Question about multi-word headwords

[drdhaval2785]

http://www.sanskrit-lexicon.uni-koeln.de/scans/awork/apidev/servepdf.php?dict=MCI&key=Santanos gives the following entry.

Would it not make better sense to have SantanostIrTam as key1 ? @funderburkjim would like to clarify the stand on multi-word headwords, so that correction submission in such case remains standardised.

[gasyoun]

SantanostIrTam maybe for key2?

[funderburkjim]

key2 should retain the space in a multi-word headword.

In determining what are the headwords in a given dictionary,

• the hw0.py program uses a procedure (usually just a regular expression, in file headword.py) to identify the 'raw' headword from the digitization. In the case of mci.txt, this identification results , for the example above, in the record 456-a+ 37:S4antanos ti1rtham:36158,36158 in hw0.txt.
• Next, it is hw1.py program which 'normalizes' the raw headwords. When the normalization does any adjustment worthy of note, it writes the case out to a log file mcihw1_note.txt. In our example case, the hw1.py program thought it should drop the second word 'ti1rtham'. If, as I think Dhaval is suggesting, this normalization is not proper, then the place to change it is in the hw1.py program. It's not a correction to the data, but a correction to our interpretation of the data. We would probably want hw1.py tp drop the space, not the second word, resulting inS4antanosti1rtham.
• Finally, the last step in generation of the 'key1' field is done by hw2.py, which changes the transliteration of the (Sanskrit) headword to slp1.
• None of the above says anything directly about key2. key2 only comes into play in the context of the generation of the xml form of the dictionary (mci.xml), which is accomplished by the program make_xml.py. This program uses the value in mcihw2.txt for key1. Then, it essentially repeats the logic of hw0.py to identify a corresponding 'key2'. However, I don't want to oversell the efficacy of this part of the existing logic, because it is tricky, and the currently existing programs may not work exactly as I am here describing. But conceptually, at least in the case if MCI, it seems that the form identified by hw0 should be the form appearing in 'key2' field of mci.xml. In particular, it would contain the space.

@drdhaval2785 may want to download pywork, etc. for mci and examine the particulars. Here is the script I've been using to get local versions of the cologne environments, based on s3 backups of those.

[gasyoun]

We would probably want hw1.py tp drop the space, not the second word, resulting inS4antanosti1rtham.

I vote for this change.

In particular, it would contain the space.

That's all that matters.

[drdhaval2785]

@funderburkjim I can correct in MCI accordingly, but issue is not MCI. It is about systemic examination of all dictionaries. Therefore, let us have a cursory look at all such adjustments in case of multi-word headwords and derive at a generic solution, and apply cross dictionaries.

[Andhrabharati]

I would suggest putting the underscore in-place-of the space in these multiple words and have it in the k1 (k2 being with the space as in print.); eliminating the space will loose the 'sense' in the print version.

What do @funderburkjim, @gasyoun and @drdhaval2785 feel about this suggestion?