Introduction
“”ஓர் ஆயிரம் கழித்தல் ஐந்து பெருக்கல் (ஒன்பது கூட்டல் ஒன்று)” என்பதை எப்படி 950 என்று கணினி புரிந்து கொள்ளும் ? Open-Tamil வழியாகத்தான் – வாங்க இதை பார்க்கலாம்.
Open-Tamil is a free and open-source Tamil Text / Natural Language Processing library (NLP) written in Python, and developed since early 2010s. This blog post is cover our gaps in marketing (on our part, ofcourse) Open-Tamil regarding the capabilities and features presented by Open-Tamil python library. Open-Tamil Python package is distributed under MIT license.
Documentation
Open-Tamil project is currently in the release v0.96 – the Python package can be found here. Documentation from Sphinx for Open-Tamil package can be found on Tamilpesu.us site at this URL.
Installation
Open-Tamil can be installed by running the following command from your terminal with root privileges on Python3. You cannot use current/later versions of Open-Tamil with v2.x – please refer to older releases on pypi page which are still available to you.
$ pip install –upgrade open-tamil>=0.96
If you want to install locally, you can run this command instead,
$ pip install –upgrade open-tamil>=0.96 –user
Python Modules
Open-Tamil provides the following Python modules with various classess and functionalities. You can learn more about them from browsing our documentation.
In summary we can list them as follows,
Calculator Example
One fun use of the API is to build interactive websites like tamilpesu.us but a more interesting use is to build a text based calculator taking inputs in form of Tamil numerals with Tamil words for arithmetic operations like add, subtract, multiply and divide.
“”ஓர் ஆயிரம் கழித்தல் ஐந்து பெருக்கல் (ஒன்பது கூட்டல் ஒன்று)” என்பதை எப்படி 950 என்று கணினி புரிந்து கொள்ளும் ?
Using Open-Tamil we can parse the numerals back to numbers and form a simple mathematical infix expression;
1000 – 5 * ( 9 + 1 )
then use Python eval() we can use Python parser and evaluator to build the expression tree and compute the post-order traversal and execute the expression to receive the result, which in our case is not surprisingly,
950.
Well, all this can be done in less than 30 lines of code using Open-Tamil; if you don’t believe me check-out the code here.
Credits
A large number of individuals have contributed their time and effort over the years in many capacities to the project; should you wish to acknowledge this project in your research, products you can choose to cite any of our published works of relevance to you; our publications are listed here.
As always if you wish to report bugs, or share your comments please write an email to me ezhillang AT gmail DOT com or file an issue at github.