GitPKM

GitPKM is a plain-text Personal Knowledge Management system that combines:

• Markdown for narrative notes
• CSV for relational data
• Python scripts for automation
• Git/GitHub for history and sync

The project is model-first:

1. Define the data model (entities, relationships, ID rules).
2. Define the workflow (generation, indexing, validation).
3. Apply the same model and workflow to any repository.

Start Here

• Quickstart
• CLI
• Data-Model-and-IDs
• Schema-Contract
• Automation-and-Validation
• Automation config: see schema/automation.json
• Example automation config: see schema/automation.example.json
• Import mappings: see schema/import_mappings
• Publishing-Safety
• Operational-Conventions

Documentation Structure

• User-facing docs are in the project wiki: gitpkm wiki
• Internal AI context and full design notes live in _instructions.md
• Release notes are tracked in CHANGELOG.md
• License is defined in LICENSE

Core Principles

• Plain text first: Markdown + CSV
• Stable descriptive IDs as the source of truth
• Exact dataset names are part of the contract; the tooling does not singularize or pluralize them
• Separation of concerns: structured data vs narrative notes
• Safe automation using generated markers to avoid overwriting manual text

Repository Layout

data/        # CSV tables (source of truth)
notes/       # Markdown entity notes and indexes
scripts/     # generation, indexing, validation scripts
schema/      # schema documentation
notebooks/   # optional analysis
pkm.wiki/    # user-facing documentation