ci/docs: split building html | linkcheck (#281)

Author: Borda

.github/workflows/ci_docs.yml

@@ -10,11 +10,19 @@ on: # Trigger the workflow on push or pull request
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref }}
-  cancel-in-progress: true
+  cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+
+defaults:
+  run:
+    shell: bash
 
 jobs:
-  build-docs:
+  make-docs:
     runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        check: ["html", "linkcheck"]
     env:
       PUB_BRANCH: publication
       PATH_DATASETS: ${{ github.workspace }}/.datasets
@@ -35,16 +43,17 @@ jobs:
           key: pip-${{ hashFiles('requirements.txt') }}-${{ hashFiles('_requirements/docs.txt') }}
           restore-keys: pip-
 
-      - name: Install dependencies
+      - name: Install Texlive & tree
         run: |
           sudo apt-get update --fix-missing
-          sudo apt-get install -y tree
           # install Texlive, see https://linuxconfig.org/how-to-install-latex-on-ubuntu-20-04-focal-fossa-linux
-          sudo apt-get install -y cmake pandoc texlive-latex-extra dvipng texlive-pictures
+          sudo apt-get install -y cmake tree pandoc texlive-latex-extra dvipng texlive-pictures
+
+      - name: Install dependencies
+        run: |
           pip --version
           pip install -q -r requirements.txt -r _requirements/docs.txt
           pip list
-        shell: bash
 
       - name: Process folders
         run: |
@@ -54,21 +63,21 @@ jobs:
           python .actions/assistant.py group-folders master-diff.txt
           printf "Changed folders:\n"
           cat changed-folders.txt
-        shell: bash
 
-      - name: ">> output"
-        id: changed
-        run: python -c "lines = open('changed-folders.txt').readlines(); print(f'::set-output name=nb_dirs::{len(lines)}')"
+      - name: Count changed notebooks
+        run: python -c "lines = open('changed-folders.txt').readlines(); print(f'NB_DIRS={len(lines)}')" >> $GITHUB_ENV
 
       - uses: oleksiyrudenko/gha-git-credentials@v2.1
         with:
-          token: "${{ secrets.GITHUB_TOKEN }}"
+          token: ${{ secrets.GITHUB_TOKEN }}
           global: true
       - name: Sync to pub
         run: git merge -s resolve origin/$PUB_BRANCH
 
       - name: Generate notebooks
-        if: steps.changed.outputs.nb_dirs != 0
+        if: ${{ env.NB_DIRS != 0 }}
+        env:
+          DRY_RUN: 1
         run: |
           # second half with || [...] is needed for reading the last line
           while read -r line || [ -n "$line" ]; do
@@ -77,12 +86,9 @@ jobs:
               cat .actions/_ipynb-render.sh
               bash .actions/_ipynb-render.sh
           done <<< $(cat changed-folders.txt)
-        env:
-          DRY_RUN: 1
-        shell: bash
 
       - name: Copy notebooks
-        if: steps.changed.outputs.nb_dirs != 0
+        if: ${{ env.NB_DIRS != 0 }}
         run: |
           # second half with || [...] is needed for reading the last line
           while read -r line || [ -n "$line" ]; do
@@ -91,22 +97,26 @@ jobs:
               cp .notebooks/${line}.ipynb changed-notebooks/${dir}/
           done <<< $(cat changed-folders.txt)
           tree changed-notebooks
-        shell: bash
 
       - uses: actions/upload-artifact@v3
-        if: steps.changed.outputs.nb_dirs != 0
+        if: ${{ matrix.check == 'html' && env.NB_DIRS != 0 }}
         with:
           name: notebooks-${{ github.sha }}
           path: changed-notebooks/
 
+      - name: Link check
+        working-directory: ./_docs
+        if: ${{ matrix.check == 'linkcheck' }}
+        run: make linkcheck --jobs $(nproc) --debug SPHINXOPTS="--keep-going"
+
       - name: Make Documentation
         working-directory: ./_docs
-        run: make html --jobs $(nproc) --debug SPHINXOPTS="-W --keep-going" linkcheck
+        if: ${{ matrix.check == 'html' }}
+        run: make html --jobs $(nproc) --debug SPHINXOPTS="-W --keep-going"
 
       - name: Upload built docs
+        if: ${{ matrix.check == 'html' }}
         uses: actions/upload-artifact@v3
         with:
           name: docs-html-${{ github.sha }}
           path: _docs/build/html/
-        # Use always() to always run this step to publish test results when there are test failures
-        if: success()

.github/workflows/ci_schema.yml

@@ -0,0 +1,15 @@
+name: Check Schema
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+jobs:
+  check:
+    uses: Lightning-AI/utilities/.github/workflows/check-schema.yml@main
+    with:
+      # skip azure due to the wrong schema file by MSFT
+      # https://github.com/Lightning-AI/lightning-flash/pull/1455#issuecomment-1244793607
+      azure-dir: ".azure"

.github/workflows/ci_test-acts.yml

@@ -2,7 +2,8 @@ name: CI internal
 
 # see: https://help.github.com/en/actions/reference/events-that-trigger-workflows
 on: # Trigger the workflow on push or pull request, but only for the main branch
-  push: {}
+  push:
+    branches: [main]
   pull_request:
     branches: [main]
 

course_UvA-DL/07-deep-energy-based-generative-models/Deep_Energy_Models.py

@@ -172,8 +172,7 @@
 # The trick is that we approximate $Z_{\theta}$ by a single Monte-Carlo sample.
 # This gives us the exact same objective as written above.
 #
-# Visually, we can look at the objective as follows (figure credit
-# - [Stefano Ermon and Aditya Grover](https://deepgenerativemodels.github.io/assets/slides/cs236_lecture11.pdf)):
+# Visually, we can look at the objective as follows (figure credit - Stefano Ermon and Aditya Grover: lecture cs236/11):
 #
 # <center width="100%"><img src="contrastive_divergence.svg" width="700px"></center>
 #
@@ -206,8 +205,7 @@
 # Modeling the probability distribution for sampling new data is not the only application of energy-based models.
 # Any application which requires us to compare two elements is much simpler to learn
 # because we just need to go for the higher energy.
-# A couple of examples are shown below (figure credit
-# - [Stefano Ermon and Aditya Grover](https://deepgenerativemodels.github.io/assets/slides/cs236_lecture11.pdf)).
+# A couple of examples are shown below (figure credit - Stefano Ermon and Aditya Grover: lecture cs236/11).
 # A classification setup like object recognition or sequence labeling can be considered as an energy-based
 # task as we just need to find the $Y$ input that minimizes the output $E(X, Y)$ (hence maximizes probability).
 # Similarly, a popular application of energy-based models is denoising of images.