[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
#!/usr/bin/env python3
|
|
|
|
|
|
"""
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
Bootstrap Git pre‑push hook with isolated virtual environment.
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
|
|
|
|
|
✓ Requires uv to be installed (fails if not available)
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
✓ Creates isolated venv in .git/hooks/linter/.venv/ for hook dependencies
|
|
|
|
|
|
✓ Installs lintrunner only in the isolated environment
|
|
|
|
|
|
✓ Creates direct git hook that bypasses pre-commit
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
|
|
|
|
|
Run this from the repo root (inside or outside any project venv):
|
|
|
|
|
|
|
|
|
|
|
|
python scripts/setup_hooks.py
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
|
|
|
|
|
|
IMPORTANT: The generated git hook references scripts/lintrunner.py. If users checkout
|
|
|
|
|
|
branches that don't have this file, git push will fail with "No such file or directory".
|
|
|
|
|
|
Users would need to either:
|
|
|
|
|
|
1. Re-run the old setup_hooks.py from that branch, or
|
|
|
|
|
|
2. Manually delete .git/hooks/pre-push to disable hooks temporarily, or
|
|
|
|
|
|
3. Switch back to a branch with the new scripts/lintrunner.py
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
"""
|
|
|
|
|
|
|
|
|
|
|
|
from __future__ import annotations
|
|
|
|
|
|
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
import shlex
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
import shutil
|
|
|
|
|
|
import subprocess
|
|
|
|
|
|
import sys
|
|
|
|
|
|
from pathlib import Path
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# Add scripts directory to Python path so we can import lintrunner module
|
|
|
|
|
|
scripts_dir = Path(__file__).parent
|
|
|
|
|
|
sys.path.insert(0, str(scripts_dir))
|
|
|
|
|
|
|
|
|
|
|
|
# Import shared functions from lintrunner module
|
|
|
|
|
|
from lintrunner import find_repo_root, get_hook_venv_path
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# Restore sys.path to avoid affecting other imports
|
|
|
|
|
|
sys.path.pop(0)
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ───────────────────────────────────────────
|
|
|
|
|
|
# Helper utilities
|
|
|
|
|
|
# ───────────────────────────────────────────
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
def run(cmd: list[str], cwd: Path = None) -> None:
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
print(f"$ {' '.join(cmd)}")
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
subprocess.check_call(cmd, cwd=cwd)
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def which(cmd: str) -> bool:
|
|
|
|
|
|
return shutil.which(cmd) is not None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def ensure_uv() -> None:
|
|
|
|
|
|
if which("uv"):
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
return
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
|
|
|
|
|
sys.exit(
|
|
|
|
|
|
"\n❌ uv is required but was not found on your PATH.\n"
|
|
|
|
|
|
" Please install uv first using the instructions at:\n"
|
|
|
|
|
|
" https://docs.astral.sh/uv/getting-started/installation/\n"
|
|
|
|
|
|
" Then rerun python scripts/setup_hooks.py\n"
|
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if sys.platform.startswith("win"):
|
|
|
|
|
|
print(
|
|
|
|
|
|
"\n⚠️ Lintrunner is not supported on Windows, so there are no pre-push hooks to add. Exiting setup.\n"
|
|
|
|
|
|
)
|
|
|
|
|
|
sys.exit(0)
|
|
|
|
|
|
|
|
|
|
|
|
# ───────────────────────────────────────────
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
# 1. Setup isolated hook environment
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
# ───────────────────────────────────────────
|
|
|
|
|
|
|
|
|
|
|
|
ensure_uv()
|
|
|
|
|
|
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
# Find repo root and setup hook directory
|
|
|
|
|
|
repo_root = find_repo_root()
|
|
|
|
|
|
venv_dir = get_hook_venv_path()
|
|
|
|
|
|
hooks_dir = venv_dir.parent.parent # Go from .git/hooks/linter/.venv to .git/hooks
|
|
|
|
|
|
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
print(f"Setting up isolated hook environment in {venv_dir}")
|
|
|
|
|
|
|
|
|
|
|
|
# Create isolated virtual environment for hooks
|
|
|
|
|
|
if venv_dir.exists():
|
|
|
|
|
|
print("Removing existing hook venv...")
