Video Sorter

This repository contains a Python utility that sorts classroom recording files into semester/course folders and can optionally upload them to Kaltura before moving them.

If you are taking the project over, start here:

• AGENTS.md
• docs/DEEP_DIVE.md

What The App Expects

The sorter depends on three local inputs:

1. config.ini in the repo root
2. a schedule spreadsheet matching the expected column names
3. recording filenames that match one of the supported parser formats

For Upload mode it also needs a repo-root .env file with Kaltura credentials.

Setup

1. Create a virtual environment

python3 -m venv .venv

2. Install dependencies

.venv/bin/python -m pip install -r requirements.txt

Note: the current requirements.txt file is UTF-16LE encoded.

3. Create config.ini

Use config-EXAMPLE.ini as a reference, but do not copy the inline comments on the same line as values. With the current Python ConfigParser usage, those inline comments become part of the value.

4. Add .env if using upload mode

kaltura_uploader.py expects:

• PARTNER_ID
• TOKEN
• TOKEN_ID

These values should stay local and never be committed.

5. Run the app

.venv/bin/python video_sorter.py

The app processes immediately on first launch, then continues running and checks again when the local time reaches 3 AM.

Running Tests

Run:

.venv/bin/python -m pytest -q

The tests depend on:

• test_courses.xlsx
• the folder structure configured by [Paths].test_folder

Status verified locally on April 21, 2026:

• 17 tests passed
• 1 test failed because it hard-codes a Windows-style path expectation and does not pass on macOS/Linux

Build To Executable

The repo includes video_sorter.spec.

Build with:

.venv/bin/pyinstaller video_sorter.spec

Schedule Spreadsheet Expectations

The current importer expects columns including:

• Course
• Section #
• Course Title
• Meeting Pattern
• Instructor LAST
• Room (cleaned)
• Instructor

Operational notes:

• Room (cleaned) should contain the bare room number
• Meeting Pattern needs to contain recognizable day/time text
• Instructor is parsed using a strict LAST, FIRST (00123456) pattern

Supported Recording Sources

The active parser pipeline supports:

• Extron filenames with room/date/time information
• Extron 2100 filenames
• CaptureCast filenames with course/section/date information

Details and examples live in docs/DEEP_DIVE.md.

Roadmap / Future Cleanup Ideas

• Support unscheduled/manual recordings that start with one or more uNIDs
• Make config examples safer and more copy-pasteable
• Reduce Windows-only assumptions in tests and docs
• Make spreadsheet and instructor parsing more defensive