flow_inference.write_inference_to_raw_xml.InferenceToRawXMLWriter

class flow_inference.write_inference_to_raw_xml.InferenceToRawXMLWriter(raw_xml_repo, inference_repo, token, allow_source_repo_update=False)[source]

Write inference output into raw XML records and upload the result.

The writer downloads a raw XML dataset and an inference result dataset from the Hugging Face Hub, builds a lookup from line identifiers to inferred text, inserts the text into matching XML records, and writes the updated data to a new or existing target dataset repository.

Parameters:

  • raw_xml_repo (str)

  • inference_repo (str)

  • token (str)

  • allow_source_repo_update (bool)

__init__(raw_xml_repo, inference_repo, token, allow_source_repo_update=False)[source]

Initialize the raw XML writeback pipeline.

Parameters:

  • raw_xml_repo (str) – Source Hugging Face dataset repository containing raw XML records.

  • inference_repo (str) – Source Hugging Face dataset repository containing inference results.

  • token (str) – Hugging Face token used for downloading and uploading datasets.

  • allow_source_repo_update (bool) – Whether writing back into the raw XML source repository is allowed.

Methods

__init__(raw_xml_repo, inference_repo, token)

Initialize the raw XML writeback pipeline.

process_and_upload([output_repo, ...])

Run the raw XML writeback pipeline and upload the result.