API documentation for loader
module.
Classes
PostgresDocumentSaver
PostgresDocumentSaver(
key: object,
engine: langchain_google_cloud_sql_pg.engine.PostgresEngine,
table_name: str,
content_column: str,
schema_name: str = "public",
metadata_columns: typing.List[str] = [],
metadata_json_column: typing.Optional[str] = None,
)
A class for saving langchain documents into a PostgreSQL database table.
PostgresLoader
PostgresLoader(
key: object,
engine: langchain_google_cloud_sql_pg.engine.PostgresEngine,
query: str,
content_columns: typing.List[str],
metadata_columns: typing.List[str],
formatter: typing.Callable,
metadata_json_column: typing.Optional[str] = None,
)
Load documents from PostgreSQL`.
Each document represents one row of the result. The content_columns
are
written into the content_columns
of the document. The metadata_columns
are written
into the metadata_columns
of the document. By default, first columns is written into
the page_content
and everything else into the metadata
.
Modules Functions
_parse_doc_from_row
_parse_doc_from_row(content_columns: typing.Iterable[str], metadata_columns: typing.Iterable[str], row: dict, metadata_json_column: typing.Optional[str] = 'langchain_metadata', formatter: typing.Callable = <function text_formatter>) -> langchain_core.documents.base.Document
Parse row into document.
_parse_row_from_doc
_parse_row_from_doc(
doc: langchain_core.documents.base.Document,
column_names: typing.Iterable[str],
content_column: str = "page_content",
metadata_json_column: typing.Optional[str] = "langchain_metadata",
) -> typing.Dict
Parse document into a dictionary of rows.
csv_formatter
csv_formatter(row: dict, content_columns: typing.List[str]) -> str
CSV document formatter.
json_formatter
json_formatter(row: dict, content_columns: typing.List[str]) -> str
JSON document formatter.
text_formatter
text_formatter(row: dict, content_columns: typing.List[str]) -> str
txt document formatter.
yaml_formatter
yaml_formatter(row: dict, content_columns: typing.List[str]) -> str
YAML document formatter.