Skip to content

strip_to_docstrings #640

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 7 commits into
base: main
Choose a base branch
from
Draft

strip_to_docstrings #640

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
dcb955d
strip_to_docstrings
mishig25 Sep 30, 2025
34b265f
wip
mishig25 Sep 30, 2025
b31b1b5
wip
mishig25 Sep 30, 2025
a90170f
wip
mishig25 Sep 30, 2025
1747a9e
wip
mishig25 Sep 30, 2025
3df3f49
fix: simplify uv sync command in documentation build workflow
mishig25 Sep 30, 2025
b7d9105
wip
mishig25 Sep 30, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
144 changes: 92 additions & 52 deletions .github/workflows/build_pr_documentation.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,8 +30,6 @@ on:
type: string
install_rust:
type: boolean
install_libgl1:
type: boolean
# Command to execute before building the documentation
pre_command:
type: string
Expand All @@ -46,12 +44,6 @@ on:
convert_notebooks:
type: boolean
description: "Convert notebooks to markdown files before building docs."
# Docker image to use for the build. Set custom_container="huggingface/transformers-doc-builder" if you need
# a complete install. Default to `""` which means CI will run on the runner directly.
custom_container:
type: string
default: ""
description: "Docker image to use for the build."
# Debug purposes only!
doc_builder_revision:
type: string
Expand All @@ -62,8 +54,6 @@ on:
jobs:
build_pr_documentation:
runs-on: ubuntu-22.04
container:
${{ inputs.custom_container }}
env:
UV_HTTP_TIMEOUT: 900 # max 15min to install deps (shouldn't take more than 5min)

Expand Down Expand Up @@ -91,29 +81,6 @@ jobs:
node-version: '20'
cache-dependency-path: "kit/package-lock.json"

- name: Export ROOT_APT_GET ('apt-get' or 'sudo apt-get')
# Use `sudo` only if running on the base runner.
# When using a container, `sudo` is not needed (and not installed)
run: |
if [ -z "${{ inputs.custom_container }}" ]
then
echo "ROOT_APT_GET=sudo apt-get" >> $GITHUB_ENV
else
echo "ROOT_APT_GET=apt-get" >> $GITHUB_ENV
fi

- name: Export PIP_OR_UV ('pip' or 'uv pip')
# Use `uv` only if running on the base runner.
# When using a container, `pip` has already been used to installed existing deps
# and is therefore quicker to resolve (already have some cached stuff).
run: |
if [ -z "${{ inputs.custom_container }}" ]
then
echo "PIP_OR_UV=uv pip" >> $GITHUB_ENV
else
echo "PIP_OR_UV=pip" >> $GITHUB_ENV
fi

- name: Set env variables
run: |
if [ -z "${{ inputs.path_to_docs }}" ]
Expand All @@ -138,47 +105,120 @@ jobs:
echo "package_name=${{ inputs.package_name }}" >> $GITHUB_ENV
fi

- name: Install libgl1
if: inputs.install_libgl1
run: $ROOT_APT_GET install -y libgl1

- name: Install Rust
uses: actions-rs/toolchain@v1
if: inputs.install_rust
with:
toolchain: stable

# Use venv in both cases
- name: Install uv
uses: astral-sh/setup-uv@v4
with:
version: "latest"

- name: Setup doc-builder environment
working-directory: doc-builder
run: |
pip install -U uv
uv venv
git pull origin ${{ inputs.doc_builder_revision }}
uv sync

- name: Setup environment
- name: Strip package to docstrings only
shell: bash
run: |
source .venv/bin/activate
$PIP_OR_UV uninstall doc-builder
cd doc-builder
git pull origin ${{ inputs.doc_builder_revision }}
$PIP_OR_UV install .
cd ..
source doc-builder/.venv/bin/activate

# Only strip if it's a Python module (skip for non-Python packages like HF Course)
if [[ "${{ inputs.additional_args }}" != *"--not_python_module"* ]]; then
echo "Running strip_to_docstrings.py on ${{ inputs.package }}"
python doc-builder/strip_to_docstrings.py ${{ inputs.package }}
echo "✅ Stripping complete"
else
echo "Skipping stripping for non-Python module"
fi

