Module ollama_harness 

Optional Ollama HTTP harness for inference when the native Qualia GGUF path is unavailable or the principal explicitly selects Ollama.

Not the primary engine. Qualia remains Local GGUF / wgpu in-process. This harness is opt-in so chat, ETL scaffolding, and later CML/logic gates can still run against a reachable Ollama (or OpenAI-compatible) endpoint while native inference is brought up.

Wire format: Ollama REST (/api/tags, /api/generate, /api/chat). Network I/O is confined to this module; callers pass prompts already augmented by Qualia retrieval / ontology routing.

Structs

OllamaGenerateResult

One completed generation (non-streaming).

OllamaHarness

OllamaModelInfo

OllamaStatus

Operator-visible probe of a configured Ollama endpoint.

Constants

DEFAULT_OLLAMA_BASE_URL

Default local Ollama base URL (override via settings or OLLAMA_HOST).

Functions

probe_configured_ollama

Convenience: probe using currently persisted settings.

probe_configured_ollama_async