered pages, and highlighted source files will be cached in the cache directory. (Another directory can be specified by setting the cache-dir configuration option.)

Cached pages are updated when pages are modified using the web interface. They are not updated when pages are modified directly through git or darcs. However, the cache can be refreshed manually by pressing Ctrl-R when viewing a page, or by sending an HTTP GET or POST request to /_expire/path/to/page, where path/to/page is the name of the page to be expired.

Users who frequently update pages using git or darcs may wish to add a hook to the repository that makes the appropriate HTTP request to expire pages when they are updated. To facilitate such hooks, the gitit cabal package includes an executable expireGititCache. Assuming you are running gitit at port 5001 on localhost, and the environment variable CHANGED_FILES contains a list of the files that have changed, you can expire their cached versions using

expireGititCache http://localhost:5001 $CHANGED_FILES

Or you can specify the files directly:

expireGititCache http://localhost:5001 "Front Page.page" foo/bar/baz.c

This program will return a success status (0) if the page has been successfully expired (or if it was never cached in the first place), and a failure status (> 0) otherwise.

The cache is persistent through restarts of gitit. To expire all cached pages, simply remove the cache directory.

Idle

By default, GHC's runtime will repeatedly attempt to collect garbage when an executable like Gitit is idle. This means that gitit will, after the first page request, never use 0% CPU time and sleep, but will use ~1%. This can be bad for battery life, among other things.

To fix this, one can disable the idle-time GC with the runtime flag -I0:

gitit -f my.conf +RTS -I0 -RTS

Note:

To enable RTS, cabal needs to pass the compile flag -rtsopts to GHC while installing.

cabal install --reinstall gitit --ghc-options="-rtsopts"

Using gitit with apache

Most users who run a public-facing gitit will want gitit to appear at a nice URL like http://wiki.mysite.com or http://mysite.com/wiki rather than http://mysite.com:5001. This can be achieved using apache's mod_proxy.

Proxying to http://wiki.mysite.com

Set up your DNS so that http://wiki.mysite.com maps to your server's IP address. Make sure that the mod_proxy, mod_proxy_http and mod_rewrite modules are loaded, and set up a virtual host with the following configuration:

<VirtualHost *>
    ServerName wiki.mysite.com
    DocumentRoot /var/www/
    RewriteEngine On
    ProxyPreserveHost On
    ProxyRequests Off

    <Proxy *>
       Order deny,allow
       Allow from all
    </Proxy>

    ProxyPassReverse /    http://127.0.0.1:5001
    RewriteRule ^(.*) http://127.0.0.1:5001$1 [P]

    ErrorLog /var/log/apache2/error.log
    LogLevel warn

    CustomLog /var/log/apache2/access.log combined
    ServerSignature On

</VirtualHost>

Reload your apache configuration and you should be all set.

Using nginx to achieve the same

Drop a file called wiki.example.com.conf into /etc/nginx/conf.d (or where ever your distribution puts it).

server {
    listen 80;
    server_name wiki.example.com
    location / {
        proxy_pass        http://127.0.0.1:5001/;
        proxy_set_header  X-Real-IP  $remote_addr;
        proxy_redirect off;
    }
    access_log /var/log/nginx/wiki.example.com.log main;
}

Reload your nginx config and you should be all set.

Proxying to http://mysite.com/wiki

Make sure the mod_proxy, mod_headers, mod_proxy_http, and mod_proxy_html modules are loaded. mod_proxy_html is an external module, which can be obtained [here] (http://apache.webthing.com/mod_proxy_html/). It rewrites URLs that occur in web pages. Here we will use it to rewrite gitit's links so that they all begin with /wiki/.

First, tell gitit not to compress pages, since mod_proxy_html needs uncompressed pages to parse. You can do this by setting the gitit configuration option

compress-responses: no

Second, modify the link in the reset-password-message in the configuration file: instead of

http://$hostname$:$port$$resetlink$

set it to

http://$hostname$/wiki$resetlink$

Restart gitit.

