canvas-ledger¶

A local, queryable ledger of Canvas LMS metadata for historical analysis.

Why canvas-ledger?¶

Canvas LMS is effective for answering questions about the current state of a course. But it's poor at answering historical questions:

What courses was I involved in three years ago, and in what capacity?
When did a student drop a course, and what was their status before dropping?
Who was the lead instructor for a co-taught course?
How has a student's performance changed across semesters?

canvas-ledger gives instructors a durable, local, queryable historical record of Canvas metadata—independent of Canvas's retention policies, UI limitations, and API constraints.

Key Features¶

Local-first: Single SQLite database you own and control
Historical accuracy: Tracks changes over time (adds, drops, grade changes)
Dual-truth model: Preserves Canvas data exactly as reported, while allowing your annotations
CLI-first: Scriptable, composable commands with JSON/CSV output
Incremental ingestion: Safe to run repeatedly without duplicating data

Quick Example¶

# Configure and initialize
cl config init
cl db migrate

# Ingest your Canvas courses
cl ingest catalog

# See your involvement timeline
cl query my-timeline

# Deep ingest a specific course
cl ingest offering 12345

# Query the roster
cl query offering 12345

Getting Started¶

See the Quickstart guide to get up and running.

Documentation¶

Concepts — Understand the dual-truth model and design principles
Data Model — Database schema, file locations, and what gets tracked
Quickstart — Install, configure, and run your first queries
Workflows — Common usage patterns by time of semester
CLI Reference — Complete command documentation