csl-inflect status

[funderburkjim]

This csl-inflect repository provides a replacement for the 'MW inflected form' link on the
sanskrit-lexicon home page.

• old link : https://www.sanskrit-lexicon.uni-koeln.de/work/fflexphp/web/index.php
• new link: https://www.sanskrit-lexicon.uni-koeln.de/scans/csl-inflect/web/index.php

The csl-homepage repository has been changed to use the new link.

csl-inflect is also an optional module to the local installations.

The repository contains code required to recompute conjugations and declensions, and to
make sqlite files used by the display. Once csl-inflect is cloned to an installation, the
'redo.sh' script recomputes everything required.

As of now, here are the known limitations:

• declensions: stems_problem contains examples
  believed to provide incorrect, or at least unverified, declensions. These declensions are
  not currently part of the display.
• conjugations: Currently only a few conjugations are computed:
  • present, imperfect, imperative, and optative conjugations of verbs of classes 1,4,6 and 10,
    for active and middle voices as gleaned from MW data.
  • present, imperfect, imperative, and optative conjugations in passive voice for verbs of all
    classes
  • restriction to about 700 roots, those which MW classifies as 'genuine'.

My intention is to remove many of these limitations in the next couple of weeks.

[funderburkjim]

@drdhaval2785 I think csl-inflect is the last existing component of the sanskrit-lexicon that needed to
be converted to repository form. Also, I think all these 'csl-xxx' repository components are now
in a stable shape. In other words, I think we now have a 'minimal plan', as we were thinking in the discussions with @artforlife (sanskrit-lexicon/COLOGNE#261). These are the
10 repositories:

csl-orig
csl-westergaard
csl-pywork
csl-doc
csl-kale
csl-inflect
csl-whitroot
csl-apidev
csl-homepage
csl-websanlexicon

Also, the scanned images are in repositories in the
sanskrit-lexicon-scans Github organization.

Have I missed anything? or is the minimal plan now satisfied?

[gasyoun]

For 9 days I checked every day and now it comes, it's huge.

declensions: stems_problem contains examples believed to provide incorrect, or at least unverified, declensions

As of 2019 I guess there is no tool that has verified forms computed anywhere. Please remind me of the link where you compared Huet, Dhavala and Amba's code.

or is the minimal plan now satisfied?

Not that I see now, thanks, Jim. Let's plan for a call on Sunday with @drdhaval2785, @artforlife and @YevgenJohn ?

[funderburkjim]

tool that has verified forms

An interesting idea to make the comparisons themselves in a form accessible to others.

Amba's code

I have done no comparisons involving code from Amba. Is there a link to such code?

Not that I see now

Are you saying that the minimal plan is NOT satisfied? If not satisfied, what is missing?

call on Sunday.

I don't mind participating, but will let you and others lead the discussion. On Sundays, time of day would need to be after Noon , New York time.

[gasyoun]

comparisons themselves in a form accessible to others.

Indeed.

Is there a link to such code?

Let me inquire.

minimal plan is NOT satisfied?

It's done as far as I see it.