PERSONAL RAG TOOLKIT — REVIEWED SOURCE EXCERPTS
Reviewed: 29 October 2026
Source: rag_system/core/query_engine.py
Independent project. Source inspection, not a runtime demonstration.
Two excerpts; indentation normalized; intervening code omitted.

Request relevant passages from a selected collection (lines 78–86)
```python
query_embedding = self.embedding_generator.generate_embeddings([query])[0]

# Query vector store
results = self.vector_store.query(
    collection_name=domain,
    query_embeddings=[query_embedding],
    n_results=top_k,
    where=metadata_filter,
)
```

Keep the passage and its source metadata together (lines 115–123)
```python
search_results.append(
    SearchResult(
        content=results["documents"][0][i],
        score=similarity,
        metadata=metadata,
        document_id=results["ids"][0][i],
        source_file=metadata.get("file_path", "unknown"),
    )
)
```

These excerpts show collection selection, a top-k request, an optional metadata filter and source metadata on results. They are not a complete runnable example.
Metadata filtering is not an access-control system. Source metadata does not prove an answer is correct.
No provider calls, ingestion or end-to-end retrieval were run for this review. No accuracy, latency, savings or production-readiness result is claimed.