verbdata_map: mapping from verbdata to MW

[funderburkjim]

In order to compare inflected forms generated by SanskritVerb to those generated by pysanskrit, it is necessary to represent a SanskritVerb root with a Monier-Williams root. The section under _verbdatamap in the pysanskrit/dhaval/roots/readme.md describes the problem.

It also describes the first part of what seems a practical solution to the mapping, namely that presented in verbdata_map. With the correspondence of this file, we can transform the extracted SanskritVerb present tense conjugations conj_pre to a form that can be directly compared to a Pysanskrit generated conjugation.

As an example, the conj_pre conjugation table

aki! pre 01.0092A:[aNkate aNkete aNkante aNkase aNkeTe aNkaDve aNke aNkAvahe aNkAmahe]

can, in light of the verbdata_map correspondence aki!:01:0092:aNk be transformed to

aNk pre 01A:[aNkate aNkete aNkante aNkase aNkeTe aNkaDve aNke aNkAvahe aNkAmahe]

Now the header information aNk pre 01A is sufficient for Pysanskrit to generate its own conjugation table, which can then be compared SanskritVerb table.

[funderburkjim]

Because of other work comparing the root-class-pada information from verbdata of SanskritVerb to that from MW (this work is described in others of these issues), we expect that there will need to be adjustments to the mapping presented in sanverb_map.txt.

In fact, we already know of one such adjustment that is needed, namely that involving the class 10 roots whose SanskritVerb spelling ends in a. The relevant correspondence is in roots_a.txt.

We modify sanverb_map.txt using roots_a.txt , the result being in sanverb_map1.txt. This is further described in the sanverb_map1 section of the documentation for pysanskrit1/dhaval/roots.