Convert python samples to use UV with uv.lock and .toml files also remove conda dependency

[eleemhui]

Existing Sample Changes

Description

ONSAM 1917
In these changes we are using UV to manage the python dependencies, kernels, and venv.
Each test will be isolated from all other tests and packages added/changed/removed will not affect later tests.
pyproject.toml files have been added to each sample test which list all the dependencies and required versions (takes the place of the requirements.txt).
Adds uv.lock file to maintain consistency of packages being run on any environment.
Removes all calls to 'conda' in the sample.json files as well as dependence on the AI Tools Offline Installer

The following test cases are completed and tested:

Features-and-Functionality:
IntelPyTorch_GPU_InferenceOptimization_with_AMP
IntelPyTorch_TrainingOptimizations_AMX_BF16
IntelPython_GPU_dpnp_Genetic_Algorithm
IntelPython_Numpy_Numba_dpnp_kNN
IntelPython_XGBoost_Performance
IntelTensorFlow_AMX_BF16_Inference
IntelTensorFlow_AMX_BF16_Training
IntelTensorFlow_InferenceOptimization
IntelTensorFlow_Transformer_AMX_bfloat16_MixedPrecision
IntelTransformers_Quantization

Getting-Started-Samples:
INC-Quantization-Sample-for-PyTorch
INC-Sample-for-Tensorflow
Intel_Extension_For_SKLearn_GettingStarted
Intel_Extension_For_TensorFlow_GettingStarted

External Dependencies

none.

Type of change

Please delete options that are not relevant. Add a 'X' to the one that is applicable.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Implement fixes for ONSAM Jiras
  ONSAM 1917

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration

• Command Line
• oneapi-cli
• Visual Studio
• Eclipse IDE
• VSCode
• When compiling the compliler flag "-Wall -Wformat-security -Werror=format-security" was used

[Ankur-singh]

Hi @eleemhui thanks the amazing PR. It looks really good. Would love to know how you tested the samples. I use a combination of Makefile and Docker, TBH it's quite clumsy. Would love to know your setup.

We can merge the PR after testing.

[eleemhui]

@Ankur-singh. We tested these using a sapphire rapids server we had access to. All these samples are python so we did not need to use Makefile on any of them we just dealt with adding all the correct python dependencies to the pyproject.toml file.
While we were working on them we checked the notebooks were working correctly in VSCode and once they worked we would then test each one by running the commands listed in sample.json and verified the correct output (for both the .py and the .ipynb).
Some of the samples were for GPU and we tested those on Intel Tiber Cloud, which was more clumsy.

[Ankur-singh]

Can relate to it. Since we have to do it very frequently, I wrote some python script to parse the sample.json file, dynamically create bash script base on commands from sample.json file and run them inside docker. Still quite clumsy and breaks every now and then.

I have the setup environment ready. Will be testing this PR over the weekend, I have a few things to take care of tomorrow. Hope thats fine.

[aemorabr]

@Ankur-singh just wanted to add that there is a second part PR for some other python samples: #2623, it is the same solution but with the rest of samples we left behind of this one.