How to convert json into data type that is supported by Azure Form Recognizer

Question

I'd like to convert a json into the data type that is supported by Azure Form Recognizer. I'm able to convert the data type into a dic and then into a json but I'm not able to do the opposite without analysing once again the document. How could I use the data type supported by Azure Form Recognizer without having to analyse the document more than one time?

Here is what I have.

endpoint = "endpoint"
key = "key"


# create your `DocumentAnalysisClient` instance and `AzureKeyCredential` variable
document_analysis_client = DocumentAnalysisClient(endpoint=endpoint, credential=AzureKeyCredential(key))

# Extract text from doc using "prebuilt-document"
with open("file.pdf", "rb") as f:
    poller = document_analysis_client.begin_analyze_document(
            "prebuilt-document", document=f)
result = poller.result()

import json
form_pages = poller.result()
d = form_pages.to_dict()
json_string = json.dumps(d)
print(json_string)
data = json.loads(json_string)
poller1 = form_pages.from_dict(data)

Answer 1

What's the scenario for converting the JSON model representation back to the SDK model? The operations don't take the result models as an input, as a solution the original result model be stored somewhere until it needs to be used again in that case.

Also, for converting the model to JSON, it would be better to use the AzureJSONEncoder to that the SDK can properly serialize all types. For example:

from azure.core.serialization import AzureJSONEncoder

# save the dictionary as JSON content in a JSON file, use the AzureJSONEncoder
# to help make types, such as dates, JSON serializable
# NOTE: AzureJSONEncoder is only available with azure.core>=1.18.0.
with open('data.json', 'w') as f:
    json.dump(analyze_result_dict, f, cls=AzureJSONEncoder)

Here is a link to the full sync sample: https://github.com/Azure/azure-sdk-for-python/blob/main/sdk/formrecognizer/azure-ai-formrecognizer/samples/v3.2/sample_convert_to_and_from_dict.py