wip feat(Documents): add support for classification #711
Conversation
github-actions
bot
added
test:uipath-langchain
alexandrujircan
changed the base branch from
main
to
feat/documents/add-support-for-digitization
| _are_mutually_exclusive(file=file, file_path=file_path) | ||
|
|
||
| if _is_provided(project_name): | ||
| _must_be_provided(project_type=project_type) | ||
| _must_not_be_provided(digitization_result=digitization_result) | ||
| else: | ||
| _must_not_be_provided(project_type=project_type) | ||
| _must_not_be_provided(file=file, file_path=file_path) | ||
| if _is_provided(digitization_result): | ||
| project_type = digitization_result.project_type | ||
| else: | ||
| _must_be_provided(classification_result=classification_result) | ||
| _must_not_be_provided(document_type_name=document_type_name) | ||
| project_type = ProjectType.MODERN | ||
|
|
||
| if _is_not_provided(classification_result): | ||
| if project_type == ProjectType.MODERN: | ||
| _must_be_provided(document_type_name=document_type_name) | ||
| else: | ||
| _must_not_be_provided(document_type_name=document_type_name) | ||
|
|
||
| return project_type |
There was a problem hiding this comment.
Painful to read through this. We really need to split into more granular functionalities per project type. But leaving that aside, isn't it missing some cases that are invalid? e.g. passing digitization with an IXP project and also a classification result.
There was a problem hiding this comment.
Yes, I refined these checks in a later PR, so I also need to update them here. But as you said, there are lots of conditions and branches, it's even worse trying to describe them in words, like in docstrings
| project_id: str = Field(alias="ProjectId") | ||
|
|
||
|
|
||
| class ClassificationResponse(BaseModel): |
There was a problem hiding this comment.
This is not used. I guess the purpose of this class was to be used as response type for the classify method, right?
There was a problem hiding this comment.
Yes, I forgot to delete it. At first, I thought about returning the entire response, but I think it’s more intuitive to just return a list of classification results, to keep it consistent with the other functions
2 participants
add support for classification