Difference between revisions of "Task ideas for Google Code-in"
|Line 308:||Line 308:|
|title=script to test coverage of analyser over corresponding wikipedia corpus
|mentors=Jonathan, Wei En
|description=Write a script (in python or ruby) that tests coverage of an Apertium analyser over the latest Wikipedia corpus in that language. One mode of this script should check out a specified language module to a given directory, compile it (or updates it if already existant), and then get the most recently Wikipedia nightly archive for that language and run coverage over it (as much in RAM if possible). In another mode, it should compile the language pair in a docker instance that it then disposes of after successfully running coverage. Scripts exist in Apertium already for finding where a Wikipedia is, extracting a Wikipedia archive into a text file, and running coverage. Ask a mentor for help finding these scripts.
|tags=python, ruby, wikipedia
Revision as of 06:14, 28 October 2019
This is the task ideas page for Google Code-in, here you can find ideas on interesting tasks that will improve your knowledge of Apertium and help you get into the world of open-source development.
The people column lists people who you should get in contact with to request further information. All tasks are 2 hours maximum estimated amount of time that would be spent on the task by an experienced developer, however:
- this is the time expected to take by an experienced developer, you may find that you spend more time on the task because of the learning curve.
- code: Tasks related to writing or refactoring code
- documentation: Tasks related to creating/editing documents and helping others learn more
- research: Tasks related to community management, outreach/marketting, or studying problems and recommending solutions
- quality: Tasks related to testing and ensuring code is of high quality.
- design: Tasks related to user experience research or user interface design and interaction
Clarification of "multiple task" types
- multi = number of students who can do a given task
- dup = number of times a student can do the same task
You can find descriptions of some of the mentors here.
The current task ideas here are for 2019. See Talk:Task ideas for Google Code-in for task ideas from previous years.
|research, quality, documentation||Adopt a Wiki page||Request an Apertium wiki account and adopt a wiki page by updating and fixing any issues with it. Examples of things to update might be documentation that still refers to our SVN repo (we're on GitHub now), documentation of new features, clarification of unclear things, indicating that a page no longer reflects how things are done, or updating documentation to reflect the current options and defaults of various tools.||wiki||*||yes||150|
|research, code||expand coverage of Kyrgyz to English structural transfer||Find a sentence in Kyrgyz that once the lexical items are added to the bilingual dictionary is not fully (or correctly) parsed by the kir-eng-transfer Apertium mode. Determine what rule(s) need(s) to be added (or fixed) to cover this structure, and update apertium-eng-kir.kir-eng.rtx accordingly. You will first want to clone and compile apertium-eng-kir.||Kyrgyz, English, recursive transfer, pairs||JNW, popcorndude||150||10|
|code||Add recursive transfer support to a language pair that doesn't support it||Make a branch of an Apertium language pair that doesn't support recursive transfer and call it "recursive transfer". Add vanilla .rtx files for both directions, and modify Makefile.am and modes.xml so that the branch compiles and runs. See this page for instructions on how to do this.||recursive transfer, pairs||JNW, popcorndude||yes||150||10|
|code||Add 2 recursive transfer rules to a language pair||Add two recursive transfer rules to an Apertium language pair. These rules consist of, at minimum, a syntactic pattern to match, a phrase to combine them into, and an output pattern (more documentation here). If the language pair does not support recursive transfer, make sure to set it up first.||recursive transfer, pairs||JNW, popcorndude||150||20|
|code||Use apertium-init to bootstrap a new language pair||Use the Apertium-init script to bootstrap a new translation pair between two languages which have monolingual modules already in Apertium. To see if a translation pair has already been made, search our repositories on github, and especially ask on IRC. Add 100 common stems to the dictionary. Your submission should be in the form of a repository on github that we can fork to the Apertium organisation.||languages, bootstrap, dictionaries, translators||JNW, wei2912, padth4i||yes||25|
|code||Use apertium-init to bootstrap a new language module||Use the Apertium-init script to bootstrap a new language module that doesn't currently exist in Apertium. To see if a language is available, search our repositories on github, and especially ask on IRC. Add enough stems and morphology to the module so that it analyses and generates at least 100 correct forms. Your submission should be in the form of a repository on github that we can fork to the Apertium organisation. Read more about adding stems...||languages, bootstrap, dictionaries||JNW, wei2912, padth4i||yes||25|
|code||Write 10 lexical selection for an existing translation pair||Add 10 lexical selection rules to an existing translation pair. Submit your work as a github pull request to that pair. Read more...||languages, bootstrap, lexical selection, translators||JNW, sevilay, Unhammer, marcriera, padth4i||25||5|
