While this does not aim to have anywhere near the complete functionality of Zyzzyva, it does aim to provide a relatively simple interface to the Zyzzyva database, allowing for querying and so-on. It will not include cardbox and other such functionality, primarily because I do not use these, but also because that would obligate a layer of persistence and user log-ins.

The interface "serving" the Zyzzyva database will be written in Python, and will be a CGI-based JSON service. Thus, while POST and GET access is available, all the results will be returned as plain JSON.

In future, this may be extended to serve as a JSON-RPC-based server instead, but for the time being this will hopefully suffice.

The initial interface will be written with jQuery and jQuery UI in mind; these will have access to the information in the database via AJAX, and this information will be presented in a (hopefully) clean manner to the user.

Initial to-do list consists on the following:

1. "Search" option, using all of the options presented in the current Zyzzyva interface.
2. "Adjudication" option, providing a similar interface to the current Zyzzyva "Word Judge" interface.
3. A "Quiz" mode, identical to the current Zyzzyva "Quiz" interface.

The primary lexicon choice will be SOWPODS, via CSW; this data package will not be included in the software for licensing reasons. To generate this database, you will need to install a copy of Zyzzyva and select the lexicon desired; once it has generated, you will find a copy in ~/Zyzzyva, where the lexicon file is ACRONYM.db.

Currently, the software supports all of the lexicons that are generated by the Zyzzyva main executable.

A series of 'maintenance' scripts are available within the directory of the same name:

• 'compact' removes probability and playability columns, compressing the size of the database.
• 'database' inserts new indexes on alphagrams and hook status; this increases the size of the database, but it does help to improve the speed of Anagram searches.
• 'consogram' creates a new column for each word of the database, storing the 'consogram' of each word. Similar to alphagrams, these are an alphabetically ordered representation of a word, only grouping consonants together at the start, and vowels together at the end. The idea comes mainly from Barry Harridge's excellent Consograms book.

All of the code, where applicable, is licensed under the terms of the MIT License; furhter information on this subject can be found in the COPYING.rst file.