Lilak is an open source project for generating Persian dictionary for hunspell spell checker based on Persian Morphology. In Persian language affixes can change the meaning of the word. Some suffixes attached to a word as short form of verbs. For detecting these suffixes we need to know the part-of-speech and in some cases the exact pronunciation of the word. Lilak has a lexicon of Persian words with part-of-speech tags. Lilak builds a dictionary for hunspell to predict the best form of compound words based on morphological rules. Lilak works under Python3.

lilak
  |-- build           : Build folder. 
  |
  |-- src
  |   |-- lilak.py      : Python script for building lilak dictionary
  |   |-- lilak-test.py : Python script to test lilak accuracy 
  |   |-- lexicon       : Lexicon of Persian words with part-of-speech tags
  |   |-- affixes       : Affix (prefix or suffix) rules
  |   |-- dic_users     : List of words without POS tag.
  |   |-- dic_delta     : List of foreign words in Persian.
  |   \-- verbs.htm     : List of Persian verbs (unstemmed)
  |
  |-- test
  |   |-- text1       : 'Farsi(Persian) is Sugar', A short story by Mohammad-Ali Jamalzadeh
  |   |-- text2       : 'A Hekayat' By Saadi 
  |   |-- text3       : 'A Ghazal' By Hafez 
  |   |-- text4       : 'Yazdgerd Kingdom' By Ferdowsi
  |   |-- text5       : 'A Ghazal' By Muhammad Husayn Tabataba'i 
  |   |-- text6       : 'Have a Safe Trip' A poem by Shafii Kadkani
  |   |-- text7       : 'Se Tar' A short story by Jalal Al-e-Ahmad
  |   |-- text8       : 'End of Shahname' By Mehdi Akhavan-Sales
  |   |-- text9       : 'The Water's Footsteps' By Sohrab Sepehri
  |   |-- text10      : 'Nei Name' By Rumi
  |   \-- verbs       : Some inflected verbs
  |
  |-- README.md       : 
  \-- LICENCE         : License file

To build dictionary, run lilak.py from 'src' folder:

python lilak.py

You can find the compiled dictionary at the 'build' folder.

Before running test make sure you have installed hunspell:

pip install hunspell

To test Lilak, run dotest.py from 'test' folder:

python lilak-test.py

You can find the result at 'result.log' file

The best way that you can contribute on this project is collecting words with part-of-speech tags. It will help me to make dictionary more useful and accurate.

Part-of-speech is important to build lilak. It should classified in main types like: verb, noun, adjective, ... Also some other tags will be useful. like tense of verb, singular or plural, ... Check Lexicon for more information

Also if you found a mistakes please report it via email to me.

If you have any questions about the project, please don't hesitate to ask.

You can download compiled dictionaries at sourceforge.

For installing Lilak on OpenOffice just check here.

For installing Lilak on Mozilla Firefox just check here.

For installing Lilak on Google Chrome: Go to Settings. Find Language and input settings. Add Persian language and make sure you have enabled the specll checker option.

lilac in English came from French lilac "shrub of genus Syringa with mauve flowers" from Spanish lilac, from Arabic lilak, from Persian lilak, variant of nilak "bluish"

Abolhassan Najafi was an associate member of Iran's Academy of Persian Language and Literature. His most famous books is "Ghalat Nanevisim" (Let’s not write incorrect).

Special thanks to

• Dr. Hamid Farrroukh : alefbaye2om
• Dr. Hamid Hassani : wikipedia
• 5j9 : github
• Reza Moshksar : github