Convert Gsheet slides to the new txt format

[gpsaggese]

• Read the doc about the documentation toolchain in //helpers/docs/tools/documentation_toolchain
• Read the code in ./dev_scripts_helpers/documentation/
• Check out the new format of the slides using the new markdown -> Latex -> PDF flow which converts txt files from
  https://github.com/gpsaggese/umd_msml610/tree/master/lectures_source

into

https://github.com/gpsaggese/umd_msml610/tree/master/lectures

with a command like:

notes_to_pdf.py --input lectures_source/Lesson01-Intro.txt --output tmp.pdf --type slides --skip_action cleanup_after --debug_on_error --toc_type navigation

• Convert slides from Gsheet https://drive.google.com/drive/folders/1RvfyFIqOSQxs_APEFeGbLjSM3_irQ3lz into the new format

• Make sure the flow works on your set up (this is 90% of the complexity)

  • We should add //helpers to the data605 repo and make it a runnable dir like https://github.com/gpsaggese/umd_msml610/tree/master (but I want to keep data605 github as simple as possible)

• I would start with the first lecture and see what happens

• I would use LLMs as much as possible (e.g., download the Gsheet as markdown and then ask LLMs to convert into txt files)

  • I don't think the images can be handled automatically. I would save them as png and then add them in the txt file

Let's do baby steps. We have 2-3 months to convert all the slides so there is time.

[protocorn]

Testing Documentation Toolchain - Issues Encountered

I've been testing the documentation toolchain on my device and encountered the following issues:

LaTeX Compilation Failure

Issue: When running notes_to_pdf.py on files with numbered lists (like Lesson01-Intro.txt), getting "TeX capacity exceeded" error with recursive \labelenumi definitions.

Root Cause: The enumitem package in latex_abbrevs.sty conflicts with pandoc's \tightlist command when processing numbered lists. Pandoc converts numbered lists (1. 2. 3.) to LaTeX enumerate environments, but \tightlist inside these environments causes recursive errors with enumitem.

Solution: Removed the enumitem package configuration entirely from latex_abbrevs.sty and let pandoc handle list environments naturally. This resolves the compilation failure for files containing numbered lists.

Color Commands Issue

Issue: Color commands like \blue{}, \red{}, \green{} are not rendering properly - only seeing maroon/magenta color, while other colors appear black.

Status: Currently trying to fix this by adding proper color command definitions to latex_abbrevs.sty.

Current Status

• Lesson00-Class.txt - Works (no numbered lists)
• Lesson01-Intro.txt - Now compiles after LaTeX fix, images render correctly
• Color rendering - Work in progress

The toolchain is functional after these fixes. Will update once color issue is resolved.

Note : I noticed that in /umd_msml610/tree/master/lectures existing Lesson01-Intro PDF, the images appear to be missing (other lecture PDFs have images correctly).

@gpsaggese @HarshitGadge

[gpsaggese]

There must be some "misconfiguration" (e.g., maybe some code from my branch that is not in master), since I run the flow without these errors.

Let's coordinate on TG and / or through a PP

In the meanwhile I would keep going, even if things are not rendered perfectly, e.g., commenting and adding TODOs.

The goal is to always make progress putting patches and bandaids.

[gpsaggese]

• First step do a PR with everything you have converted
• Ignore the pieces that are failing (comment them out)