[WIP] Add GitHub Actions CI workflow for mboxer

[Copilot]

• Create .github/workflows/ci.yml with:
  • Triggers on push to master and PRs targeting master
  • ubuntu-latest runner
  • Python 3.11 and 3.12 matrix
  • Steps: checkout → setup Python → install .[dev] → ruff check → mypy → pytest
  • Well-commented, descriptive job/step names for branch protection status checks

Original prompt

Create a GitHub Actions CI workflow at .github/workflows/ci.yml for the uscient/mboxer Python project.

Requirements

The workflow should:

• Trigger on:

  • Every push to master
  • Every pull request targeting master

• Run on ubuntu-latest

• Use a matrix of Python 3.11 and 3.12 (matching the classifiers in pyproject.toml)

• Steps:

  1. Checkout code
  2. Set up Python using the matrix version
  3. Install dependencies: pip install -e ".[dev]"
  4. Run ruff linting: ruff check src/
  5. Run mypy type checking: mypy src/
  6. Run pytest: pytest

Project details

• Package manager: pip with pyproject.toml
• Dev extras group: .[dev] which includes pytest>=8.0, ruff>=0.5.0, mypy>=1.10
• Source is in src/mboxer/
• Tests are in tests/
• ruff line-length is 100, target is py311 (configured in pyproject.toml)
• mypy strict is false (configured in pyproject.toml)
• Python requires >=3.11

Notes

• The job names should be clear and descriptive so they can be referenced as required status checks in branch protection rules
• Keep lint, type-check, and test as separate steps within the same job, or split into separate jobs — whichever is cleaner for status check granularity
• The workflow file should be well-commented

The following is the prior conversation context from the user's chat exploration (may be truncated):

User: Give me a high level overview of this repo

Assistant: ## uscient/mboxer — High-Level Overview

mboxer is a local-first Python CLI tool that transforms Gmail/Google Takeout MBOX exports into structured, AI-ready knowledge assets — primarily targeting NotebookLM, RAG pipelines, and local LLM workflows.

---

🎯 Core Purpose

Gmail MBOX exports are raw and unwieldy. mboxer bridges that gap by converting them into:

Output	Purpose
Markdown source packs	Organized by category/year for NotebookLM upload
SQLite database	Local, durable index for all messages, threads, classifications
JSONL exports	RAG pipelines, embeddings, local LLM tools
CSV exports (planned)	Spreadsheet review and auditing

---

⚙️ Pipeline Stages

The full workflow is a sequential CLI pipeline:

ingest → normalize → classify → security-scan → scrub → review → export

1. Ingest — Parses .mbox files using Python's mailbox stdlib, with resumable checkpoints and deduplication
2. Classify — Rule-based categorization (sender, domain, subject keywords) at message or thread level; LLM via Ollama is config-wired but not yet fully connected
3. Review — Interactive CLI to approve/reject proposed categories
4. Security scan — Flags sensitive content; configurable redaction of phone numbers, SSNs, credit cards, etc.
5. Export — Outputs Markdown, JSONL, and manifests; five NotebookLM limit profiles (standard → ultra_safe)

---

🗂️ Source Layout

src/mboxer/
  cli.py            # argparse CLI entry point
  ingest.py         # MBOX ingest pipeline
  normalize.py      # message normalization
  classify.py       # rule-based classification
  taxonomy.py       # category management
  db/               # SQLite schema + init helpers
  exporters/        # notebooklm, jsonl, manifest writers
  security/         # scan, scrub, policy
config/             # annotated YAML config example
tests/              # 14 test modules covering all pipeline stages

---

🔑 Key Design Principles

• Local-first & privacy-conscious — nothing leaves your machine by default
• Resumable — large MBOX ingests can be interrupted and continued
• Multi-account — supports multiple Gmail accounts in one project
• Inspectable — dry-run export, category review, and local SQLite before any cloud upload
• NotebookLM-optimized — exports respect source limits with named profiles

---

📊 Implementation Status

The core pipeline is complete and working. Still in progress:

• CSV export
• Full LLM classification via Ollama
• Local web UI for category review
• Incremental export tracking
• PII scrub profiles

Tech stack: Pure Python (≥3.11), SQLite, YAML config, pyproject.toml packaging with optional pdf and dev extras. Licensed under MIT.

User: Evaluate the Repo settings

Assistant: Here's a full evaluation of the repository settings for uscient/mboxer:

---

✅ What's Configured Well

Setting	Status	Notes
License	✅ MIT	Clearly declared
Description	✅ Set	Clear and descriptive
Topics	✅ 14 tags	Well-tagged (gmail, rag, notebooklm, mbox, etc.)
Issues enabled	✅ On	Good for tracking work
Dependabot	✅ Configured	Weekly pip updates, grouped by prod/dev, with commit prefix
Default branch named	✅ master	Consistent and in use

---

⚠️ Areas of Concern

🔴 High Priority

| Issue | Detail | Recommendation |
|---|---|--...

This pull request was created from Copilot chat.