|code||Write 10 constraint grammar rules for an existing language module||Add 10 constraint grammar rules to an existing language module for a language that you know. Submit your work as a github pull request to that pair. Read more...||languages, bootstrap, constraint grammar||JNW, Unhammer, padth4i||25||5|
|research||Syntactic annotation of text||Pick a text of about 200 words and make a syntactic annotation for it according to the Universal Dependencies treebank. UD Annotatrix can be used for visualisation. Consult with your mentor about the language.||UD, trees, annotation||anakuz, fotonzade, khannatanmai|
|research||Create a UD-Apertium morphology mapping||Choose a language that has a Universal Dependencies treebank and tabulate a potential set of Apertium morph labels based on the (universal) UD morph labels. See Apertium's list of symbols and UD's POS and feature tags for the labels.||morphology, ud, dependencies||JNW,ftyers, fotonzade, khannatanmai, anakuz||5|
|research||Create an Apertium-UD morphology mapping||Choose a language that has an Apertium morphological analyser and adapt it to convert the morphology to UD morphology||morphology, ud, dependencies||JNW, ftyers, fotonzade, khannatanmai, anakuz||5|
|quality||Install Apertium and verify that it works||See Installation for instructions and if you encounter any issues along the way, document them and/or improve the wiki instructions!||bash||ftyers, JNW, Unhammer, anakuz, Josh, fotonzade, sevilay, eirien, wei2912, padth4i||yes||150|
|research||Write a contrastive grammar||Document 6 differences between two (preferably related) languages and where they would need to be addressed in the Apertium pipeline (morph analysis, transfer, etc). Use a grammar book/resource for inspiration. Each difference should have no fewer than 3 examples. Put your work on the Apertium wiki under Language1_and_Language2/Contrastive_grammar. See Farsi_and_English/Pending_tests for an example of a contrastive grammar that a previous GCI student made.||wiki, languages||Mikel, JNW, Josh, xavivars, fotonzade, sevilay, khannatanmai, dolphingarlic, padth4i||yes||40|
|quality||Add 200 new entries to the bilingual dictionary of language pair %AAA%-%BBB%||Our translation systems require large lexicons so as to provide production-quality coverage of any input data. This task requires the student to add 200 new words to a bidirectional dictionary. Read more...||xml, dictionaries||Mikel, anakuz, xavivars, fotonzade, sevilay, Unhammer, eirien, dolphingarlic, wei2912, marcriera, padth4i||yes||40|
|quality||Add 500 new entries to the bilingual dictionary of language pair %AAA%-%BBB%||Our translation systems require large lexicons so as to provide production-quality coverage of any input data. This task requires the student to add 500 new words to a bidirectional dictionary. Read more...||xml, dictionaries||Mikel, anakuz, xavivars, fotonzade, ftyers, sevilay, eirien, dolphingarlic, wei2912, marcriera, padth4i||no||10|
|quality||Post-edit 500 sentences of any public domain text from %AAA% to %BBB%||Many of our systems benefit from statistical methods used with (ideally public domain) bilingual data. For this task, you need to translate a public domain text from %AAA% to %BBB% using any available machine translation system and clean up the translations yourself manually. Commit the post-edited texts (in plain text format) to an existing (via pull request) or if needed new github repository for the pair in dev/ or texts/ folder. The texts are subject to mentor approval.||xml, dictionaries||fotonzade, JNW, ftyers, anakuz, xavivars, Mikel, shardulc, sevilay, Unhammer, khannatanmai, dolphingarlic, wei2912, marcriera||yes||10|
|quality||Disambiguate 500 tokens of text in %AAA%||Run some text through a morphological analyser and disambiguate the output. Discuss with the mentor beforehand to approve the choice of language and text. Read more...||disambiguation||Mikel, anakuz, xavivars, fotonzade, sevilay, khannatanmai, dolphingarlic, wei2912, marcriera, padth4i||yes|
|research||Categorise 100 words from frequency list in %AAA%||Categorise words by frequency into one of the major part-of-speech categories. You will receive a frequency list. Work from top to bottom. At the beginning of each line you should put a letter which categorises the word form by its part-of-speech. For example n for noun, v for verb, etc. Read more...||dictionaries||eirien, khannatanmai, anakuz, marcriera, padth4i||yes||yes|
|research||Categorise 500 words from frequency list in %AAA%||Categorise words by frequency into one of the major part-of-speech categories. You will receive a frequency list. Work from top to bottom. At the beginning of each line you should put a letter which categorises the word form by its part-of-speech. For example n for noun, v for verb, etc. Read more...||dictionaries||eirien, anakuz, sevilay. marcriera||yes|
|research||Evaluate an existing apertium pair %AAA% to %BBB% on a text||Pick an existing apertium pair and get a parallel text for that language pair. Translate %AAA% to %BBB% and evaluate the translation using an automatic evaluation metric like BLEU or/and evaluate it manually (?) Read more...||evaluation||khannatanmai, sevilay, padth4i|
|code, research||Add apertium-anaphora support to a new language pair||Make a branch of an Apertium language pair that doesn't use apertium-anaphora yet, and call it "anaphora". Manually add the correct antecedent to side ref of the anaphors in the output of biltrans, and modify t1x to change the anaphor based on its antecedent. Verify that it runs and gives the correct anaphor. See the documentation of the apertium-anaphora module for help.||apertium-anaphora, transfer||khannatanmai|
