feat(ci): add prod CI workflows for test, manifest PR, and release pr…

.github/actions/detect-changes/action.yml

@@ -0,0 +1,158 @@
+name: Detect CV changes
+description: Detect bump type, fetch registry version, compute next version, generate changelog
+
+inputs:
+  base_sha:
+    description: Base SHA to diff against
+    required: true
+  github_token:
+    description: GitHub token for registry API calls
+    required: true
+  fetch_universe:
+    description: Whether to fetch universe version from registry
+    required: false
+    default: "false"
+
+outputs:
+  bump_type:
+    description: "none, patch, minor, or major"
+    value: ${{ steps.bump.outputs.bump_type }}
+  warning:
+    description: Warning message (e.g. for spec changes)
+    value: ${{ steps.bump.outputs.warning }}
+  current_version:
+    description: Current version from registry
+    value: ${{ steps.registry.outputs.current_version }}
+  next_version:
+    description: Computed next version
+    value: ${{ steps.next.outputs.next_version }}
+  universe_version:
+    description: Universe version from registry (only if fetch_universe=true)
+    value: ${{ steps.registry.outputs.universe_version }}
+  changelog:
+    description: Markdown changelog
+    value: ${{ steps.changelog.outputs.changelog }}
+
+runs:
+  using: composite
+  steps:
+    # ---------------------------------------------------------------
+    # 1. Detect bump type from diff
+    # ---------------------------------------------------------------
+    - name: Detect bump type
+      id: bump
+      shell: bash
+      env:
+        BASE_SHA: ${{ inputs.base_sha }}
+      run: |
+        set -e
+
+        echo "Comparing ${BASE_SHA}..HEAD"
+
+        CHANGED=$(git diff --name-only --diff-filter=ACDMR "${BASE_SHA}..HEAD" -- \
+          ':(exclude)_src' ':(exclude)_tests' ':(exclude)_archive' ':(exclude)scripts' \
+          ':(exclude)*.md' ':(exclude)*.toml' ':(exclude)*.lock' ':(exclude)*.py' \
+          ':(exclude).github' ':(exclude)LICENSE*' ':(exclude).gitignore' \
+          ':(exclude).pre-commit-config.yaml' ':(exclude)esgvoc_manifest.yaml' || true)
+
+        if [ -z "${CHANGED}" ]; then
+          echo "No CV files changed"
+          echo "bump_type=none" >> "$GITHUB_OUTPUT"
+          echo "warning=" >> "$GITHUB_OUTPUT"
+          exit 0
+        fi
+
+        echo "Changed files:"
+        echo "${CHANGED}"
+
+        BUMP="patch"
+        WARN=""
+
+        # New collections → major
+        NEW_CONTEXTS=$(git diff --name-only --diff-filter=A "${BASE_SHA}..HEAD" -- '*/000_context.jsonld' || true)
+        if [ -n "${NEW_CONTEXTS}" ]; then
+          BUMP="major"
+        fi
+
+        # Spec changes → minor minimum
+        if [ "${BUMP}" != "major" ]; then
+          SPEC_CHANGES=$(git diff --name-only --diff-filter=ACDMR "${BASE_SHA}..HEAD" -- '*_specs.yaml' || true)
+          if [ -n "${SPEC_CHANGES}" ]; then
+            BUMP="minor"
