Data Classes

All public methods on LocalLens return one of these dataclasses. Each has a to_dict() method for JSON serialization.

from locallens import IndexResult, SearchResult, AskResult, StatsResult, FileInfo, DoctorCheck

IndexResult

Returned by LocalLens.index().

Field	Type	Description
total_files	int	Total files in index
new_files	int	Newly indexed files
updated_files	int	Re-indexed files
skipped_files	int	Skipped (unchanged) files
total_chunks	int	Total chunks in index
duration_seconds	float	Time taken

{
  "total_files": 42,
  "new_files": 5,
  "updated_files": 2,
  "skipped_files": 35,
  "total_chunks": 1284,
  "duration_seconds": 3.47
}

SearchResult

Returned by LocalLens.search() (as a list).

Field	Type	Description
file_path	str	Absolute path to the file
file_name	str	File name only
file_type	str	File extension (e.g. .pdf)
chunk_text	str	The matched text chunk
chunk_index	int	Chunk position within the file
score	float	Relevance score
extractor	str | None	Extractor that produced this chunk

{
  "file_path": "/Users/me/Documents/report.pdf",
  "file_name": "report.pdf",
  "file_type": ".pdf",
  "chunk_text": "Q3 revenue exceeded projections by 12%...",
  "chunk_index": 3,
  "score": 0.8742,
  "extractor": "pymupdf"
}

AskResult

Returned by LocalLens.ask().

Field	Type	Description
answer	str	The generated answer
sources	list[SearchResult]	Source chunks used for context
model	str	Ollama model used
duration_seconds	float	Time taken

{
  "answer": "Q3 revenue was $4.2M, exceeding projections by 12%.",
  "sources": [
    {
      "file_path": "/Users/me/Documents/report.pdf",
      "file_name": "report.pdf",
      "file_type": ".pdf",
      "chunk_text": "Q3 revenue exceeded projections...",
      "chunk_index": 3,
      "score": 0.8742,
      "extractor": "pymupdf"
    }
  ],
  "model": "qwen2.5:3b",
  "duration_seconds": 5.23
}

AskStreamEvent

Yielded by LocalLens.ask_stream().

Field	Type	Description
event_type	str	"token" or "sources"
data	Any	Token string or list of SearchResult

When event_type is "token", data is a string (one token of the answer). When event_type is "sources", data is a list[SearchResult] with the context chunks.

StatsResult

Returned by LocalLens.stats().

Field	Type	Description
total_files	int	Number of indexed files
total_chunks	int	Number of chunks
file_types	dict[str, int]	Chunk count per file type
collection_name	str	Qdrant collection name
data_dir	str	Data directory path

{
  "total_files": 42,
  "total_chunks": 1284,
  "file_types": {
    ".pdf": 450,
    ".py": 320,
    ".md": 280,
    ".txt": 234
  },
  "collection_name": "locallens",
  "data_dir": "/Users/me/.locallens"
}

FileInfo

Returned by LocalLens.files() (as a list).

Field	Type	Description
file_path	str	Absolute path
file_name	str	File name
file_type	str	File extension
chunk_count	int	Number of chunks
indexed_at	str | None	ISO timestamp

{
  "file_path": "/Users/me/Documents/report.pdf",
  "file_name": "report.pdf",
  "file_type": ".pdf",
  "chunk_count": 15,
  "indexed_at": "2024-12-15T10:30:00Z"
}

DoctorCheck

Returned by LocalLens.doctor() (as a list).

Field	Type	Description
name	str	Check name
status	str	"ok", "fail", or "warn"
message	str	Detail message

{
  "name": "Qdrant Edge",
  "status": "ok",
  "message": "1284 points in shard"
}