Skip to content

Releases: eduardocornelsen/full-funnel-ai-analytics

Release list

v0.9.1 — Beta

v0.9.1 — Beta Pre-release
Pre-release

Choose a tag to compare

View all tags
@eduardocornelsen eduardocornelsen released this 21 May 20:23
v0.9.1
ae85e2a
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

CI/CD validated end-to-end with real Olist-anchored data

What's included

Production Readiness & CI/CD

  • 4 GitHub Actions workflows: PR gate (DuckDB, no cloud creds), warehouse deploy to BigQuery + Snowflake in parallel, daily scheduled refresh, daily synthetic data append
  • Warehouse adapters: uniform DuckDB / BigQuery / Snowflake connection layer (scripts/_warehouse_adapters.py)
  • validate_metrics.py — re-queries warehouse and diffs against golden_metrics.json; exits 1 on any drift
  • generate_golden_metrics.py — --target and --live flags for multi-warehouse support

dbt MetricFlow Fixes

  • dbt_project.yml vars block for window dates and time spine bounds
  • fct_pipeline.sql rewritten to deal-level UNION of HubSpot + Salesforce
  • fct_marketing_attribution.sql — order_date added to GROUP BY for time-window filtering
  • metricflow_time_spine.sql — hardcoded dates replaced with var() references
  • metrics.yml — fill_nulls_with: 0 and NULLIF on all ratio denominators

Golden Layer Foundation

  • CLAUDE.md §9 and §14: mandatory golden-layer-first data sourcing rules for AI agents
  • Mock server date filtering for HubSpot and Salesforce

Streamlit Data Sources UI

  • File Upload tab: drag-and-drop CSV/Excel to replace any mock table
  • Warehouse Connection tab: test and save DuckDB / BigQuery / Snowflake connections
  • MCP Server Status tab: live row counts and latest date for all 6 servers

Daily Synthetic Data

  • scripts/daily_synthetic_append.py — date-seeded reproducible data generation with seasonal factors and day-of-week adjustments

Test Suite

  • 20+ pytest assertions on golden metrics structure, metric sanity, funnel ordering, and attribution percentages
  • FastAPI lead scoring endpoint tests

Documentation

  • docs/architecture.md — full medallion architecture and MetricFlow internals
  • docs/guides/production_readiness_guide.md — CI/CD setup and GitHub secrets
  • docs/guides/data_import_guide.md — import data via UI, CSV, or live warehouse
  • docs/guides/connector_ui_guide.md — live dashboard capabilities by platform
  • Updated README with CI badges, architecture v2, and full project structure

GitHub Secrets needed for warehouse deploys

GCP_PROJECT_ID, GCP_SERVICE_ACCOUNT_KEY_JSON, SNOWFLAKE_ACCOUNT, SNOWFLAKE_USER, SNOWFLAKE_PASSWORD, SNOWFLAKE_WAREHOUSE, SNOWFLAKE_DATABASE, SNOWFLAKE_SCHEMA

What's Changed

  • fix: CI validates real Olist data + standalone fallback + data architecture docs by @eduardocornelsen in #5

Full Changelog: v0.9.0...v0.9.1

Contributors

eduardocornelsen
Loading