Workflow to check we have a paper trail on répertoire changes (#629)

.github/workflows/pipeline.yml

@@ -0,0 +1,87 @@
+name: Pipeline
+
+
+on:
+  pull_request:
+    branches: ['*']
+    types: [opened, synchronize, reopened, labeled, unlabeled, ready_for_review, converted_to_draft, edited]
+
+jobs:
+  labels-for-repertoire-changes:
+    name: Labels for repertoire changes
+    outputs:
+      repertoire-changed: ${{ steps.compare-repertoire.outputs.repertoire-changed }}
+      pipeline-label: ${{ steps.check-labels.outputs.pipeline-label }}
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout merged UnicodeData.txt
+      uses: actions/checkout@v3
+      with:
+        path: merged
+        sparse-checkout: unicodetools/data/ucd/dev/UnicodeData.txt
+    - name: Checkout base UnicodeData.txt
+      uses: actions/checkout@v3
+      with:
+        ref: ${{ github.event.pull_request.base.sha }}
+        path: base
+        sparse-checkout: unicodetools/data/ucd/dev/UnicodeData.txt
+    - name: Compare repertoire
+      id: compare-repertoire
+      run: |
+        # Look for changes affecting the first two fields of UnicodeData.txt (code point and name).
+        sed 's/^\([^;]*;[^;]*\);.*$/\1/' merged/unicodetools/data/ucd/dev/UnicodeData.txt > merged-repertoire.txt
+        sed 's/^\([^;]*;[^;]*\);.*$/\1/' base/unicodetools/data/ucd/dev/UnicodeData.txt > base-repertoire.txt
+        if diff base-repertoire.txt merged-repertoire.txt
+        then echo "repertoire-changed=false" >> "$GITHUB_OUTPUT"
+        else echo "repertoire-changed=true" >> "$GITHUB_OUTPUT"
+        fi
+    - name: Checkout Python scripts
+      uses: actions/checkout@v3
+      with:
+        sparse-checkout: py/pipeline-workflow
+    - name: Check pipeline labels
+      id: check-labels
+      if: steps.compare-repertoire.outputs.repertoire-changed == 'true'
+      run: python3 py/pipeline-workflow/compare-repertoire.py
+  l2-document:
+    needs: labels-for-repertoire-changes
+    if: ${{ always() && needs.labels-for-repertoire-changes.outputs.repertoire-changed == 'true' }}
+    name: Proposal document
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout Python scripts
+      uses: actions/checkout@v3
+      with:
+        sparse-checkout: py/pipeline-workflow
+    - name: Check L2 document
+      run: |
+        python3 py/pipeline-workflow/check-l2-document.py
+  utc-decision:
+    needs: labels-for-repertoire-changes
+    if: needs.labels-for-repertoire-changes.outputs.repertoire-changed == 'true' && needs.labels-for-repertoire-changes.outputs.pipeline-label != 'pipeline-recommended-to-UTC'
+    name: UTC decision
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout Python scripts
+      uses: actions/checkout@v3
+      with:
+        sparse-checkout: py/pipeline-workflow
+    - name: Check UTC decision
+      run: python3 py/pipeline-workflow/check-utc-decision.py
+  draft-unless-approved:
+    needs: labels-for-repertoire-changes
+    if: needs.labels-for-repertoire-changes.outputs.repertoire-changed == 'true'
+    name: Draft unless approved
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout Python scripts and DerivedAge.txt
+      uses: actions/checkout@v3
+      with:
+        sparse-checkout: |
+          py/pipeline-workflow
+          unicodetools/data/ucd/dev/DerivedAge.txt
+    - name: Check draft status
+      env:
+        PIPELINE_LABEL: ${{ needs.labels-for-repertoire-changes.outputs.pipeline-label }}
+      run: python3 py/pipeline-workflow/check-draft-status.py
+

docs/pipeline.md

@@ -62,6 +62,9 @@ PR preparation:
 - [ ] UTC decisions — Cite _in extenso_
 - [ ] data-for-new — Set label
 - [ ] pipeline-* — Set label
+- [ ] PR button — Set to DRAFT pull request
+  - unless approved for the upcoming version
+- [ ] PR button — Press
 
 ## Scripts
 

py/pipeline-workflow/check-draft-status.py

@@ -0,0 +1,19 @@
+
+import os
+import json
+
+pipeline_label = os.environ['PIPELINE_LABEL']
+
+with open("unicodetools/data/ucd/dev/DerivedAge.txt", 'r') as f:
+    version = f.readline().strip().replace("# DerivedAge-", "").replace(".0.txt", "")
+
+if pipeline_label != "pipeline-" + version:
+    with open(os.environ['GITHUB_EVENT_PATH'], 'r') as f:
+        event = json.load(f)
+    print(event)
+    draft = event['pull_request']['draft']
+    if not draft:
+        print("::error title=PR must be draft::"
+              "PRs for character additions must be draft unless approved for "
+              "the upcoming version of Unicode.")
+        exit(1)

py/pipeline-workflow/check-l2-document.py

@@ -0,0 +1,14 @@
+
+import os
+import json
+import re
+
+with open(os.environ['GITHUB_EVENT_PATH'], 'r') as f:
+    event = json.load(f)
+print(event)
+pr_body = event['pull_request']['body']
+if not re.search(r"L2/\d\d-\d\d\d", pr_body):
+    print("::error title=Need proposal document::"
+          "PRs for character additions must include a link to an L2 document in"
+          " the PR description.")
+    exit(1)

py/pipeline-workflow/check-utc-decision.py

@@ -0,0 +1,14 @@
+
+import os
+import json
+import re
+
+with open(os.environ['GITHUB_EVENT_PATH'], 'r') as f:
+    event = json.load(f)
+print(event)
+pr_body = event['pull_request']['body']
+if not re.search(r"UTC-\d\d\d-[MC]\d", pr_body):
+    print("::error title=Need UTC decision::"
+          "PRs for approved or provisionally assigned must include a link to a "
+          "UTC decision.")
+    exit(1)

py/pipeline-workflow/compare-repertoire.py

@@ -0,0 +1,21 @@
+
+import os
+import json
+
+with open(os.environ['GITHUB_EVENT_PATH'], 'r') as f:
+    event = json.load(f)
+print(event)
+labels = event['pull_request']['labels']
+pipeline_labels = [label for label in labels if label['name'].startswith('pipeline-')]
+if not pipeline_labels:
+    print("::error title=Missing pipeline label::"
+          "PRs for character additions must have a pipeline label.")
+    exit(1)
+if len(pipeline_labels) > 1:
+    print("::error title=Multiple pipeline labels::"
+          "Only one pipeline-* label must be applied.")
+    exit(1)
+label = pipeline_labels[0]
+print("Labeled", label)
+with open(os.environ['GITHUB_OUTPUT'], 'a') as f:
+    print("pipeline-label=" + label['name'], file=f)