remove some warnings

[EdwardLi-coder]

When I run the tests, I often encounter warnings related to library dependencies. In my PR, I've ignored warnings coming from third-party libs. And if they come from our own code, I'm rewriting it.
These warnings are need to rewrite code.

tests/unit/lib/test_text.py: 1 warning
  /Users/lifei/PycharmProjects/datachain/.nox/tests-3-12/lib/python3.12/site-packages/datachain/lib/text.py:36: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    tokens = torch.tensor(tokens)

tests/unit/lib/test_clip.py::test_similarity_scores_hf
tests/unit/lib/test_image.py::test_convert_image_hf
  /Users/lifei/PycharmProjects/datachain/.nox/tests-3-12/lib/python3.12/site-packages/datachain/lib/image.py:37: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    img = torch.tensor(img.pixel_values[0])  # type: ignore[assignment,attr-defined]

tests/unit/lib/test_clip.py::test_similarity_scores_hf
  /Users/lifei/PycharmProjects/datachain/.nox/tests-3-12/lib/python3.12/site-packages/datachain/lib/clip.py:21: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    return lambda x: method(torch.tensor(x))

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.31%. Comparing base (ca36654) to head (3943e4f).
Report is 3 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #404   +/-   ##
=======================================
  Coverage   87.31%   87.31%           
=======================================
  Files          92       92           
  Lines        9982     9982           
  Branches     2041     2041           
=======================================
  Hits         8716     8716           
  Misses        911      911           
  Partials      355      355           

Flag	Coverage Δ
datachain	87.26% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[shcheklein]

can we ignore warning that come from third-party libs (e.g. some lib that we depend on is using torch in a deprecated way) and fix warnings on our end by rewriting the code, wdyt?

[EdwardLi-coder]

can we ignore warning that come from third-party libs (e.g. some lib that we depend on is using torch in a deprecated way) and fix warnings on our end by rewriting the code, wdyt?

I think I've already done this.
In my PR, I've ignored warnings coming from third-party libs. And if they come from our own code, I'm rewriting it.

[shcheklein]

In my PR, I've ignored warnings coming from third-party libs. And if they come from our own code, I'm rewriting it.

will it ignore new warnings (let's say someone adds it by mistake tomorrow) in our code? just to double check ...

[EdwardLi-coder]

In my PR, I've ignored warnings coming from third-party libs. And if they come from our own code, I'm rewriting it.

will it ignore new warnings (let's say someone adds it by mistake tomorrow) in our code? just to double check ...

These ignore settings are specific to warnings from certain third-party libraries. They will not affect new warnings in our own code. Each ignore has a specific module path, so even if someone accidentally adds new warnings in our code in the future, these settings will not ignore them. New warnings in our code will still be captured and displayed.
However, to ensure we don't miss important warnings, I suggest we periodically review these ignore settings and re-evaluate their necessity after updates to the third-party libraries.

[EdwardLi-coder]

I think maybe only ignore DeprecationWarning is good.

[EdwardLi-coder]

This is the warning:

tests/unit/lib/test_clip.py::test_similarity_scores_hf
  /Users/lifei/PycharmProjects/datachain/.nox/tests-3-12/lib/python3.12/site-packages/datachain/lib/clip.py:21: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    return lambda x: method(torch.tensor(x))

Update Reason:
We use torch.as_tensor() to create tensors, which avoids copying data whenever possible. We then call clone().detach() to create a new tensor that is detached from the original data and will not affect the computation graph of the original data.

[EdwardLi-coder]

This is the warning:

  /Users/lifei/PycharmProjects/datachain/.nox/tests-3-12/lib/python3.12/site-packages/torchvision/transforms/v2/_deprecated.py:42: UserWarning: The transform `ToTensor()` is deprecated and will be removed in a future release. Instead, please use `v2.Compose([v2.ToImage(), v2.ToDtype(torch.float32, scale=True)])`.Output is equivalent up to float precision.
    warnings.warn(

update reason：
Compatibility: ToTensor() has been deprecated and may be removed in future versions. Using the new methods ensures your code will continue to work in future versions.

Functional equivalence: The new method combination (v2.ToImage() and v2.ToDtype()) provides the same functionality as ToTensor(), but with more flexibility and clarity.

Precision: The new methods provide output equivalent to ToTensor() up to float precision.

[EdwardLi-coder]

@shcheklein I only remain ignore the DeprecationWarning. I have provided some explanations for my modifications. If you have any other suggestions, I would be happy to make further changes.

[dreadatour]

Looks good to me! 👍

[shcheklein]

thanks @EdwardLi-coder !