OCR0036: Find remaining glyphs of complex ligatures

[10kalden]

Description: Around 300 glyphs from OPF text and 100 glyphs or complex ligatures are yet to be found. We have methods and different approaches to extract all the glyphs required.

Implementation: There are two approaches we have developed for this,

1. Use the already obtained individual glyphs, stack them according to the ligature and apply morph and augmentation to obtain the best possible result.
2. Use annotators to draw the missing glyphs as closely as the glyphs from the publication and scan them.

Subtask:

• [x] write a script to join the glyphs and apply morph to form a ligature

• [x] parse the derge tengyur opf

Completion Criteria: To obtain the remaining missing glyphs

[kaldan007]

Please reach out to gen Lopa on 100 missing glyph.

[10kalden]

• We already have glyphs for Tibetan consonants. Gen Lopa says we need Photoshop to align the glyphs to form the ligatures.
• Currently exploring ways to do this programmatically, we can break down a complex ligature according to the occurrence of individual glyphs unicode, and then we can take the glyphs and join them together, this would require some augmentation to the joined glyphs since individual glyphs would be of different sizes.
• this is the individual glyph we have

[kaldan007]

please update the card with new approach.

[10kalden]

ligatures obtained from joining glyphs

[kaldan007]

Kindly scan https://github.com/OpenPecha-Data/P000002 for the missing glphys. Use all the glyphs u font and generate a font.

[10kalden]

The OPF data for P000002 is incomplete, there are no image references for where the char occurs.

[kaldan007]

Please check commit history of the opf

[10kalden]

Found 73 glyphs from Derge tengyur

[kaldan007]

we are left with 23 missing glyphs.