issue #51: implement pipeline switching in back end

[MaxenceGui]

• Added model_name parameter handling in inference_request function.
• Created pipelines_Endpoints dictionary to map pipeline names to their respective processing functions.
• Integrated dynamic model selection based on the model_name parameter.
• Ensured consistent data structure for process_inference_results function across different models.
• Validate swin and seed detector pipeline work
• Implement model categories with their task (Classification, Object detection, Segmentation)
• Implement API call function per model and logic to process the inference result per model
• Set up a JSON with the pipelines in the blob storage while the database is not yet ready
• Create blob container for the pipelines JSON files
• Implement test on inference requests for different categories
• issue Refactor test to implement asyncio run and test error detection instead of False being return #59
• issue Refactor function from module in the backend to catch errors and raise custom exceptions #61
• issue Create the pipeline insertion module #60

[rngadam]

you have missing EOF newline. please add

please add

https://github.com/ai-cfia/github-workflows/blob/main/.github/workflows/workflow-repo-standards-validation.yml

to your repo

[rngadam]

As discussed yesterday, the code here does not seem to be going in the right direction in terms of readability, testability and simplicity

[MaxenceGui]

I've added model documentation from the backend. It is still a work in progress. Following this doc, here is a list of what's left to do in the code:

• Implement model categories with their task (Classification, Object detection, Segmentation)
• Implement test on inference requests for different categories
• Set up a JSON with the pipeline in the blob storage while the database is not yet ready

I'm asking for reviews from:

• William (for your English skill and to keep you in the loop on Nachet)
• Leron (to keep you in the loop)
• Amir (to verify that I implemented your suggestion the good way)
• François (to look at the metadata structure of model and pipeline)
• Ricky

To access the new documentation: docs/nachet-model-documentation.md

[rngadam]

you have a number of lint failures reported from the Github Workflows.

Suggest managing your TODO list as a checklist in the description of this pull request (it's easy to lose track in the comments of the PR

[rngadam]

there's still basic issues with EOF newline missing so looks like the Github workflows are not all running.

[rngadam]

@MaxenceGui
New issues were created related to the work here so we can postpone some of the work and get this huge PR merged in main.

Modify the description to add checklist items as completed by referencing postponed tasks to new issues; this gives us a trace between the work that should have been done here and follow up issues.

[MaxenceGui]

@MaxenceGui New issues were created related to the work here so we can postpone some of the work and get this huge PR merged in main.

Modify the description to add checklist items as completed by referencing postponed tasks to new issues; this gives us a trace between the work that should have been done here and follow up issues.

Perfect, will do that. I push my last commit so all checks pass. @RussellJimmies and @rngadam if there is any last-minute review you wanna add, go ahead. If you approve this PR I will rebase and push it to main.

[rngadam]

Seems like the issue links are created manually in the description? I think they look nicer if you use this notation:

* issue #59 
* issue #61 
* issue #60

output:

• issue Refactor test to implement asyncio run and test error detection instead of False being return #59
• issue Refactor function from module in the backend to catch errors and raise custom exceptions #61
• issue Create the pipeline insertion module #60

[MaxenceGui]

Seems like the issue links are created manually in the description? I think they look nicer if you use this notation:

• issue Refactor test to implement asyncio run and test error detection instead of False being return #59
• issue Refactor function from module in the backend to catch errors and raise custom exceptions #61
• issue Create the pipeline insertion module #60

It is I will absolutely change it

[MaxenceGui]

@RussellJimmies just missing your approval before merging in with main

[RussellJimmies]

Gargantuan work, great job! Code LGTM.