User:N0nick/GSoC Journal

Bonding Period week 1: 4/25-5/1

• Got the development environment ready. apertium, lttoolbox and other tools and tests all working properly.
• Filled the Pending Tests page with some translations (based on the ones in the mt-en page).
• Started working on a script to generate a Maltese monodix from external sources. Nothing to show yet.
  • Some Maltese newspapers (suggested by spectie): [1] [2] [3] [4]
  • In the chatroom, User:zfe noticed the Bitextor project. Might be useful, started experimenting with it tonight.
• Notified 2 TAU professors (both specializing in CL) about my project, both agreed to offer help if necessary.
• Wrote to a contact related to the MaltiLex project, looking for better contact (perhaps through my university's faculty).

Bonding Period week 2: 5/2 - 5/8

• Picked up and started reading the Teach Yourself Maltese grammar book.
• Wrote the framework for a script that generates fullform Maltese verb lists. [5]
  • We worked on splitting verbs into categorizes and optional subclasses, writing rules (based on stem affixes, roots and vowels) in a python script for each class.
  • Found out about the way [Wiktionary] stores conjugation data about the verbs it has; very useful for creating new rule groups. [6]. Finished converting these tables this week except for [7] [8] (that are identical to strong.py apart for a transformation in imperfect forms).
• spectre contacted Prof. Adam Ussishkin and he provided us with Maltese verb lists that we need to look over. [9]
• Contacted John J. Camilleri regarding his Maltese morphology slideshow, asking for data on verb conjugation.
• Will add the verbs already added to the Hebrew and bidi dix files.

Bonding Period week 3: 5/9 - 5/15

• Continued studying Maltese from the grammar book.
• Added verbs from John Camilleri's slideshow to the Maltese analyser program.

Bonding Period week 4: 5/16 - 5/22

• Added closed categories to the Maltese analyser (pronouns, prepositions, conjunctions, determiners, numerals)
• Added closed categories to the Hebrew dictionary
• Added closed categories to the bidix

Week 1: 5/23 - 5/29

• Fixed bugs in some of the environment tools I was using
• Added missing Hebrew determiners and pronouns
• Added missing closed categories to the bidix

Week 2: 5/30 - 6/5

• Generate Hebrew verb speling file from hspell output
• Format Hebrew verbs speling file as Apertium dix file
• Research handling of attached/clitic pronouns on both Maltese & Hebrew

Week 3: 6/6 -6/12

• Fixed Hebrew noun paradigms (automatically generated from hspell)
• Add existing verbs to bidix
• Add missing determiners to Maltese file (as per Fran's email)

Week 4: 6/13 - 6/19

--- (Studied for exam, haven't achieved much)

Week 5: 6/20 - 6/26

• Fully analyse all words in sample paragraph 1
• Added all paragraph words to bidix, then update Hebrew dix accordingly
• Added a list of Maltese proper nouns, adverbs and updated dix, bidix
• Fixed bugs in hspell output for plural nouns, verbs

Week 6: 6/27 - 7/3

• Worked on mt.dix to achieve better coverage of Maltese corpus
• Added lists of adjectives generated from 'suspected' lists (according to suffixes, etc)
• Used Wiktionary extracting script to load nouns, adjectives from Maltese items in English Wiktionary
• Used wider Maltese corpus received from Kevin Scnanell
• Added more Maltese nouns, adjectives and verbs from top of frequency list
• Contacted Michael Spagnol re Maltese Broken Plurals to receive list of nouns with broken plural form.
• Fixed several noun paradigms
• Searched for documentation on the kien/ikun verb, differences and forms. Contacted Adam Ussishkin & john Camilleri for help with this.

Week 7: 7/4 - 7/10

• Added more nouns, adjectives, adverbs from frequency list
• Updated verbs.py interface, added option to set dictionary restriction
• Added -x negative suffix for all verbs (in verbs.py)
• Fixed all forms of kien/ikun (with help from Kevin & John)
• Added verbs, verb classes from frequency list + corpus
• Added all verified 630 broken plural nouns & adjectives from Tamra Schembri's thesis with sg, pl and gender=GD
• Fixed existing words with gender we have in mt-en dictionary
• Categorized top 1900 words from hitparade frequency list
• Finally acquired the Maltese Descriptive Grammar book!

Week 8: 7/11 - 7/17

• Added more verb paradigms and stems from grammar book
• Wrote gen_stems.py for updating the stems file with ones handled by the new verbs script (temporary solution)
• Added many Maltese nouns, adjectives from mt-en dictionary
• Added '@' terms to bidix by frequency: closed-cats, nouns, adjectives, toponyms and some verbs

Week 9: 7/18 - 7/24

• Added most determiners to bidix
• Added all nouns that has only masc. form to bidix
• Fixed gender transfer for verbs (copied to pronouns)
• Added most (~550) adjectives to bidix
• Fixed some bad / wrong entries in mt.dix

Week 10: 7/25 - 7/31

• Fixed bugs in our modification to hspell that outputs the Hebrew verb dix
• Added most (~150) adverbs to bidix
• Added most (~480) proper nouns to bidix
• Added some determiners
• Fixed some bad entries in the bidix