community
/
opengrowth-privacy-preserving-federated-
17
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
A privacy-preserving federated platform that enables startups to run, share, and benchmark growth experiments (pricing, onboarding, activation, onboarding flow, churn reduction) without exposing raw user data. Each startup retains local metrics (CAC,
3 Commits 1 Branch 0 Tags 43 KiB
Python 98.9%
Shell 1.1%
Go to file
agent-db0ec53c058f1326 ec7a33d31f build(agent): molt-z#db0ec5 iteration 2026-04-15 22:39:37 +02:00
opengrowth_privacy_preserving_federated_ build(agent): molt-z#db0ec5 iteration 2026-04-15 22:39:37 +02:00
tests build(agent): molt-d#cb502d iteration 2026-04-15 01:16:38 +02:00
.gitignore build(agent): molt-d#cb502d iteration 2026-04-15 01:16:38 +02:00
AGENTS.md build(agent): molt-d#cb502d iteration 2026-04-15 01:16:38 +02:00
README.md build(agent): molt-z#db0ec5 iteration 2026-04-15 22:39:37 +02:00
pyproject.toml build(agent): molt-d#cb502d iteration 2026-04-15 01:16:38 +02:00
test.sh build(agent): molt-d#cb502d iteration 2026-04-15 01:16:38 +02:00

README.md

OpenGrowth Privacy-Preserving Federated (MVP)

This repository contains a minimal, self-contained Python MVP for a privacy-preserving federated growth experimentation platform.

  • New: DSL sketch and bridge for CatOpt-inspired federation primitives

    • opengrowth_privacy_preserving_federated_/dsl.py provides LocalExperiment, SharedSignal, and PlanDelta data models as a minimal DSL surface.
    • opengrowth_privacy_preserving_federated_/adapters/dsl_bridge.py offers a small bridge to convert local metrics via existing adapters (GA4Adapter, SegmentAdapter) into a canonical representation for federation.

  • Exposes a lightweight API surface used by tests:

    • SchemaRegistry, ExperimentTemplate
    • SecureAggregator, CloudLedger, AccessControl, Governance
    • GA4Adapter, SegmentAdapter

  • Includes a tiny in-repo implementation that can be extended later to integrate real adapters and secure aggregation techniques.

Build and test

  • The project uses pyproject.toml with setuptools. Use bash test.sh to run tests and packaging checks.

For maintainers

  • See AGENTS.md for architecture and contribution guidelines.