insta2table 0.1.0

Fetch Instagram captions (with optional OCR) and convert them to a normalized table via Gemini.

insta2table

A tiny toolkit to:

1. Crawl Instagram links from a text file, extract captions and (optionally) OCR text from images → output.csv
2. Convert those rows into a clean single-row table per link using Gemini → result.csv

Credentials & keys via env:

• IG_USER, IG_PASS (optional): for Instaloader login (reduces 403s)
• GOOGLE_API_KEY: required for Gemini

Quickstart

# 1) (Recommended) Create and activate a virtualenv
python -m venv .venv
# Linux/macOS
source .venv/bin/activate
# Windows (PowerShell)
# .venv\Scripts\Activate.ps1

# 2) Install (with OCR & Gemini extras if you need them)
pip install -e .[ocr,genai]

# 3) Prepare links
cp examples/links.txt .

# 4) Crawl -> output.csv
export IG_USER="your_user"
export IG_PASS="your_pass"
insta2csv --links links.txt --out output.csv

# 5) Process with Gemini -> result.csv
export GOOGLE_API_KEY="your_key"
insta2table --in output.csv --out result.csv

CLI

insta2csv --links links.txt --out output.csv [--no-ocr]
insta2table --in output.csv --out result.csv

Notes

• OCR requires Tesseract installed on your system if you opt in.
• Instagram scraping without login can trigger 403s. Supplying IG credentials helps.
• Gemini formatting expects a single-row Markdown table per input.

Publishing

We use trusted publishing from GitHub to PyPI (no API token needed).

1. Create the project on PyPI (only first time) and enable 'Manage publishing' with GitHub OIDC for your repo.
2. In GitHub: create a new Release with a tag like v0.1.0.
3. The Publish to PyPI workflow will build and upload the release automatically.
4. Alternatively (manual): make build then twine upload dist/* (requires a PyPI token).