Skip to content

DocumentIntelligenceConverter should not set a default api_version #1904

@evozondatascience

Description

@evozondatascience

I wanted to use markitdown[az-doc-intel] and I encountered a problem on api_version.

  1. Apparently the Azure document intelligence has a default version of '2024-11-30' set inside the DocumentIntelligenceClient from azure.
  2. Inside the DocumentIntelligenceConverter class the api version is set on default as being api_version: str = "2024-07-31-preview".
  3. For most of azure document intelligence deployments they don't even specify an api_version endpoint anymore and the default is always 2024-11-30. So if you pass an empty one the DocumentIntelligenceConverter will set the default to the wrong one which will cause azure to give 404 resource not found.

The fix: would be that DocumentIntelligenceConverter accepts api_version as null and it can safely let the DocumentIntelligenceClient take care of setting the right api_version.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions