sanskrit-parser 0.0.1.dev1

Tools for lexical and morphological analysis of Sanskrit

Parsers for Sanskrit / संस्कृतम्

NOTE: This project is still under development. Both over-generation (invalid forms/splits) and under-generation (missing valid forms/splits) are quite likely. Please see the Sanskrit Parser Stack section below for detailed status. Report any issues here.

Please feel free to ping us if you would like to collaborate on this project.

Installation

This project has been tested and developed using Python 2.7. (Python 3 support is in progress)

pip install git+https://github.com/kmadathil/sanskrit_parser/

Usage

Lexical Analyzer

Use the SanskritLexicalAnalyzer to split a sentence and retrieve the top 10 splits:

>>> from sanskrit_parser.lexical_analyzer.SanskritLexicalAnalyzer import SanskritLexicalAnalyzer
>>> analyzer = SanskritLexicalAnalyzer()
>>> splits = analyzer.getSandhiSplits(sentence).findAllPaths(10)
>>> for split in splits:
...    print split
...
[u'asti', u'uttarasyAm', u'diSi', u'devatA', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'devat', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'devata', u'AtmA']
[u'asti', u'uttara', u'syAm', u'diSi', u'devatA', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'devatA', u'at', u'mA']
[u'asti', u'uttarasyAm', u'diSi', u'de', u'vatA', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'devata', u'at', u'mA']
[u'asti', u'uttas', u'rasyAm', u'diSi', u'devat', u'AtmA']
[u'asti', u'uttara', u'syAm', u'diSi', u'devat', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'de', u'avatA', u'AtmA']

The lexical_analyzer can also be used to look up the tags for a given word form in the INRIA database: (Note that the database stores words ending in visarga with an ‘s’ at the end)

>>> word = SanskritObject('hares')
>>> tags = analyzer.getLexicalTags(word)
>>> for tag in tags:
...    print tag
...
('hf#1', set(['cj', 'snd', 'prim', 'para', 'md', 'sys', 'prs', 'v', 'np', 'sg', 'op']))
('hari#1', set(['na', 'mas', 'sg', 'gen']))
('hari#1', set(['na', 'mas', 'abl', 'sg']))
('hari#1', set(['na', 'fem', 'sg', 'gen']))
('hari#1', set(['na', 'fem', 'abl', 'sg']))
('hari#2', set(['na', 'mas', 'sg', 'gen']))
('hari#2', set(['na', 'mas', 'abl', 'sg']))
('hari#2', set(['na', 'fem', 'sg', 'gen']))
('hari#2', set(['na', 'fem', 'abl', 'sg']))

InriaXMLWrapper

The InriaXMLWrapper utility class can also be used to lookup tags:

>>> from sanskrit_parser.util.inriaxmlwrapper import InriaXMLWrapper
>>> db = InriaXMLWrapper()
>>> db_tags = db.get_tags('hares')
>>> tags == db_tags
True

Sandhi

The Sandhi class can be used to join/split words:

>>> from sanskrit_parser.lexical_analyzer.sandhi import Sandhi
>>> sandhi = Sandhi()
>>> word1 = SanskritObject('te')
>>> word2 = SanskritObject('eva')
>>> joins = sandhi.join(word1, word2)
>>> for join in joins:
...    print join
...
teeva
taeva
ta eva
tayeva

To split at a specific position, use the Sandhi.split_at() method:

>>> w = SanskritObject('taeva')
>>> splits = sandhi.split_at(w, 1)
>>> for split in splits:
...    print split
...
(u'tar', u'eva')
(u'tas', u'eva')
(u'taH', u'eva')
(u'ta', u'eva')

To split at all possible locations, use the Sandhi.split_all() method:

>>> splits_all = sandhi.split_all(w)
>>> for split in splits_all:
...    print split
...
(u't', u'aeva')
(u'tar', u'eva')
(u'taev', u'a')
(u'to', u'eva')
(u'ta', u'eva')
(u'te', u'eva')
(u'taH', u'eva')
(u'tae', u'va')
(u'taeva', u'')
(u'tas', u'eva')

Note: As mentioned previously, both over-generation and under-generation are possible with the Sandhi class.

