Search API

What I’m planning to build

A Python FastAPI service backed by OpenSearch for indexing documents and running full-text queries. It is a stripped-down version of the kind of search layer that shows up behind internal tools, support portals, and log exploration UIs.

Why OpenSearch

I have worked on OpenSearch migrations in production. This project keeps the fundamentals sharp: index mappings, analyzers, bulk indexing, and query DSL — without the baggage of a full product codebase.

What it covers

• /health, /documents, and /search endpoints
• Docker Compose for OpenSearch 2.11
• scripts/seed.py for sample data
• Pytest coverage for request/response models

Connection to my day job

Search infrastructure sits at the intersection of data modeling and ops — shard sizing, refresh intervals, and relevance tuning matter as much as API design. A small demo repo is a safe place to experiment before those decisions hit a shared cluster.

Repo

Source lives in the sandbox monorepo under projects/search-api/.