drdhaval2785
commented
8 years ago http://sanskrit-lexicon.github.io/CORRECTIONS/abnormending/abnorm.html is the output.
https://github.com/sanskrit-lexicon/CORRECTIONS/tree/master/abnormending is the code.
Execution
Run this shell file to regenerate the results.
Logic -
- The last two letters of each word in sanhw1.txt is stored as 'endings'.
- The words in sanhw1.txt are checked and the count of endings are shown e.g.
(2, u'hE')i.e. the words ending with 'hE' are only 2. See this for full list. - Only endings having less than 50 entries (thereby meaning less frequent ones) are kept.
- This list is sorted in ascending order (1,2,3.....50).
- sanhw1.txt is checked for three criteria (1) words ending in this sorted list of point 4. (2) the word should be seen only in one dictionary and (3) the word should not be seen in nochange.
To put in regex terms,
if re.search(end+':[^,]*$',datum) and datum not in noc. - Words passing the above mentioned criteria are stored in
abnorm.txt. - Webpage and PDF are linked by
link.phpand stored inabnorm.html. abnorm.htmlis put on github.io for potential errors and submit corrections.
gasyoun
As desired by @gasyoun the following is list of headword endings which are found in <50 headwords in sanhw1.txt. Not sorted anyhow. Just a rough list to play with.
How to use ?
177 was found by the same method.