My Dictionary for Chrome

My Dictionary is a Chrome extension that allows you to quickly look up definitions of words and phrases in a clean inline box. The definitions come from Wiktionary, but have been preprocessed to reduce formatting and provide faster lookup. Hold Alt and double-click or select any word to look it up - an inline box will open to the dictionary entry of that word, usually including examples, audio pronunciation and synonyms.

Parts

The application is powered by three separate parts:

• The actual Chrome extension that is embedded in pages, queries the lookup server for definitions and formats them for display. Written in Javascript.
• The lightweight lookup server script that receives word or phrase queries and returns matches from a local database. It also does some trivial spelling suggestions based on a a Soundex index. Written in PHP. The canonical server is hosted on http://wiktionary.org.
• The scripts which parse a Wiktionary dump and produce a database of machine readable definitions appropriate for quick lookup. Written mainly in Python.

Building

The extension itself, located in the extension subfolder, does not require building - it can be loaded or packed into a CRX by Chrome as is.

The lookup server script, located in the lookup_server subfolder, does not require building either. However, it expects the lookup database built by the Wiktionary parser to be accessible - the database settings can be edited at the top of the file (the $MYSQL_* variables).

The Wiktionary parser is a little more complicated. Here's a step-by-step guide to running the whole thing:

1. Run wiktionary_setup/build-mediawiki.sh to create a MediaWiki installation with the ExpandTemplates and ParserFunctions extensions and the additions needed by the Wiktionary parser. This is a dummy server that is only used to expand Wiktionary templates. You may need to edit the database settings in wiktionary_setup/build-mediawiki.sh before running the script. Note that this script requires the following commands to be available:
   • wget: for getting the MediaWiki source tarball.
   • tar: for unpacking the source tarball.
   • patch: for applying parser-specific patches.
   • svn: for getting the required mediawiki extensions.
   • mysql: for creating the databases (mysqld must be running).
2. Setup a PHP-enabled Apache server and make sure it serves the mediawiki folder created in the previous step. It is not necessary (or recommended) to serve directly from the DocumentRoot.
3. Edit mediawiki/LocalSettings.php to set the database connection details ($wgDB*) and the URL path from which the mediawiki folder is served ($wgScriptPath).
4. Test that the MediaWiki server is working by pointing your browser at the path from which MediaWiki is being served. It should return an empty page with the HTTP status code 200.
5. Download and extract a Wiktionary dump, e.g. from: http://dumps.wikimedia.org/enwiktionary/latest/enwiktionary-latest-pages-articles.xml.bz2
6. Adjust the database, dump path and URL settings in wiktionary/__main__.py.
7. Run `wiktionary/main.py. This should execute the whole pipeline.
8. If everything went right, the final database in the format used by the lookup server will be in LOOKUP_DB (as specified in wiktionary/__main__.py).