| name | fit-process |
| description | Build the knowledge indexes from HTML sources: process resources, then graphs and vectors. Use when populating the retrieval indexes an agent queries, as a build-pipeline step run in order before search or graph queries work.
|
Build the Knowledge Indexes
fit-process is the write surface that turns HTML knowledge sources into the
indexes fit-rag queries. One binary, three subcommands run in order:
resources parses HTML into typed resources, graphs derives an RDF graph from
them, and vectors embeds them into a vector index.
When to Use
- Ingest HTML sources into typed resources —
fit-process resources
- Derive the graph index from resources —
fit-process graphs
- Embed resources into the vector index —
fit-process vectors
Usage
fit-process resources --base https://example.invalid/
fit-process graphs
fit-process vectors
resources accepts --base (-b) to set the base URI for generated
identifiers (default https://example.invalid/). resources and graphs run
offline; vectors calls the embedding service. Run resources before graphs
and vectors, which both read the resource index it writes. Once the indexes
are built, query them with fit-rag.