ToUnicode patch corrupts Tibetan text into Latin/ASCII gibberish (GID mismatch)

[10zYmon]

Description

When running pdf-cmap-fix on the document TI1844-01-001.pdf, the script degrades the text extraction rather than fixing it.
Folder name: IE3KG730
File name: TI1844-01-001.pdf
Fonts: BookmanOldStyle, Ededris-a, Ededris-a1, Ededris-b, Ededris-b1, Ededris-vowa, Jomolhari, Kailasa, MonlamUniOuChan2, Cambria, Kailasa, Calibri, #e6#96#87#e9#bc#8e#e7#b2#97#e6#af#9b#e6#a5#

RAW Output:

(RAW): ༄༅། ། ༧ ས་་དམ་པ་ད་བན་ར་་འགས་ད་ན་གས་

PATCHED Output:

(PATCHED): ༄༅# # ( Tú!Ö!f∞!u!çf!£Áq!të!¤!ÎV<ú!≥f!yq!∫<ú!
The script overwrites the existing ToUnicode map with incorrect Latin/ASCII characters, turning the text into complete gibberish.

Sample page for testing :
sample.pdf

Subtasks

• implement gname to unicode mapping if gid to unicode fails
• iimplement hash or glyph curve to unicode mapping , gshape-> unicode if both of above fails.
• using gname find unicode for pua cases.
• do joining on gname and gshape to updated gshape mappings.
• update cli endpoints
• run test
• documentation of repo

[gangagyatso4364]

• failed using font gid mapping. id mismatch.
• subset lacks gid mapping to unicode information. failed.