MaheshvaraSutras

Get varnas in a pratyahara:

>>> from sanskrit_parser.base.MaheshvaraSutras import MaheshvaraSutras
>>> MS = MaheshvaraSutras()
>>> jaS = SanskritObject('jaS', encoding=SLP1)
>>> print MS.getPratyahara(jaS)
jabagaqada

Check if a varna is in a pratyahara:

>>> g = SanskritObject('g')
>>> print MS.isInPratyahara(jaS, g)
True
>>> k = SanskritObject('k')
>>> print MS.isInPratyahara(jaS, k)
False

SanskritObject

SanskritObject is a base class used in all modules. It supports automatic detection of input encoding and transcoding to any encoding supported by the indic_transliteration package.

>>> from sanskrit_parser.base.SanskritBase import SanskritObject, SLP1
>>> sentence = SanskritObject("astyuttarasyAMdishidevatAtmA")
>>> print sentence.transcoded(SLP1)
astyuttarasyAMdiSidevatAtmA

Command Line Usage

All the classes described above can also be used from the command line. The corresponding examples are below. Please run the tools with --help/-h to get help on the options

SanskritLexicalAnalyzer

$ python -m sanskrit_parser.lexical_analyzer.SanskritLexicalAnalyzer astyuttarasyAMdishidevatAtmA --split
Splits:
[u'asti', u'uttarasyAm', u'diSi', u'devat', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'devata', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'devatA', u'AtmA']
[u'asti', u'uttara', u'syAm', u'diSi', u'devat', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'devata', u'at', u'mA']
[u'asti', u'uttarasyAm', u'diSi', u'de', u'vatAt', u'mA']
[u'asti', u'uttarasyAm', u'diSi', u'devatA', u'at', u'mA']
[u'asti', u'uttas', u'asyAm', u'diSi', u'devat', u'AtmA']
[u'asti', u'uttara', u'syAm', u'diSi', u'devata', u'AtmA']
[u'asti', u'uttarasyAm', u'diSi', u'de', u'vatA', u'AtmA']

$ python -m sanskrit_parser.lexical_analyzer.SanskritLexicalAnalyzer hares
Input String: hares
Input String in SLP1: hares
[('hf#1', set(['cj', 'snd', 'prim', 'para', 'md', 'sys', 'prs', 'v', 'np', 'sg', 'op'])), ('hari#1', set(['na', 'mas', 'sg', 'gen'])), ('hari#1', set(['na', 'mas', 'abl', 'sg'])), ('hari#1', set(['na', 'fem', 'sg', 'gen'])), ('hari#1', set(['na', 'fem', 'abl', 'sg'])), ('hari#2', set(['na', 'mas', 'sg', 'gen'])), ('hari#2', set(['na', 'mas', 'abl', 'sg'])), ('hari#2', set(['na', 'fem', 'sg', 'gen'])), ('hari#2', set(['na', 'fem', 'abl', 'sg']))]

InriaXMLWrapper

$ python -m sanskrit_parser.util.inriaxmlwrapper hares
INFO:root:Pickle file found, loading at 2017-07-31 14:35:56.093000
INFO:root:Loading finished at 2017-07-31 14:35:59.159000, took 3.066000 s
INFO:root:Cached 666994 forms for fast lookup
Getting tags for hares
('hf#1', set(['cj', 'snd', 'prim', 'para', 'md', 'sys', 'prs', 'v', 'np', 'sg', 'op']))
('hari#1', set(['na', 'mas', 'sg', 'gen']))
('hari#1', set(['na', 'mas', 'abl', 'sg']))
('hari#1', set(['na', 'fem', 'sg', 'gen']))
('hari#1', set(['na', 'fem', 'abl', 'sg']))
('hari#2', set(['na', 'mas', 'sg', 'gen']))
('hari#2', set(['na', 'mas', 'abl', 'sg']))
('hari#2', set(['na', 'fem', 'sg', 'gen']))
('hari#2', set(['na', 'fem', 'abl', 'sg']))

Sandhi

$ python -m sanskrit_parser.lexical_analyzer.sandhi --join te eva
Joining te eva
set([u'teeva', u'taeva', u'ta eva', u'tayeva'])

