nv_ingest_api.internal.extract.pdf package

Subpackages

• nv_ingest_api.internal.extract.pdf.engines package
  • Subpackages
    • nv_ingest_api.internal.extract.pdf.engines.pdf_helpers package
      • Module contents
  • Submodules
  • nv_ingest_api.internal.extract.pdf.engines.adobe module
    • adobe_extractor()
  • nv_ingest_api.internal.extract.pdf.engines.llama module
    • async_llama_parse()
    • llama_parse_extractor()
  • nv_ingest_api.internal.extract.pdf.engines.nemoretriever module
    • nemoretriever_parse_extractor()
  • nv_ingest_api.internal.extract.pdf.engines.pdfium module
    • pdfium_extractor()
  • nv_ingest_api.internal.extract.pdf.engines.tika module
    • tika_extractor()
  • nv_ingest_api.internal.extract.pdf.engines.unstructured_io module
    • unstructured_io_extractor()
  • Module contents
    • adobe_extractor()
    • llama_parse_extractor()
    • nemoretriever_parse_extractor()
    • pdfium_extractor()
    • tika_extractor()
    • unstructured_io_extractor()

Submodules

nv_ingest_api.internal.extract.pdf.pdf_extractor module

nv_ingest_api.internal.extract.pdf.pdf_extractor.extract_primitives_from_pdf_internal(

df_extraction_ledger: DataFrame,

task_config: Dict[str, Any],

extractor_config: Any,

execution_trace_log: List[Any] | None = None,

) → Tuple[DataFrame, Dict]

Process a DataFrame of PDF documents by orchestrating extraction for each row.

This function applies the row-level orchestration function to every row in the DataFrame, aggregates the results, and returns a new DataFrame with the extracted data along with any trace information.

Parameters:

• df_extraction_ledger (pd.DataFrame) – A pandas DataFrame containing PDF documents. Must include a ‘content’ column with base64-encoded PDF data.

• task_config (dict) – A dictionary of configuration parameters. Expected to include ‘task_properties’ and ‘validated_config’ keys.

• extractor_config (Any) – A dictionary of configuration parameters for the extraction process.

• execution_trace_log (list, optional) – A list for accumulating trace information during extraction. Defaults to None.

Returns:

A tuple where the first element is a DataFrame with the extracted data (with columns: document_type, metadata, uuid) and the second element is a dictionary containing trace information.

Return type:

tuple of (pd.DataFrame, dict)

Raises:

Exception – If an error occurs during the extraction process on any row.

Module contents