scienceskillscommon

What problem does it solve?

It provides a shared, dependable HTTP client layer for science skills, so API-heavy workflows can handle rate limits, retries, and backoff consistently.

Core Features & Use Cases

• Unified HTTP client: One entry-point for fetching JSON, text, and bytes across scientific web APIs.
• Cross-process rate limiting: Coordinates QPS using a shared file lock to avoid burst traffic.
• Automatic retries with backoff: Retries transient HTTP statuses (429, 5xx) and network errors using exponential backoff and optional jitter.
• Server-directed throttling: Honors Retry-After and interprets X-Throttling-Control to proactively slow down (e.g., PubChem/NCBI).
• Practical use: Skills importing it can reliably query endpoints like NCBI E-utilities and other rate-limited services without re-implementing HTTP robustness.

Quick Start

Use it from another science skill by importing the shared client and calling fetch_json or fetch_bytes for the target endpoint.