Some errors about AI.

[c-rei]

I wnat to inspect the details abount the ai_agent, so I get the AI agent default configuration with this:

a2 = client.ai.get_ai_agent_default_config(GetAiAgentDefaultConfigMode.EXTRACT_STRUCTURED, language="ja-JP")

And I tried to use tihs configration with this:

ans4 = client.ai.create_ai_extract_structured(
... [AiItemBase(id="................")],
... fields=[field5, field6],
... ai_agent=a2,
... )

but I got some errors: (When I do not use ai_agent=a2, I can get the collect result.)

Traceback (most recent call last):
File "", line 1, in
File "/Users/jscuser01/.pyenv/versions/3.10.16/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/box_sdk_gen/managers/ai.py", line 453, in create_ai_extract_structured
response: FetchResponse = self.network_session.network_client.fetch(
File "/Users/jscuser01/.pyenv/versions/3.10.16/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/box_sdk_gen/networking/box_network_client.py", line 134, in fetch
self._raise_on_unsuccessful_request(
File "/Users/jscuser01/.pyenv/versions/3.10.16/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/box_sdk_gen/networking/box_network_client.py", line 267, in _raise_on_unsuccessful_request
raise BoxAPIError(
box_sdk_gen.box.errors.BoxAPIError:
Timestamp: 2025-06-10 14:58:27.791832
Underlying error: None
Message: 400 [{"keyword":"enum","dataPath":".long_text.embeddings.model","schemaPath":"#/properties/long_text/properties/embeddings/properties/model/enum","params":{"allowedValues":["azure__openai__text_embedding_ada_002"]},"message":"should be equal to one of the allowed values"}]; Request ID:
Request:
Method: POST
URL: https://api.box.com/2.0/ai/extract_structured
Query params:
{}
Headers:
{ 'Authorization': '---[redacted]---',
'Content-Type': 'application/json',
'User-Agent': 'box-python-generated-sdk-1.14.0',
'X-Box-UA': 'agent=box-python-generated-sdk/1.14.0; env=python/3.10.16'}
Body:
('{"items": [{"id": "1887265902551", "type": "file"}], "fields": [{"key": '
'"\u6587\u66f8\u306e\u7a2e\u985e", "description": '
'"\u6587\u66f8\u306e\u5206\u985e", "displayName": '
'"\u6587\u66f8\u306e\u7a2e\u985e", "type": "string"}, {"key": '
'"\u5951\u7d04\u95a2\u4fc2\u4f1a\u793e", "description": '
'"\u5951\u7d04\u53cc\u65b9", "displayName": '
'"\u5951\u7d04\u4f1a\u793e", "type": "string"}], "ai_agent": '
'{"basic_image": {"model": "azure__openai__gpt_4_1_mini", "system_message": '
'"Respond only in valid json. You are extracting metadata that is name, value '
'pairs from a document. Only output the metadata in valid json form, as '
'{\"name1\":\"value1\",\"name2\":\"value2\"} and nothing else. You '
'will be given the document data and the schema for the metadata, that '
'defines the name, description and type of each of the fields you will be '
'extracting. Schema is of the form {\"fields\": [{\"key\": '
'\"key_name\", \"prompt\": \"prompt to extract the value\", \"type\": '
'\"date\"}]}. Leverage prompt for each key to identify where the key and '
'value pairs are in the document. In certain cases, prompt can also indicate '
'the instructions to perform on the document to obtain the value. Prompt will '
'be in the form of Schema is schema \n document is document", '
'"prompt_template": "If you need to know today's date to respond, it is '
'{current_date}. Schema is {user_question} \n document is '
'{content}", "num_tokens_for_completion": 8400, '
'"llm_endpoint_params": {"temperature": 0, "top_p": 1, "frequency_penalty": '
'0.75, "presence_penalty": 0.75, "stop": "<|im_end|>", "type": '
'"openai_params"}}, "type": "ai_agent_extract_structured", "long_text": '
'{"model": "google__gemini_2_0_flash_001", "num_tokens_for_completion": 4096, '
'"llm_endpoint_params": {"type": "google_params", "temperature": 0, "top_p": '
'1, "top_k": 1}, "system_message": "Respond only in valid json. You are '
'extracting metadata that is name, value pairs from a document. Only output '
'the metadata in valid json form, as '
'{\"name1\":\"value1\",\"name2\":\"value2\"} and nothing else. You '
'will be given the document data and the schema for the metadata, that '
'defines the name, description and type of each of the fields you will be '
'extracting. Schema is of the form {\"fields\": [{\"key\": '
'\"key_name\", \"prompt\": \"prompt to extract the value\", \"type\": '
'\"date\"}]}. Leverage prompt for each key to identify where the key and '
'value pairs are in the document. In certain cases, prompt can also indicate '
'the instructions to perform on the document to obtain the value. Prompt will '
'be in the form of Schema is schema \n document is document", '
'"prompt_template": "If you need to know today's date to respond, it is '
'{current_date}. Schema is {user_question} \n document is '
'{content}", "embeddings": {"model": "google__text_embedding_005", '
'"strategy": {"id": "basic", "num_tokens_per_chunk": 128}}}, "basic_text": '
'{"model": "google__gemini_2_0_flash_001", "num_tokens_for_completion": 4096, '
'"llm_endpoint_params": {"type": "google_params", "temperature": 0, "top_p": '
'1, "top_k": 1}, "system_message": "Respond only in valid json. You are '
'extracting metadata that is name, value pairs from a document. Only output '
'the metadata in valid json form, as '
'{\"name1\":\"value1\",\"name2\":\"value2\"} and nothing else. You '
'will be given the document data and the schema for the metadata, that '
'defines the name, description and type of each of the fields you will be '
'extracting. Schema is of the form {\"fields\": [{\"key\": '
'\"key_name\", \"prompt\": \"prompt to extract the value\", \"type\": '
'\"date\"}]}. Leverage prompt for each key to identify where the key and '
'value pairs are in the document. In certain cases, prompt can also indicate '
'the instructions to perform on the document to obtain the value. Prompt will '
'be in the form of Schema is schema \n document is document", '
'"prompt_template": "If you need to know today's date to respond, it is '
'{current_date}. Schema is {user_question} \n document is '
'{content}"}}}')
Response:
Status code: 400
Headers:
{ 'Alt-Svc': 'h3=":443"; ma=2592000,h3-29=":443"; ma=2592000',
'Transfer-Encoding': 'chunked',
'access-control-expose-headers': 'Server-Timing',
'box-request-id': '08eb03a203b7740bbcd67d916857d4610',
'cache-control': 'no-cache',
'content-type': 'application/json; charset=utf-8',
'date': 'Tue, 10 Jun 2025 05:58:27 GMT',
'server-timing': 'traceparent;desc="00-8f6281ababb00b01e028ed87babf7f43-681d4affb1b5f88b-00"',
'set-cookie': 'csrf-token=OqX0kRQSnF7KNq9fSPc4e0RUTXQGNtVLaxMMBwV32Lg; '
'Secure; SameSite=None; Path=/',
'strict-transport-security': 'max-age=31536000',
'via': '1.1 google',
'x-content-type-options': 'nosniff',
'x-download-options': 'noopen',
'x-envoy-upstream-service-time': '85',
'x-frame-options': 'SAMEORIGIN',
'x-xss-protection': '0'}
Code: bad_request
Context Info:
{}
Request Id:
Help Url: https://developer.box.com/guides/api-calls/permissions-and-errors/common-errors/
Body:
{ 'code': 'bad_request',
'help_url': 'https://developer.box.com/guides/api-calls/permissions-and-errors/common-errors/',
'message': '[{"keyword":"enum","dataPath":".long_text.embeddings.model","schemaPath":"#/properties/long_text/properties/embeddings/properties/model/enum","params":{"allowedValues":["azure__openai__text_embedding_ada_002"]},"message":"should '
'be equal to one of the allowed values"}]',
'request_id': '',
'status': 400,
'type': 'error'}

