README.md

A number of components for an English to Kinyarwanda translator(Focus - rules +English tagging)

current supported features:

1. Detecting kinyarwanda_ness of txt by using ratio consonant combinations that confirm to kinyarwanda consonant constraints(ibihekane)
2. Autogenerating numbers wording in order of 10^15(Kinyarwanda), and English(10^6)
   • test kinyarwanda number2word at http://g.havugimana.com//count/1010000, change the numeric value and see the output
3. Autoconjugating (present, past, and future) in kinyarwanda ---template()
4. Plural formation(kinyarwanda) --still in development(it can be used as noun class detector with some tweak)

Currently, I am doing

1. English tense and verb detector (on simple sentences)
2. webscrawling +webscraping tool
3. studying IBM models(for SMT)

for testing features (use,edit)

testing_features.py

• edit testing_features by uncommenting and commenting

• print(function_number(input))

  • print(thousands(100)) ===> one hundred

• Error?

  • make sure you have imported all neccessary code
    • from code.number2words import *, inport everything in number2words(including bugs if any) +install any missing lib +BS4 +NLTK(if used)

• No need to install for testing some features:

  • http://g.havugimana.com/count/224
    • works for numbers and return wording for a number(this case 224)
  • <http://g.havugimana.com/ubwinshi/umunyeshuri + return the plural form of word icyuki + currently in development + You can look at the code and suggest tweaks

• word list

  • http:///www.havugimana.com/rw_word_list.txt

• command line calling of functions(while in the main folder)

  • python main.py function(word) ### no space between function and words
    • python main.py million(34007)
  • python main.py function word
    • python main.py million 34007
  • python main.py "[(x, tiliyoni(x), million(x) )for x in range(10000)]"
  • etc --- you can suggest the best way to build an interface(or push one)