+            WARN="Spec files changed — at minimum a minor bump, could warrant major if it breaks downstream tooling."
+          fi
+        fi
+
+        # Modified collection models → minor
+        if [ "${BUMP}" = "patch" ]; then
+          MOD_CONTEXTS=$(git diff --name-only --diff-filter=M "${BASE_SHA}..HEAD" -- '*/000_context.jsonld' || true)
+          if [ -n "${MOD_CONTEXTS}" ]; then
+            BUMP="minor"
+          fi
+        fi
+
+        echo "bump_type=${BUMP}" >> "$GITHUB_OUTPUT"
+        echo "warning=${WARN}" >> "$GITHUB_OUTPUT"
+        echo "Detected bump type: ${BUMP}"
+
+    # ---------------------------------------------------------------
+    # 2. Fetch current version from registry
+    # ---------------------------------------------------------------
+    - name: Fetch registry version
+      if: steps.bump.outputs.bump_type != 'none'
+      id: registry
+      shell: bash
+      env:
+        GH_TOKEN: ${{ inputs.github_token }}
+        FETCH_UNIVERSE: ${{ inputs.fetch_universe }}
+      run: |
+        set -e
+
+        CURRENT=$(gh api repos/WCRP-CMIP/esgvoc_registry/contents/obs4ref.json \
+          --jq '.content' | base64 -d | python3 .github/scripts/ci_version.py fetch-current)
+
+        echo "current_version=${CURRENT}" >> "$GITHUB_OUTPUT"
+        echo "Current registry version: ${CURRENT}"
+
+        if [ "${FETCH_UNIVERSE}" = "true" ]; then
+          UNIVERSE=$(gh api repos/WCRP-CMIP/esgvoc_registry/contents/universe.json \
+            --jq '.content' | base64 -d | python3 .github/scripts/ci_version.py fetch-universe)
+          echo "universe_version=${UNIVERSE}" >> "$GITHUB_OUTPUT"
+          echo "Universe version: ${UNIVERSE}"
+        fi
+
+    # ---------------------------------------------------------------
+    # 3. Compute next version
+    # ---------------------------------------------------------------
+    - name: Compute next version
+      if: steps.bump.outputs.bump_type != 'none'
+      id: next
+      shell: bash
+      env:
+        CURRENT: ${{ steps.registry.outputs.current_version }}
+        BUMP: ${{ steps.bump.outputs.bump_type }}
+      run: |
+        NEXT=$(python3 .github/scripts/ci_version.py compute-next --current "${CURRENT}" --bump "${BUMP}")
+        echo "next_version=${NEXT}" >> "$GITHUB_OUTPUT"
+        echo "Next version: ${NEXT}"
+
+    # ---------------------------------------------------------------
+    # 4. Generate changelog
+    # ---------------------------------------------------------------
+    - name: Generate changelog
+      if: steps.bump.outputs.bump_type != 'none'
+      id: changelog
+      shell: bash
+      env:
+        BASE_SHA: ${{ inputs.base_sha }}
+      run: |
+        CHANGELOG=$(python3 .github/scripts/ci_changelog.py "${BASE_SHA}")
+
+        {
+          echo "changelog<<CHANGELOG_EOF"
+          echo "${CHANGELOG}"
+          echo "CHANGELOG_EOF"
+        } >> "$GITHUB_OUTPUT"
+
+        echo "Changelog:"
+        echo "${CHANGELOG}"

