itemsuch

itemsuch is a software library that provides fast and flexible search over collections of items. It is designed for catalogs, inventories, and datasets where items have metadata such as name, category, price, tags, and ratings. The project emphasizes modular indexing, configurable ranking, and support for multiple data sources.

Key features include an inverted index for text fields, support for exact and partial matches, and configurable fuzzy matching. It also offers facet-based filtering, range queries for numeric and date fields, and pluggable backends for in-memory or disk-backed indexes. Ranking combines query relevance with field weights and item metadata.

The architecture centers on modular components: a core search engine, indexers that convert records into index structures, a query parser, and a scorer. Data connectors allow ingestion from databases, JSON stores, or remote APIs. The index can reside in memory for small datasets or on disk for larger collections, and may be distributed or sharded for scalability.

Typical usage involves defining an item schema, feeding records to an indexer, and building the index. Users submit queries such as name:chair AND category:furniture or price:[20 TO 100], optionally requesting sorted results by relevance or price. Queries yield items with scores and associated metadata.

Performance considerations include indexing time, memory footprint, and update latency. The system supports batch and real-time indexing, caching, and configurable refresh intervals. Scaling is achieved through sharding, replication, or multiple indices handling separate data domains.

Applications span e-commerce catalogs, library inventories, asset management, and game item databases.