|
|
|
|
|
|
shutil.rmtree(venv_dir)
|
|
|
|
|
|
|
2025-10-10 11:09:13 -07:00
|
|
|
|
run(["uv", "venv", str(venv_dir), "--python", "3.10"])
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
|
|
|
|
|
|
# Install lintrunner in the isolated environment
|
|
|
|
|
|
print("Installing lintrunner in isolated environment...")
|
|
|
|
|
|
run(
|
|
|
|
|
|
["uv", "pip", "install", "--python", str(venv_dir / "bin" / "python"), "lintrunner"]
|
|
|
|
|
|
)
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
|
|
|
|
|
# ───────────────────────────────────────────
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
# 2. Create direct git pre-push hook
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
# ───────────────────────────────────────────
|
|
|
|
|
|
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
pre_push_hook = hooks_dir / "pre-push"
|
|
|
|
|
|
python_exe = venv_dir / "bin" / "python"
|
|
|
|
|
|
lintrunner_script_path_quoted = shlex.quote(
|
|
|
|
|
|
str(repo_root / "scripts" / "lintrunner.py")
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
)
|
|
|
|
|
|
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
hook_script = f"""#!/bin/bash
|
|
|
|
|
|
set -e
|
|
|
|
|
|
|
|
|
|
|
|
# Check if lintrunner script exists (user might be on older commit)
|
|
|
|
|
|
if [ ! -f {lintrunner_script_path_quoted} ]; then
|
|
|
|
|
|
echo "⚠️ {lintrunner_script_path_quoted} not found - skipping linting (likely on an older commit)"
|
|
|
|
|
|
exit 0
|
|
|
|
|
|
fi
|
|
|
|
|
|
|
|
|
|
|
|
# Run lintrunner wrapper using the isolated venv's Python
|
|
|
|
|
|
{shlex.quote(str(python_exe))} {lintrunner_script_path_quoted}
|
|
|
|
|
|
"""
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
print(f"Creating git pre-push hook at {pre_push_hook}")
|
|
|
|
|
|
pre_push_hook.write_text(hook_script)
|
|
|
|
|
|
pre_push_hook.chmod(0o755) # Make executable
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
|
|
|
|
|
|
print(
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
"\n✅ Isolated hook environment created and pre‑push hook is active.\n"
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
" Lintrunner will now run automatically on every `git push`.\n"
|
[BE] Isolate pre-push hook dependencies in dedicated virtual environment (#160048)
This adds two changes:
- Isolates pre-push hook dependencies into an isolated venv, no longer affect your system environment
- Lets you manually run the pre-push lintrunner (including with lintrunner -a) by invoking `python scripts/lintrunner.py [-a]` (it's ugly, but better than nothing...for now)
This is a follow up to:
- https://github.com/pytorch/pytorch/pull/158389
## Problem
The current pre-push hook setup installs lintrunner and related dependencies globally, which makes developers nervous about system pollution and can cause version conflicts with existing installations.
Also, if the pre-push lintrunner found errors, you had to hope your normal lintrunner could fix them (which wasn't always the case, e.g. if those errors only manifested in certain python versions)
## Key Changes:
- Isolated Environment: Creates .git/hooks/linter/.venv/ with Python 3.9 (the python used in CI) and an isolated lintrunner installation
- User-Friendly CLI: New python scripts/lintrunner.py wrapper allows developers to run lintrunner (including -a auto-fix) from any environment
- Simplified Architecture: Eliminates pre-commit dependency entirely - uses direct git hooks
File Changes:
- scripts/setup_hooks.py: Rewritten to create isolated uv-managed virtual environment
- scripts/lintrunner.py: New wrapper script with shared hash management logic
- scripts/run_lintrunner.py: Removed (functionality merged into lintrunner.py)
- .pre-commit-config.yaml: Removed (no longer needed)
## Usage:
```
# Setup (run once)
python scripts/setup_hooks.py
# Manual linting (works from any environment)
python scripts/lintrunner.py # Check mode
python scripts/lintrunner.py -a # Auto-fix mode
# Git hooks work automatically
git push # Runs lintrunner in isolated environment
# Need to skip the pre-push hook?
git push --no-verify
```
## Benefits:
- ✅ Zero global dependency installation
- ✅ Per-repository isolation prevents version conflicts
- ✅ Full lintrunner functionality is now accessible
## Implementation Notes:
- Virtual env is kept in a dedicated dir in .git, to keep per-repo mechanics
- lintrunner.py does not need to be invoked from a specific venv. It'll invoke the right venv itself.