.github/scripts/ci_changelog.py

@@ -0,0 +1,130 @@
+"""
+CI helper: generate a markdown changelog from git diff.
+
+Usage:
+  python .github/scripts/ci_changelog.py <base_sha>
+"""
+
+import argparse
+import collections
+import subprocess
+import sys
+
+# Paths excluded from CV change detection (infrastructure, docs, etc.)
+EXCLUDE_PATTERNS = [
+    ":(exclude)_src",
+    ":(exclude)_tests",
+    ":(exclude)_archive",
+    ":(exclude)scripts",
+    ":(exclude)*.md",
+    ":(exclude)*.toml",
+    ":(exclude)*.lock",
+    ":(exclude)*.py",
+    ":(exclude).github",
+    ":(exclude)LICENSE*",
+    ":(exclude).gitignore",
+    ":(exclude).pre-commit-config.yaml",
+    ":(exclude)esgvoc_manifest.yaml",
+]
+
+
+def get_diff(base_sha):
+    """Run git diff and return the name-status output."""
+    cmd = [
+        "git", "diff", "--name-status", "--diff-filter=ACDMR",
+        f"{base_sha}..HEAD", "--",
+    ] + EXCLUDE_PATTERNS
+
+    result = subprocess.run(cmd, capture_output=True, text=True, check=True)
+    return result.stdout.strip()
+
+
+def parse_diff(diff_output):
+    """Parse git diff name-status output into categorized changes."""
+    changes = collections.defaultdict(lambda: {"A": [], "M": [], "D": []})
+    model_changes = []
+    spec_changes = []
+
+    for line in diff_output.split("\n"):
+        if not line.strip():
+            continue
+
+        parts = line.split("\t")
+
+        if parts[0].startswith(("R", "C")):
+            # Renames/copies: "R100\told/path\tnew/path" — treat as add of the new path
+            if len(parts) < 3:
+                continue
+            status = "A"
+            path = parts[2]
+        elif len(parts) >= 2:
+            status = parts[0]
+            path = parts[1]
+        else:
+            continue
+
+        segments = path.split("/")
+        filename = segments[-1]
+
+        # Spec files at root level
+        if len(segments) == 1 and filename.endswith("_specs.yaml"):
+            action = "new" if status == "A" else "updated"
+            spec_changes.append((filename, action))
+            continue
+
+        if len(segments) < 2:
+            continue
+
+        collection = segments[0]
+
+        if filename == "000_context.jsonld":
+            action = "new" if status == "A" else "updated"
+            model_changes.append((collection, action))
+        else:
+            term_id = filename.replace(".json", "")
+            changes[collection][status[0]].append(term_id)
+
+    return spec_changes, model_changes, changes
+
+
+def format_changelog(spec_changes, model_changes, term_changes):
+    """Format changes into markdown."""
+    lines = []
+
+    for filename, action in spec_changes:
+        lines.append(f"- {filename}: {action}")
+
+    for collection, action in model_changes:
+        lines.append(f"- {collection}: {action} collection model")
+
+    for collection in sorted(term_changes):
+        parts = []
+        counts = term_changes[collection]
+        if counts["A"]:
+            parts.append(f'{len(counts["A"])} added')
+        if counts["M"]:
+            parts.append(f'{len(counts["M"])} modified')
+        if counts["D"]:
+            parts.append(f'{len(counts["D"])} removed')
+        if parts:
+            lines.append(f"- {collection}: {', '.join(parts)}")
+
+    return "\n".join(lines) if lines else "No CV changes."
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Generate changelog from git diff")
+    parser.add_argument("base_sha", help="Base SHA to diff against")
+    args = parser.parse_args()
+
+    diff_output = get_diff(args.base_sha)
+    if not diff_output:
+        print("No CV changes.")
+        sys.exit(0)
+
+    spec_changes, model_changes, term_changes = parse_diff(diff_output)
+    print(format_changelog(spec_changes, model_changes, term_changes))
+
+
+if __name__ == "__main__":
+    main()

.github/scripts/ci_update_manifest.py

@@ -0,0 +1,63 @@
+"""
+CI helper: update esgvoc_manifest.yaml with new version info.
+
+Usage:
+  python .github/scripts/ci_update_manifest.py \
+    --version 1.2.0 \
+    --universe-version 1.0.3 \
+    --changelog-file changelog.md
+"""
+
+import argparse
+import sys
+
+import yaml
+
+
+def _literal_str_representer(dumper, data):
+    """Force multi-line strings to use YAML literal block scalar (|)."""
+    if "\n" in data:
+        return dumper.represent_scalar("tag:yaml.org,2002:str", data, style="|")
+    return dumper.represent_scalar("tag:yaml.org,2002:str", data)
+
+
+yaml.add_representer(str, _literal_str_representer)
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Update esgvoc manifest")
+    parser.add_argument("--version", required=True, help="New CV version")
+    parser.add_argument("--universe-version", required=True, help="Universe version")
+    parser.add_argument(
+        "--changelog-file",
+        default="-",
+        help="Path to changelog file, or - for stdin (default: -)",
+    )
+    parser.add_argument(
+        "--manifest",
+        default="esgvoc_manifest.yaml",
+        help="Path to manifest file (default: esgvoc_manifest.yaml)",
+    )
+    args = parser.parse_args()
+
+    if args.changelog_file == "-":
+        changelog = sys.stdin.read().strip()
+    else:
+        with open(args.changelog_file) as f:
+            changelog = f.read().strip()
+
+    with open(args.manifest) as f:
+        manifest = yaml.safe_load(f)
+
+    manifest["cv_version"] = args.version
+    manifest["universe_version"] = args.universe_version
+    manifest["release_notes"] = changelog + "\n"
+
+    with open(args.manifest, "w") as f:
+        yaml.dump(manifest, f, default_flow_style=False, sort_keys=False, allow_unicode=True)
+
+    print(f"Updated manifest to version {args.version}")
+
+
+if __name__ == "__main__":
+    main()