Now add the following lines to the apache configuration file for the mysite.com server:

# These commands will proxy /wiki/ to port 5001

ProxyRequests Off

<Proxy *>
  Order deny,allow
  Allow from all
</Proxy>

ProxyPass /wiki/ http://127.0.0.1:5001/

<Location /wiki/>
  SetOutputFilter  proxy-html
  ProxyPassReverse /
  ProxyHTMLURLMap  /   /wiki/
  ProxyHTMLDocType "<!DOCTYPE html PUBLIC '-//W3C//DTD XHTML 1.0 Strict//EN' 'http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd'>" XHTML
  RequestHeader unset Accept-Encoding
</Location>

Reload your apache configuration and you should be set.

For further information on the use of mod_proxy_http to rewrite URLs, see the mod_proxy_html guide.

Using gitit as a library

By importing the module Network.Gitit, you can include a gitit wiki (or several of them) in another happstack application. There are some simple examples in the haddock documentation for Network.Gitit.

Reporting bugs

Bugs may be reported (and feature requests filed) at https://github.com/jgm/gitit/issues.

There is a mailing list for users and developers at http://groups.google.com/group/gitit-discuss.

Acknowledgements

A number of people have contributed patches:

  • Gwern Branwen helped to optimize gitit and wrote the InterwikiPlugin. He also helped with the Feed module.
  • Simon Michael contributed the patch adding RST support.
  • Henry Laxen added support for password resets and helped with the apache proxy instructions.
  • Anton van Straaten made the process of page generation more modular by adding Gitit.ContentTransformer.
  • Robin Green helped improve the plugin API and interface, and fixed a security problem with the reset password code.
  • Thomas Hartman helped improve the index page, making directory browsing persistent, and fixed a bug in template recompilation.
  • Justin Bogner improved the appearance of the preview button.
  • Kohei Ozaki contributed the ImgTexPlugin.
  • Michael Terepeta improved validation of change descriptions.
  • mightybyte suggested making gitit available as a library, and contributed a patch to ifLoggedIn that was needed to make gitit usable with a custom authentication scheme.

I am especially grateful to the darcs team for using darcsit for their public-facing wiki. This has helped immensely in identifying issues and improving performance.

Gitit's default visual layout is shamelessly borrowed from Wikipedia. The stylesheets are influenced by Wikipedia's stylesheets and by the bluetrip CSS framework (see BLUETRIP-LICENSE). Some of the icons in img/icons come from bluetrip as well.

More Repositories

1

pandoc

Universal markup converter
Haskell
34,313
star
2

djot

A light markup language
HTML
1,557
star
3

peg-markdown

An implementation of markdown in C, using a PEG grammar
C
686
star
4

pandocfilters

A python module for writing pandoc filters, with a collection of examples
Python
511
star
5

pandoc-templates

Templates for pandoc, tagged to release
HTML
418
star
6

yst

create static websites from YAML data and string templates
Haskell
373
star
7

texmath

A Haskell library for converting LaTeX math to MathML.
Haskell
291
star
8

pandoc-citeproc

Library and executable for using citeproc with pandoc
Haskell
288
star
9

lunamark

Lua library for conversion between markup formats
C
192
star
10

skylighting

A Haskell syntax highlighting library with tokenizers derived from KDE syntax highlighting descriptions
Haskell
185
star
11

citeproc

CSL citation processing library in Haskell
Haskell
138
star
12

commonmark-hs

Pure Haskell commonmark parsing library, designed to be flexible and extensible
Haskell
130
star
13

djot.js

JavaScript implementation of djot
TypeScript
120
star
14

highlighting-kate

A syntax highlighting library in Haskell, based on Kate syntax definitions
HTML
109
star
15

cheapskate

Experimental markdown processor in Haskell
HTML
105
star
16

pandoc-types

types for representing structured documents
Haskell
105
star
17

gitit2

A reimplementation of gitit in Yesod
Haskell
94
star
18

lcmark

Flexible CommonMark converter
Lua
54
star
19

doctemplates

Pandoc-compatible templating system
Haskell
49
star
20

zip-archive

Native Haskell library for working with zip archives
Haskell
44
star
21

cmark-hs

Haskell bindings to libcmark commonmark parser
C
43
star
22

djot.lua

Lua parser for the djot light markup language
Lua
39
star
23

typst-hs

Haskell library for parsing and evaluating typst
Haskell
32
star
24

dotvim

My vim configuration
Vim Script
30
star
25

scripts

A collection of small scripts to do various things
Shell
28
star
26

filestore

A versioning file store backed by git, darcs, or mercurial
Haskell
28
star
27

pandoc-website

Source files for pandoc's website
Lua
28
star
28

illuminate

An efficient syntax highlighting library in Haskell, using alex-generated lexers
Haskell
26
star
29

