MCPcopy Index your code
hub / github.com/feast-dev/feast / load_parse_and_chunk

Method load_parse_and_chunk

sdk/python/feast/chunker.py:30–54  ·  view source on GitHub ↗

Load, parse, and chunk a document. Args: source: File path, raw text, bytes, etc. source_id: Document identifier. source_type: Optional type hint. source_column: The column containing the document sources. Returns:

(
        self,
        source: Any,
        source_id: str,
        source_column: str,
        source_type: Optional[str] = None,
    )

Source from the content-addressed store, hash-verified

28
29 @abstractmethod
30 def load_parse_and_chunk(
31 self,
32 source: Any,
33 source_id: str,
34 source_column: str,
35 source_type: Optional[str] = None,
36 ) -> list[dict]:
37 """
38 Load, parse, and chunk a document.
39
40 Args:
41 source: File path, raw text, bytes, etc.
42 source_id: Document identifier.
43 source_type: Optional type hint.
44 source_column: The column containing the document sources.
45
46 Returns:
47 List of chunk dicts with keys:
48 - chunk_id: str
49 - original_id: str
50 - text: str
51 - chunk_index: int
52 - (any additional metadata)
53 """
54 pass
55
56 def chunk_dataframe(
57 self,

Callers 1

chunk_dataframeMethod · 0.95

Calls

no outgoing calls

Tested by

no test coverage detected