API Reference¶
Every module, class, and function in src/ is documented here. Each
name is hyperlinked, and every rendered signature carries a [source]
button that jumps directly to the definition in GitHub.
Docstrings are extracted verbatim via sphinx.ext.autodoc; any
inaccuracies in an entry below reflect the state of the docstring in
source. Fixing them is a one-line PR: edit the module and re-push, and
the CI pipeline (see .github/workflows/docs.yml) rebuilds this
site.
Modules¶
The five ofa modules a downstream integrator is most likely to reach for first:
ofa Python client — call ofa --serve from user code. |
|
OnField Assistant (ofa) — locally hosted, RAG-augmented LLM for HPC and scientific-computing workflows. |
|
OpenAI-compatible HTTP server for ofa (BYOK shim for VS Code, etc.). |
|
Site-configuration loader for ofa. |
|
Rebuild one or more ofa RAG collections from |
Supporting modules¶
Ingest / RAG helpers that admins running rebuild_indices.py may want
to inspect, but that are not part of the ofa runtime hot path:
Build the RAG vector index over OpenFOAM tutorials, source headers, and docs. |
|
Build the OnField Assistant RAG index from scratch. |
|
PDF text extraction for RAG ingestion. |