Problem
The embedding model (e.g., nomic-embed-text for Ollama) is only validated during obsidian-rag setup. At runtime — indexing, searching, watching — the configured model name is passed directly to the API with no pre-flight check. If the model isn't available, the user gets a raw HTTP error with no actionable message.
Expected behavior
Before starting an embedding operation, verify the model is available and surface a clear error like:
Model "nomic-embed-text" is not available in Ollama. Run ollama pull nomic-embed-text to install it.
Where to fix
create_embedder() in src/obsidian_rag/indexer.py — could call get_ollama_models() / get_lmstudio_models() and validate before returning the embedder- Alternatively, each embedder's
embed() could catch the first failure and wrap it in a descriptive error
Relevant code
get_ollama_models(): src/obsidian_rag/indexer.py:477is_ollama_running(): src/obsidian_rag/indexer.py:448create_embedder(): src/obsidian_rag/indexer.py:499
Problem
The embedding model (e.g.,
nomic-embed-textfor Ollama) is only validated duringobsidian-rag setup. At runtime — indexing, searching, watching — the configured model name is passed directly to the API with no pre-flight check. If the model isn't available, the user gets a raw HTTP error with no actionable message.Expected behavior
Before starting an embedding operation, verify the model is available and surface a clear error like:
Where to fix
create_embedder()insrc/obsidian_rag/indexer.py— could callget_ollama_models()/get_lmstudio_models()and validate before returning the embedderembed()could catch the first failure and wrap it in a descriptive errorRelevant code
get_ollama_models():src/obsidian_rag/indexer.py:477is_ollama_running():src/obsidian_rag/indexer.py:448create_embedder():src/obsidian_rag/indexer.py:499