emojis

Haskell library for emojis
Haskell
25
star
30

markdown-peg

A Haskell implementation of markdown using a PEG grammar
Haskell
24
star
31

pandoc-server

Simple server app for pandoc conversions.
Haskell
20
star
32

doclayout

A prettyprinting library designed for laying out plain text documents
Haskell
20
star
33

standalone-html

Incorporates external dependencies into HTML file using data: URI scheme
Haskell
19
star
34

pandoc-tex2svg

Pandoc filter to convert math to SVG using MathJax-node's tex2svg
HTML
19
star
35

cloudlib

tools for keeping a library of books and articles on Amazon's S3 and SimpleDB
Ruby
19
star
36

cmark-lua

Lua bindings to libcmark CommonMark parser
C
17
star
37

HeX

a flexible text macro system
Haskell
17
star
38

djoths

Haskell parser for the djot light markup language
Haskell
17
star
39

unicode-collation

Haskell implementation of the Unicode Collation Algorithm
Haskell
16
star
40

sep-offprint

Creates formatted "offprints" of Stanford Encyclopedia of Philosophy entries.
15
star
41

BayHac2014

Slides for my presentation on pandoc at BayHac2014
TeX
14
star
42

cmarkpdf

Steps towards a PDF renderer for cmark using libharu
C
14
star
43

lunamark-standalone

Standalone version of lunamark (compiled with no library dependencies)
C
12
star
44

commonmarker

Ruby wrapper for libcmark (CommonMark parser)
Ruby
12
star
45

ipynb

Data structures and JSON serializer/deserializer for Jupyter notebooks (.ipynb) format.
Jupyter Notebook
11
star
46

hsb2hs

Preprocessor for inserting literals with binary blobs into Haskell programs.
Haskell
11
star
47

gogar

Computer implementation of Robert Brandom's "game of giving and asking for reasons," from Making It Explicit, chapter 3.
Ruby
10
star
48

emacsd

emacs configuration
Emacs Lisp
9
star
49

hscommonmark

pure Haskell CommonMark parser
Haskell
9
star
50

recaptcha

Haskell library for using the reCAPTCHA service
Haskell
8
star
51

select-meta

Pandoc lua filter for constructing metadata from YAML data sources using queries
Lua
8
star
52

html2cmark

Lua library to convert HTML5 to commonmark
Lua
8
star
53

citeproc-hs-bin

Command-line interface to the citeproc-hs CSL citation processing library
Haskell
8
star
54

grammata

Well-typed system for generating documents in multiple formats
Haskell
7
star
55

hw2gitit

Script to convert haskellwiki pages to a gitit wiki
Haskell
7
star
56

ecstatic

Static website management using tenjin templates and YAML data files
Ruby
7
star
57

hsgit

A higher-level interface to libgit2 functions than hlibgit2
Haskell
6
star
58

trypandoc

Live demo of pandoc
JavaScript
6
star
59

pandoc-highlight

Filter and library for using pandoc with highlighting-kate
Haskell
6
star
60

commonmark-lua

Lua binding to libcmark commonmark parser
Lua
5
star
61

rfc5051

Haskell implementation of RFC5051, simple unicode collation.
Haskell
5
star
62

jgm.github.com

jgm's web pages on github
4
star
63

rocks

luarocks repository
4
star
64

cmark-fuzz-data

A minimal fuzz test suite for cmark created by american fuzzy lop and afl-cmin
3
star
65

GHCUnicodeAlt

Improved version of GHC.Unicode, with benchmarks
Haskell
3
star
66

luacmark

Lua binding to CommonMark
C
2
star
67

typst-symbols

Defines symbols and emoji used in typst
Haskell
2
star