Versions Used

Python SDK: 3.10.16
Python: 1.14.0

Thans for any help.

[congminh1254]

Hi @c-rei

We are sorry for the late reply, this issue got lost by somehow.

This issue seem like the AI team deprecated some AI model but not updating the model in the default agent endpoint. We will contact the AI team to resolve this issue as soon as possible.

Thanks for your question.

Bests,
Minh

[congminh1254]

Hi @c-rei

I confirmed with AI team and it's by design.
So in this case, you should provide the AI Agent by yourself with the models in allowed list returned from the API.

Bests,
Minh

[c-rei]

Hi, @congminh1254
Thanks for your responses, but I don't know how to set an AI Agent with the model.
I'll try it again.

Bests,
Rei

[c-rei]

I'v known that I can get an AI agent for structured metadata extraction that is an AiAgentExtractStructured class with below code.

a2 = client.ai.get_ai_agent_default_config(GetAiAgentDefaultConfigMode.EXTRACT_STRUCTURED, language="ja-JP")

So, when I want to do structured metadata extraction, I could use below code.

client.ai.create_ai_extract_structured(
[AiItemBase(id=file.id)],
fields=[....],
ai_agent=AiAgentExtractStructured,
)

And by the document, ai_agent just is an AiAgentExtractStructured. Isn't it the return value above?

Bests,
Rei

[arjankowski]

Hi @c-rei,
As @congminh1254 mentioned, the problem is with the AI agent and especially with the long_text.embeddings.model field in it. In that case, I suggest overriding the returned AI agent config in such a way that you omit the embeddings field in AiAgentLongTextTool.

You can try using the sample below:

    ai_extract_structured_agent_config: AiAgentExtractStructured = client.ai.get_ai_agent_default_config(GetAiAgentDefaultConfigMode.EXTRACT_STRUCTURED, language='ja-JP')

    long_text_config_with_no_embeddings: AiAgentLongTextTool = AiAgentLongTextTool(
        system_message=ai_extract_structured_agent_config.long_text.system_message,
        prompt_template=ai_extract_structured_agent_config.long_text.prompt_template,
        model=ai_extract_structured_agent_config.long_text.model,
        num_tokens_for_completion=ai_extract_structured_agent_config.long_text.num_tokens_for_completion,
        llm_endpoint_params=ai_extract_structured_agent_config.long_text.llm_endpoint_params
    )

    agent_ignoring_overriding_embeddings_model: AiAgentExtractStructured = AiAgentExtractStructured(
        basic_text=ai_extract_structured_agent_config.basic_text,
        long_text=long_text_config_with_no_embeddings
    )

    response: AiExtractStructuredResponse = client.ai.create_ai_extract_structured(
        [AiItemBase(id=file.id)],
        fields=[....],
        ai_agent=agent_ignoring_overriding_embeddings_model
    )

Let us know if that helped
Regards,
Artur

[c-rei]

Hi @arjankowski,
With your presented code, I got the correct response.

Thank you very much.
Regards,
Rei