- name: Install package-specific dependencies
shell: bash
run: |
source doc-builder/.venv/bin/activate

# Install package-specific dependencies based on package name
package_name="${{ inputs.package_name }}"
if [ -z "$package_name" ]; then
package_name="${{ inputs.package }}"
fi

echo "Installing dependencies for package: $package_name"

case "$package_name" in
"transformers")
echo "📦 Installing transformers dependencies (recommended: ~170 MB)"
uv pip install torch --index-url https://download.pytorch.org/whl/cpu
uv pip install numpy pillow safetensors
uv pip install sentencepiece tokenizers protobuf regex
uv pip install huggingface-hub packaging filelock requests tqdm pyyaml
;;

"diffusers")
echo "📦 Installing diffusers dependencies (~200 MB)"
uv pip install torch --index-url https://download.pytorch.org/whl/cpu
uv pip install numpy pillow safetensors
uv pip install huggingface-hub packaging filelock requests pyyaml
;;

"peft")
echo "📦 Installing peft dependencies (~120 MB)"
uv pip install torch --index-url https://download.pytorch.org/whl/cpu
uv pip install safetensors numpy
uv pip install huggingface-hub packaging filelock requests
;;

"accelerate")
echo "📦 Installing accelerate dependencies (~150 MB)"
uv pip install torch --index-url https://download.pytorch.org/whl/cpu
uv pip install numpy safetensors psutil
uv pip install huggingface-hub packaging filelock requests pyyaml
;;

"huggingface_hub"|"huggingface-hub")
echo "📦 Installing huggingface_hub dependencies (lightweight: ~6 MB, no PyTorch!)"
uv pip install requests pyyaml packaging filelock
;;

*)
echo "📦 Installing default dependencies (full set: ~170 MB)"
# Default to full dependency set for unknown packages
uv pip install torch --index-url https://download.pytorch.org/whl/cpu
uv pip install numpy pillow safetensors
uv pip install sentencepiece tokenizers protobuf regex
uv pip install huggingface-hub packaging filelock requests tqdm pyyaml psutil
;;
esac

echo "✅ Dependencies installed"

- name: Install package
shell: bash
run: |
source doc-builder/.venv/bin/activate

# Install the stripped package
if [[ -n "${{ inputs.package_path }}" ]]
then
cd ${{ inputs.package_path }}
$PIP_OR_UV install .[dev]
uv pip install -e .
elif [[ "${{ inputs.additional_args }}" != *"--not_python_module"* ]];
then
cd ${{ inputs.package }}
$PIP_OR_UV install .[dev]
uv pip install -e .
fi
cd ..

echo "✅ Package installed"

- name: Run pre-command
shell: bash
run: |
source .venv/bin/activate
source doc-builder/.venv/bin/activate
if [ ! -z "${{ inputs.pre_command }}" ]
then
bash -c "${{ inputs.pre_command }}"
Expand All @@ -187,7 +227,7 @@ jobs:
- name: Convert notebooks to markdown files
if: inputs.convert_notebooks
run: |
source .venv/bin/activate
source doc-builder/.venv/bin/activate
branch=${GITHUB_HEAD_REF:-${GITHUB_REF#refs/heads/}}
remaining_part=$(echo "${{ env.doc_folder }}" | perl -pe 's|^[^/]+||')
remaining_part=${remaining_part%/}
Expand All @@ -200,7 +240,7 @@ jobs:
NODE_OPTIONS: --max-old-space-size=6656
shell: bash
run: |
source .venv/bin/activate
source doc-builder/.venv/bin/activate
echo "doc_folder has been set to ${{ env.doc_folder }}"
cd doc-builder
args="--build_dir ../build_dir --clean --version pr_${{ inputs.pr_number }} --html ${{ inputs.additional_args }} --repo_owner ${{ inputs.repo_owner }} --repo_name ${{ inputs.package }} --version_tag_suffix=${{ inputs.version_tag_suffix }}"
Expand Down
Loading