def extract_di_words()

in function_app/src/result_enrichment/doc_intelligence.py [0:0]

• 15 lines of code
• 3 McCabe index (conditional complexity)

def extract_di_words(di_result: AnalyzeResult) -> List[EnrichedDocumentWord]:
    """
    Extract all words from a Document Intelligence response, enriching them.

    :param raw_azure_doc_response:
        The raw Document Intelligence response for a single document.

    :return:
        A list of all words in the document.
    """
    di_words = list()
    for page in di_result.pages:
        for word in page.words:
            out_word = EnrichedDocumentWord(
                normalized_polygon=scale_flat_poly_list(
                    word.polygon,
                    existing_scale=(page.width, page.height),
                    new_scale=(1, 1),
                ),
                page_number=page.page_number,
                **word,
            )
            di_words.append(out_word)
    return di_words