Projects

A Python library for date-aware currency operations Updated 2025-08-15: v2.x with DM factory, robust fallback for historical rates, and SQLite/PostgreSQL serialization helpers. dated-money represents monetary values with an optional date and performs conversions using historical exchange rates with automatic fallback. https://github.com/juanre/dated-money Installation uv add dated-money or pip install dated-money Quickstart from dated_money import DM, DatedMoney, Currency # Factory: default currency/date Eur = DM('EUR', '2024-01-01') # Create amounts price = Eur(100) # €100.00 from_usd = Eur(50, '$') # $50 → EUR (on 2024-01-01) # Arithmetic: result in the second operand's currency/date a = DatedMoney(100, 'EUR', '2024-01-01') b = DatedMoney(50, 'USD', '2024-01-01') res = a + b # result in USD on 2024-01-01 # Convert and format print(res.to('EUR')) # €... print(str(res)) # e.g. $... (rounded to cents) print(repr(res)) # '2024-01-01 USD ...' (parseable) # Parse from repr parsed = DatedMoney.parse('2024-01-01 EUR 100.50') assert parsed == DatedMoney(100.50, 'EUR', '2024-01-01') Notes: ...

A PostgreSQL async library for modular applications pgdbm is a small, pragmatic library on top of asyncpg that gives you: Connection pooling with retries and backoff Schema-aware SQL templating ({{tables.users}}) for multi-tenant/multi-module apps A simple migration system with per-module isolation and checksums Optional monitoring (slow queries, pool stats, EXPLAIN, health checks) First-class testing utilities (ephemeral DBs, fixtures, isolation) It is designed to be used by standalone services or by libraries that need to bring their own schema and migrations. ...

A Python library for computing with units Updated 2025-01-28: magnitude has been modernized with full type hints, improved error handling, and is available on PyPI. magnitude implements efficient computation with values with units. It allows you to do mathematical operations with them as if they were numbers, taking care of the units behind the scenes. Units are specified as strings. They can be any of the SI units, plus a bunch of non-SI, bits, dollars, and any combination of them. They can include the standard SI prefixes. magnitude can operate with physical quantities, parse their units, and print them. You don’t have to worry about unit consistency or conversions; everything is handled transparently. ...

Superseded by llmring llmbridge has been superseded by llmring, which provides a cleaner architecture without database dependencies. llmring offers: Alias-based routing with lockfiles Registry-driven model discovery No database requirement Better separation of concerns Please see the llmring documentation for the new approach. PyPI: https://pypi.org/project/llmring/

Alias-first LLM routing with lockfiles llmring decouples code from model IDs. Instead of hardcoding “gpt-4o”, you use aliases like “fast” or “smart” that resolve through a lockfile—like package managers pin dependencies. Project resources Main site: https://llmring.ai - Documentation and guides Registry: https://llmring.github.io/registry/ - Live model capabilities and pricing Source: https://github.com/juanre/llmring - Python package Core architecture The system has three independent layers: Lockfile (llmring.lock) - Local alias bindings, version-controlled with your code Registry - Public model database, versioned per provider, human-curated daily Receipts - Cryptographic audit trail (optional, requires server) Your Code → Alias ("fast") → Lockfile → Registry validation → Provider ↓ ↓ ↓ Task identity Git-tracked Versioned truth Why aliases matter Models change. Prices change. New models appear. When you hardcode “gpt-4o-mini” everywhere, updates require code changes. With aliases, you change the lockfile: ...