AI-Cognitive-Leap/agnes-the-ai-analyst
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
agnes-the-ai-analyst/scripts/README.md
Petr c56905d34f Initial commit: OSS data distribution platform 
Open-source AI data analyst platform extracted from internal repo.
Includes data sync engine, Keboola adapter, Flask web portal,
server deployment scripts, and configuration templates.
2026-03-08 23:31:28 +01:00

2 KiB
Raw Blame History

Scripts

Helper scripts for working with Internal AI Data Analyst project.

These scripts are synced from the server into server/scripts/ on the analyst's machine.

Available Scripts

setup_views.sh

Initialize or refresh DuckDB views on Parquet files.

bash server/scripts/setup_views.sh

sync_data.sh

Synchronize data from server, upload user files, and refresh DuckDB.

# Recommended: update scripts first, then sync
rsync -avz data-analyst:server/scripts/ ./server/scripts/   # Linux/macOS
scp -r data-analyst:server/scripts/* ./server/scripts/      # Windows fallback
bash server/scripts/sync_data.sh

# Other options:
bash server/scripts/sync_data.sh --dry-run  # Preview what would be synced (no changes)
bash server/scripts/sync_data.sh --push     # Only upload user/ to server

What sync does:

  1. Self-update check - detects if sync_data.sh changed, asks to re-run if so
  2. Downloads server/docs/, server/scripts/, server/metadata/ from server
  3. Updates CLAUDE.md from latest template
  4. Downloads server/parquet/ data files (with --delete to remove old files)
  5. Uploads user/ directory to server (backup, no --delete)
  6. Syncs Python environment to server
  7. Validates DuckDB - if corrupted, deletes and recreates from parquets
  8. Reinitializes DuckDB views (CREATE OR REPLACE VIEW for all tables)

Self-update mechanism: The script checks its own checksum before and after syncing scripts. If it detects it was updated, it exits with a message asking you to run sync again. This ensures you always run the latest sync logic.

DuckDB corruption recovery: If DuckDB file is corrupted (e.g., interrupted sync), it's automatically detected and recreated. All data is safe in parquet files - DuckDB only contains VIEW definitions.

Typical Workflow

  1. First time setup: Follow bootstrap.yaml instructions
  2. Before analysis: Sync latest data 
    bash server/scripts/sync_data.sh
  3. Analyze: Use DuckDB database at user/duckdb/analytics.duckdb