40 VARCHAR Limit on Model Name

[mbbyn]

Self Checks

• This is only for bug report, if you would like to ask a question, please head to Discussions.
• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• Please do not modify this template :) and fill in all the required fields.

Dify version

0.6.14

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

First, add an OpenAI API compatible model with a long name.

Then, try to use the model in a Knowledge Base settings. Observe the failure and error logs.

✔️ Expected Behavior

It should work with long model names.

❌ Actual Behavior

It fails to save the model name since it's >40 VARCHAR.

sqlalchemy.exc.DataError: (psycopg2.errors.StringDataRightTruncation) value too long for type character varying(40)
dify_api.1
all values are hidden
dify_api.1
[SQL: INSERT INTO dataset_collection_bindings (provider_name, model_name, type, collection_name) VALUES (%(provider_name)s, %(model_name)s, %(type)s, %(collection_name)s) RETURNING dataset_collection_bindings.id, dataset_collection_bindings.created_at]
dify_api.1
[parameters: {'provider_name': 'openai_api_compatible', 'model_name': 'sentence-transformers/distiluse-base-multilingual-cased-v1', 'type': 'dataset', 'collection_name': 'Vector_index_df1864b1_1392_456d_ad1f_0ec6030969a3_Node'}]

[mbbyn]

For those interested, we are hosting HuggingFace Text Embedding Inference server, which exposes an OpenAI compatible API. In order to ask it to infer on a certain model, we pass the model name to the server, which are usually very long names.

[mbbyn]

Related #1857 @crazywoola

[crazywoola]

Actually, it's pretty easy to fix this, I will upgrade the migrations later.

[mbbyn]

I have tested this change, but the issue still persists, unfortunately. The PR updates provider_name in embeddings table, but we would want to update model_name in dataset_collection_bindings instead. It is also worth increading the limit of model_name in other tables, such as embeddings.

[HiroshigeAoki]

@mbbyn
I ran into the same issue. So I fixed it.