g-raffy / pymusco Goto Github PK

Latex is good, but what pymusco actually uses it for is very minimal (to generate a pdf with a table of content, and to overlay text on pdfs). If pymusco could use other ways to do this, it would make pymusco much lighter in its requirements.

there are still instrument names in pymusco to make ts-auto work. The goals of this task is:

• to write unit tests in CI to check the behavious of ts-auto against known typical examples
• to remove hardcoded instruments in pymusco
• to clarify (or get rid of) the is_rare status of instruments.

At the moment, blanck pages in scan pdfs are declared as if they were a fake track (with a disabled keyword). As track ids are dictionary keys, they can't have the same name. As a result, we have to use a meaningless name to ensure there's no duplicate keys.

As an example, I had a scan of eb clarinet which had 2 blank pages : one at the beginning of the eb clarinet, and one at its end. Because blank pages are treated like disabled tracks, I've named them 'eb clarinet 10 and eb clarinet 11', as they need to also obey track naming conventions. This is suboptimal to say the least...

the current test.py uses samples that are not included in the project. Find a small copyright free sample and include it in the project so that the test is fully usable to anyone.

sometimes, one page of the resulting stub pdf contains a corrupted image (the bottom part contains wring pixel values).

The input images are not corrupted though. The corruption happens during the pdflatex part. Double check that the input images are completely written at the moment they are used as input