fraiseql-confiture 0.8.20

PostgreSQL schema evolution with built-in multi-agent coordination 🍓

Confiture 🍓

PostgreSQL migrations, sweetly done.

Build fresh databases in <1 second. Apply incremental migrations. Sync production data with PII anonymization. Zero-downtime schema swaps via FDW. All from the same tool.

---

Why Confiture?

• DDL is the source of truth — your db/schema/ directory defines the database, not a chain of migrations. Fresh databases build in <1 second by executing DDL directly.
• Timestamp-based versioning — migration filenames use YYYYMMDDHHMMSS, so parallel branches never collide. No more 003_add_users merge conflicts.
• 4 strategies, one tool — build from DDL, incremental migrations, production sync with anonymization, and zero-downtime schema-to-schema via FDW. Pick the right one for each situation.
• Multi-agent coordination — built-in intent registration and conflict detection so teams and AI agents don't step on each other's schema changes.
• CI/CD-native — semantic exit codes, structured output (JSON/CSV/YAML), distributed locking, and dry-run with SAVEPOINT testing.

---

Quick Start

Installation

pip install fraiseql-confiture

# Recommended: include pglast for large-schema support
# (uses PostgreSQL's own C parser — no token limits)
pip install "fraiseql-confiture[ast]"

CLI

# Initialize project structure
confiture init

# Write your schema DDL files
vim db/schema/10_tables/users.sql

# Build database from DDL (<1 second)
confiture build --env local

# Generate and apply migrations
confiture migrate generate --name "add_bio"
confiture migrate up

Library API

from confiture import Migrator

with Migrator.from_config("db/environments/prod.yaml") as m:
    status = m.status()
    if status.has_pending:
        result = m.up()

---

The Four Strategies

Strategy	Use Case	Command
Build from DDL	Fresh databases, testing, CI	confiture build --env local
Incremental Migrations	Existing databases, production	confiture migrate up
Production Sync	Copy data with PII anonymization	confiture sync --from prod --anonymize users.email
Zero-Downtime	Complex migrations via FDW	confiture migrate schema-to-schema

---

Features

Migration Management

• migrate preflight — pre-deploy safety gate: checks reversibility, non-transactional statements, duplicate versions, and checksum tampering. JSON output for CI/CD.
• Semantic exit codes — 0 success, 2 validation error, 3 SQL failure, 6 lock contention. Script with confidence.
• migrate rebuild — drop and recreate from DDL + replay migrations in one command. Fast environment reset.
• migrate validate — naming convention checks, schema drift detection, function signature and body drift.
• migrate fix-signatures — detect and atomically fix stale function overloads and body drift.
• Dry-run with SAVEPOINT testing — --dry-run-execute runs migrations inside a savepoint, then rolls back.
• Checksum verification — detect tampered migration files before applying.
• Distributed locking — safe concurrent deployments via PostgreSQL advisory locks.
• Migration hooks — run custom logic before/after each migration.

Schema Intelligence

• Schema diff detection — two-tier parser: pglast (PostgreSQL's C parser) primary, sqlparse fallback.
• Schema linting — configurable rules to catch common DDL mistakes.
• Function introspection — FunctionIntrospector, TypeMapper, and DependencyGraph for deep schema analysis.
• Grant accompaniment checker — detect permission changes without corresponding migrations.

Seed Data

• Sequential execution — handles PostgreSQL parser limits on large seed files.
• Per-file savepoint isolation — one bad seed file doesn't ruin the batch.
• 5-level prep-seed validation — static analysis through full execution, pre-commit safe at levels 1-3.

Multi-Agent Coordination

• Intent registration — declare which tables you're changing before you start.
• Conflict detection — automatic alerts when agents touch overlapping tables.
• JSON output — machine-readable for CI/CD pipelines.

Developer Experience

• Structured output — JSON, CSV, and YAML for all commands.
• Exception hierarchy — typed errors with error codes and resolution hints.
• Git-aware validation — detect schema drift vs. main branch, enforce migrations for DDL changes.
• PII anonymization — built-in strategies for production sync.
• Optional Rust extension — drop-in performance boost for SQL parsing and hashing.
• Python 3.11, 3.12, 3.13 — tested across all supported versions.
• 4,480+ tests passing.

---

Documentation

Getting Started: docs/getting-started.md

Guides:

• Build from DDL
• Incremental Migrations
• Production Data Sync
• Zero-Downtime Migrations
• Sequential Seed Execution
• Multi-Agent Coordination
• Prep-Seed Validation
• Migration Decision Tree
• Dry-Run Mode

CLI Reference: docs/reference/cli.md

API Reference: docs/reference/

Examples: examples/

---

Contributing

git clone https://github.com/fraiseql/confiture.git
cd confiture
uv sync --all-extras
uv run pytest

See CONTRIBUTING.md and CLAUDE.md.

---

Author & License

Vibe-engineered by Lionel Hamayon 🍓

MIT License - Copyright (c) 2025 Lionel Hamayon

---

Making jam from strawberries, one migration at a time. 🍓→🍯