Torchfix

- Updated libCST dependency to 1.5.0 to support running on Python 3.13
- Added `torch.matrix_rank` and `torch.lstsq` to the list of deprecated/removed APIs

- Added `torch.utils._pytree._register_pytree_node` and `torch.backends.cuda.sdp_kernel` to the deprecated APIs rules
- Enhanced rule TOR203 to support `torchvision.datasets` and `transforms` in addition to `models`
- Added rules TOR106 and TOR107 to suggest replacing `torch.log(1 + x)` and `torch.exp(x) - 1` with more numerically stable equivalents
- Multiple code refactorings, bug fixes, and quality of life and documentation improvements

- Added rule TOR203 to replace 'import torchvision.models as models' with 'from torchvision import models'
- Added rules TOR104 and TOR105 for calling and importing non-public PyTorch functions that have known public aliases
- Added rules TOR004 and TOR103 for importing removed and deprecated functions (in addition to the existing rules for calling those functions)
- Fixed loading for deprecated symbols config in zipped deployments
- Done several smaller bug fixes and refactorings

- Improvements for the standalone `torchfix` command:
- Added `--version` flag
- `--select` flag now accepts specific rules, not just `ALL`
- Fixed excessive debug output on MacOS
- Added PyTorch-internal rule TOR901
- TorchFix explicitly requires at least Python 3.9 now
- Small clean-ups and bugfixes

- Added rule TOR003 about explicitly passing `use_reentrant` to `torch.utils.checkpoint`
- Added `torch.nn.utils.weight_norm` to the list of deprecated functions flagged by TOR101
- Updated README with TOR0 rules description

This is the first release for pytorch-labs/torchfix repo, with the only differences from TorchFix 0.2.0 on PyPI are files related to repo maintenance and project metadata.