[INDOLOGY] Extension of Sanskrit Heritage tools with connection to Amba Kulkarni's parser

[Gérard Huet]

Dear indologists,

This is to announce an upgrade of the Sanskrit Heritage site <https://sanskrit.inria.fr/>, connecting its reader assistant segmenter/tagger to Pr Kulkarni’s dependency parser. It is now possible to invoke her parser from the Heritage reader interface, whenever the number of remaining solutions is not too high. This facility is also available from its corpus manager.

Here is an example. Search for entry kāka <> in the dictionary. You will see on its entry page <https://sanskrit.inria.fr/DICO/20.html#kaaka> a reference to the citation kākaḥ kṣīraṃ pītvā kokilaḥ na bhavati <https://sanskrit.inria.fr/cgi-bin/SKT/sktgraph.cgi?corpmode=reader&text=kaaka.h+k.siiram+piitvaa+kokila.h+na+bhavati&t=VH&cpts=0%2C%7BNouc%7D%2C%7Bkaakas%7D%2C%7Bt%7D%7C30%2C%7BRoot%7D%2C%7Bbhavati%7D%2C%7Bt%7D&corpdir=Heritage_citations&sentno=158>. If you click on it, you wind up in the Sanskrit Segmenter interface listing the relevant segmentation. Now a new action is proposed, mentioning UoH Analysis Mode. Click on its green tick sign. This will transport you to an interface page listing the padapāṭha. Click on the Submit button. You are now facing the saṃsādhanī analyser, listing its possible parses. Selecting the root pā in the pītvā column gives the intended parse, which may now be displayed as a dependency graph like this <https://sanskrit.inria.fr/cgi-bin/scl/SHMT/prog/interface/show_selected_parse.cgi?filename=/private/tmp/SKT_TEMP/./tmp_in14680/&sentnum=1&start=0&outscript=IAST>. 

This facility is still in an experimental phase. It works for not too complex sentences given in prose order. For poetry, it is possible to use the padapāṭha interface to reorder the components by giving manually the prose order. 

Many thanks to Pr Kulkarni for providing this wonderful grammatical analysis tool, and for her assistance in adapting it to the Sanskrit Heritage Web services.

Gérard Huet




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://list.indology.info/pipermail/indology/attachments/20190118/4a58f497/attachment.htm>