Amarakosha markup

[funderburkjim]

This issue describes changes to the PWG markup for the 'AK,' (Amarakosha) literary source references.
The main goal is to change the markup so that the AK references will be 'linkable' to some link target.

[funderburkjim]

About 1700 lines are changed. The supporting work is in the ak directory.

• change_1.txt has all the changes.
• lsextract_pwg_0.txt and _1.txt summarize the ls changes made, primarily for AK, but also for a few other sources during error correction.
• classify_1_done.txt summarizes all the various markups, according to parameter analysis, for those cases which are thought to be 'done', in the sense of being readily linkable.
• classify_1_todo.txt summarizes the markups which are, based on form, not readily linkable.
• ak_verse_list.txt lists all the AK parameters (from 'done' list). It further shows whether the parameters are consistent (+) or inconsistent (-) with the numbering in the 1891 version of AK by Colebrooke. It also shows all the verse labels in this 1891 edition and the frequency of occurrence of PWG ls references.

[funderburkjim]

possible link targets

The AK. abbreviation is said to refer to

AMARAKOṢA nach der Ausgabe von COLEBROOKE und LOISELEUR DESLONGCHAMPS (GILD. Bibl. 251. 253).

Up to this point, I have used this 3rd edition of a version by Colebrooke:
https://archive.org/details/ColebrookeAmarakosha1891_201809/
ak_index.txt is based on this edition, as well as the ak_verse_list.txt +- markup.

It seems likely that the AK. references which specifically mention 'COL.' or 'COLEBR' (see classify_1_todo.txt file) are referring to the Colebrooke edition (though no doubt to either the 1st or 2nd edition rather than the available 1891 edition.)

While writing these comments, I found what may be a better version for use as a link target for PWG: https://archive.org/details/amarakochaouvoca01amaruoft/
VOCABULAIRE D'AMARASINHA, A. LOISELEUR DESLONGCHAMPS, PREMIERE PARTIE, 1839

[funderburkjim]

index for LD version

ak_index_ld.txt is like ak_index.txt, but from the LD edition.

ak_verse_ld_list.txt is like ak_verse_list.txt, but uses the LD index.
My expectation is that there would be fewer '-' cases in ak_verse_ld_list (verse references
from PWG digitization to a verse in AK whose verse signature is not found in LD.

511 matches for "^-" in buffer: ak_verse_list.txt
306 matches for "^-" in buffer: ak_verse_ld_list.txt

While there are fewer 'wrong' verse ids using LD, the question remains why there are
still 306 'wrong' verse ids. Only about 30 of these 306 occur in Books 1 and 2. And 234
of those in Book 3 occur in Chapter 4, and 32 occur in chapter 6.

[funderburkjim]

http://amara.aupasana.com/sloka?sloka_number=1.1.53

This website provides another possible link target for Amarakosha.
However, it has some unresolved differences in verse identification when compared to the Colebrooke and Loiseleur editions.

@drdhaval2785 Perhaps you are familiar with this web site due to your work on the stardict dictionaries? Any thoughts on its appropriateness as a link target candidate?

[Andhrabharati]

The AK. abbreviation is said to refer to

AMARAKOṢA nach der Ausgabe von COLEBROOKE und LOISELEUR DESLONGCHAMPS (GILD. Bibl. 251. 253).

It seems likely that the AK. references which specifically mention 'COL.' or 'COLEBR' (see classify_1_todo.txt file) are referring to the Colebrooke edition (though no doubt to either the 1st or 2nd edition rather than the available 1891 edition.)

No doubt, PWG is referring to the 1st ed. of Colebrooke (1808), as it is mentioned to be GILD. Bibl. 251.

And this 1808 ed. is available at Google books (somewhat poor quality scan) and also at the Bavarian State Library https://opacplus.bsb-muenchen.de/title/BV001767120 (very good quality scan).