A minor bug: It tends to garble the lintrunner output a bit, like the screenshot below shows, but I haven't found a workaround so far and it remains understandable to users:
<img width="241" height="154" alt="image" src="https://github.com/user-attachments/assets/9496f925-8524-4434-8486-dc579442d688" />
## What's next?
Features that could be added:
- Check for lintrunner updates, auto-update if needed
- Depending on dev response, this could be enabled by default for all pytorch/pytorch environments
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160048
Approved by: https://github.com/seemethere
2025-08-12 01:58:44 +00:00
|
|
|
|
f" Hook dependencies are isolated in {venv_dir}\n"
|
[BE] Add pre-push hook for lintrunner to the PyTorch repo (#158389)
Adds a pre-commit hook (technically a pre-push hook) to the PyTorch repo.
**This is currently an opt-in feature**, which one can opt into by running `python scripts/setup_hooks.py` locally.
### Features
- **Run Lintrunner Before Push**: Before every `git push`, automatically runs lintrunner on your changes.
- Really need to skip the checks? Run `git push --no-verify`
- **Consistent, Isolated, Lintrunner Environment**: During pre-push, Lintrunner runs in it's own virtual en environment that contain all lintrunner dependencies in a consistent, isolated environment. No more lintrunner failures because you created a new .venv. (Did you know you needed to run `lintrunner init` every time you make a new .venv?)
- **Dependencies Automatically Updated**: If .lintrunner.toml is updated, this will automatically re-run `lintrunner init` to ensure you install the latest dependencies specified
### Installation
- Run `python scripts/setup_hooks.py`. Now every `git push` will first run lintrunner.
### Additional details
- The lintrunner used by the pre-push hook runs in a special per-repo virtual environment managed by the commit-hook tool located under `$USER/.cache/pre-commit`
- Does not affect your regularly used lintrunner
- Manual invocations of lintrunner will continue to depend on your local environment instead of the special pre-push one. If there's enough interest, we could explore consolidating them.
- Does not run `lintrunner -a` for you.
- You still need to manually run that (can be changed later though!)
- Have staged/unstaged changes? No worries
- This runs `git stash` before running the pre-commit hooks and pops back your changes afterwards, so only the changes actaully being pushed will be tested
### Downsides
- No streaming UI updates
- While you still get the same output from lintrunner that you're used to, the commit-hook framework doesn't show any output while lintrunner is actually running. Instead, it shows the entire output after linter has completed execution, which could be a few minutes (especially if it has to run `lintrunner init` first)
- `uv` installation is required to run the setup script. The setup script will ask users to install uv if it's not available.
- This is required to be able to install the pre-commit package in a safe way that's available no matter what .venv you are running in.
### Opting out
- Disable hook for a single push: Run `git push --no-verify`
- Disable hook permanently: If something goes wrong and you need to wipe your setup:
- Delete the `$USER/.cache/pre-commit` folder and the `.git/hooks/pre-push` file in your local repo.
- You can now rerun `python scripts/setup_hooks.py` to setup your git push hook again if you want.
### Potential Future Changes
Things that could be done to make this even better if folks like these ideas:
- Automatic setup
- Our `CONTRIBUTING.md` file tells devs to run `make setup-env`. That could be a good entry point to hook the installation into
- Fix the console output streaming
- Make every lintrunner invocation (including manual ones) use the same repo-specific venv that the commit-hook uses.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158389
Approved by: https://github.com/seemethere
2025-07-18 19:55:35 +00:00
|
|
|
|
)
|