$ python -m sanskrit_parser.lexical_analyzer.sandhi --split taeva 1
Splitting taeva at 1
set([(u'tar', u'eva'), (u'tas', u'eva'), (u'taH', u'eva'), (u'ta', u'eva')])

$ python -m sanskrit_parser.lexical_analyzer.sandhi --split taeva --all
All possible splits for taeva
set([(u't', u'aeva'), (u'tar', u'eva'), (u'taev', u'a'), (u'to', u'eva'), (u'ta', u'eva'), (u'te', u'eva'), (u'taH', u'eva'), (u'tae', u'va'), (u'taeva', u''), (u'tas', u'eva')])

MaheshvaraSutras

$ python -m sanskrit_parser.base.MaheshvaraSutras --encoding SLP1 --pratyahara jaS
aiuR fxk eoN EOc hayavaraw laR YamaNaRanam JaBaY GaQaDaz jabagaqadaS KaPaCaWaTacawatav kapay Sazasar hal
जश्
जबगडद

$ python -m sanskrit_parser.base.MaheshvaraSutras --encoding SLP1 --pratyahara jaS --varna k
aiuR fxk eoN EOc hayavaraw laR YamaNaRanam JaBaY GaQaDaz jabagaqadaS KaPaCaWaTacawatav kapay Sazasar hal
जश्
जबगडद
Is क् in जश्?
False

$ python -m sanskrit_parser.base.MaheshvaraSutras --encoding SLP1 --pratyahara jaS --varna g
aiuR fxk eoN EOc hayavaraw laR YamaNaRanam JaBaY GaQaDaz jabagaqadaS KaPaCaWaTacawatav kapay Sazasar hal
जश्
जबगडद
Is ग् in जश्?
True

Sanskrit Parser Stack

Stack of parsing tools

Level 0

Sandhi splitting subroutine Input: Phoneme sequence and Phoneme number to split at Action: Perform a sandhi split at given input phoneme number Ouptut: left and right sequences (multiple options will be output). No semantic validation will be performed (up to higher levels)

Current Status

Module that performs sandhi split/join and convenient rule definition is at lexical_analyzer/sandhi.py.

Rule definitions (human readable!) are at lexical_analyzer/sandhi_rules/*.txt

Level 1

• From dhatu + lakAra + puruSha + vachana to pada and vice versa

• From prAtipadika + vibhakti + vachana to pada and vice versa

• Upasarga + dhAtu forms - forward and backwards

• nAmadhAtu forms

• Krt forms - forwards and backwards

• Taddhita forms - forwards and backwards

Current Status

To be done.

However, we have a usable solution with inriaxmlwrapper + Prof. Gerard Huet’s forms database to act as queriable form database. That gives us the bare minimum we need from Level 1, so Level 2 can work.

Level 2

Input

Sanskrit Sentence #### Action * Traverse the sentence, splitting it (or not) at each location to determine all possible valid splits * Traverse from left to right * Using dynamic programming, assemble the results of all choices

To split or not to split at each phoneme

If split, all possible left/right combination of phonemes that can result

Once split, check if the left section is a valid pada (use level 1 tools to pick pada type and tag morphologically)

If left section is valid, proceed to split the right section

• At the end of this step, we will have all possible syntactically valid splits with morphological tags

Output

All semantically valid sandhi split sequences

Current Status

Module that performs sentence split is at lexical_analyzer/SanksritLexicalAnalyzer.py

Level 3

Input

Semantically valid sequence of tagged padas (output of Level 1) #### Action: * Assemble graphs of morphological constraints

viseShaNa - viseShya

karaka/vibhakti

vachana/puruSha constraints on tiGantas and subantas

• Check validity of graphs #### Output

1. Is the input sequence a morphologically valid sentence?

2. Enhanced sequence of tagged padas, with karakas tagged, and a dependency graph associated

Current Status

Not begun

Seq2Seq based Sanskrit Parser

See: Grammar as a Foreign Language : Vinyals & Kaiser et. al. Google http://arxiv.org/abs/1412.7449

• Method: Seq2Seq Neural Network (n? layers)

• Input Embedding with word2vec (optional)

Input

Sanskrit sentence ### Output Sentence split into padas with tags ### Train/Test data DCS corpus, converted by Vishvas Vasuki

Current Status

Not begun