|code, research||Add one markable to the arx file for a language pair||The arx file is where we tell the anaphora resolution algorithm which patterns to detect and score, when we want to find the antecedent of an anaphor. Add one rule in this file which can help the algorithm find out the antecedent of an anaphor - specify the pattern to detect and the positive or negative score you want to give to the noun in this pattern. See the documentation of the apertium-anaphora module for help.||apertium-anaphora, anaphora resolution||khannatanmai|
|code||Write a TextMate grammar for CG-3 syntax highlighting||We want CG-3 files to be syntax highlighted in Github. They use TextMate-compatible grammars. See issue 48 in the CG-3 repo.||cg, editors, tools||Unhammer|
|code||Make Apertium IRC bot's messaging system case-insensitive||Begiak is Apertium's IRC bot. It has a messaging system, where you can say e.g. "begiak, tell randomuser thanks for the tip!" or "begiak: ask randomuser where they filed that issue", and the bot will deliver the message next time it sees randomuser say something. There's been a request for begiak to recognise case-insensitive commands. Your job is to create a fix for this and submit a pull request to the repository with the fix.||python, irc||JNW||yes|
|code||Make Apertium's IRC bot's updating of Apertium's wiki format numbers better||Begiak is Apertium's IRC bot. It has a module that allows users on IRC to trigger a script that updates the Apertium wiki with statistics about Apertium modules. There have been complaints about the formatting of the numbers it writes. Your job is to create a fix for this and submit a pull request to the repository with the fix.||python, irc, wiki||JNW||yes|
|code||Create a new init script for Apertium's IRC bot||Begiak is Apertium IRC bot. It runs on a low-power server that runs Debian. Sometimes that server is reset, and we have to manually restart begiak. The init script that used to control begiak no longer seems to work. Your task is to create a new init script that supports the normal sort of actions that init scripts do, and also the following options for begiak: the specific path to run it from, the user to run it as, miscellaneous arguments, and a log file to log its output. This init script should live in the repository in a reasonable place. There is a github issue describing this task.||python, debian, init, irc||JNW|
|documentation||Add comments to a dictionary defining the symbols used in it||Add comments to a monolingual or bilingual dictionary file (.lexc/.dix, .dix) in the symbol definitions area that clarify what each symbol stands for. Also direct the comment reader to the Apertium wiki page on symbol definitions for more information.||dictionaries||JNW, Flammie||10|
|documentation||find symbols that aren't on the list of symbols page||Go through the symbol definitions in Apertium dictionaries on GitHub (.lexc and .dix format), and document any symbols you don't find on the List of symbols page. This task is fulfilled by adding at least one class of related symbols (e.g., xyz_*) or one major symbol (e.g., abc), along with notes about what it means.||wiki,dictionaries||JNW|
|documentation||document usage of the apertium-separable module||Document which language pairs have included the apertium-separable module in its package, which have beta-tested the lsx module, and which are good candidates for including support for lsx. Add to this wiki page||C++, transducers, lsx||JNW|
|quality||beta testing the apertium-separable||create an lsx dictionary in both directions for any relevant and existing language pair that doesn't yet support it (as a "separable" branch in its GitHub repository), adding 10-30 entries to it in one or both directions. Thoroughly test to make sure the output is as expected. Report bugs/non-supported features and add them to future work. Document your tested language pair by listing it under Lsx_module#Beta_testing and in this wiki page||C++, transducers, lsx||JNW, Fran, Wei En,||yes||yes|
|code, quality||script to test coverage of analyser over corresponding wikipedia corpus||Write a script (in python or ruby) that tests coverage of an Apertium analyser over the latest Wikipedia corpus in that language. One mode of this script should check out a specified language module to a given directory, compile it (or updates it if already existant), and then get the most recently Wikipedia nightly archive for that language and run coverage over it (as much in RAM if possible). In another mode, it should compile the language pair in a docker instance that it then disposes of after successfully running coverage. Scripts exist in Apertium already for finding where a Wikipedia is, extracting a Wikipedia archive into a text file, and running coverage. Ask a mentor for help finding these scripts.||python, ruby, wikipedia||Jonathan, Wei En|
These are languages that can be substituted for AAA and/or BBB for tasks each mentor is listed to mentor above.
|ftyers||eng, spa, cat, fra, nor, rus, por, swe, tur, gag, aze|
|JNW||eng, spa, fra, rus, tur, gag, aze, kaz, kir, kaa, tat, bak, kum, nog, kaa, uzb, uig, crh, khk, yid|
|anakuz||grn, spa, por, rus|
|fotonzade||eng, tur, aze, uig, tat, crh, kmr, ckb, fas|
|xavivars||cat, spa, eng, fra|
|unhammer||nno, nob, swe, dan, fao, sme, ovd|
|shardulc||eng, fra, mar, hin, urd, kan|
|m-alpha||eng, fra, byv|
|sevilay||eng, ara, tur, kaz, aze, tat, gag, uig, uzb, crh, kum|
|eirien||sah, rus, eng|
|flammie||fin, krl, olo, hun, nio, kpv, mdf, tlh, fra, swe, eng, est, ekk, vro|
|dolphingarlic||afr, deu, eng|
|marcriera||cat, spa, eng, ron|
|padth4i||eng, mal, hin|