fix branch parsing script

Author: Astraea Quinn S

.github/workflows/integration-tests.yml

@@ -17,20 +17,63 @@ jobs:
         python-version: ["3.13"]
 
     steps:
-    - name: Parse testing SDK branch from PR body
+    - name: Parse testing SDK branch from PR body (Python)
       id: parse
+      shell: bash
       run: |
-        # Look for a line like: TESTING_SDK_BRANCH: feature/foo
-        REF=$(printf '%s\n' '${{ github.event.pull_request.body }}' | sed -n 's/^TESTING_SDK_BRANCH:[[:space:]]*//p' | head -n1)
-        if [ -z "$REF" ]; then REF="main"; fi
-        echo "testing_ref=$REF" >> "$GITHUB_OUTPUT"
-        echo "Using testing SDK branch: $REF"
-
+        set -euo pipefail
+    
+        python - << 'PYTHON'
+        import os
+        import re
+    
+        body = os.environ.get("PR_BODY", "") or ""
+        default_ref = "main"
+    
+        # Regex:
+        # - (?i)            case-insensitive
+        # - TESTING_SDK_BRANCH
+        # - optional spaces
+        # - : or =
+        # - optional spaces
+        # - capture rest of line
+        pattern = re.compile(r"(?i)TESTING_SDK_BRANCH\s*[:=]\s*(.+)$")
+    
+        ref = None
+        for line in body.splitlines():
+          m = pattern.search(line)
+          if m:
+            # Strip trailing whitespace
+            candidate = m.group(1).strip()
+            if candidate:
+              ref = candidate
+              break
+    
+        if not ref:
+          ref = default_ref
+    
+        github_output = os.environ.get("GITHUB_OUTPUT")
+        if not github_output:
+          raise SystemExit("GITHUB_OUTPUT not set")
+    
+        with open(github_output, "a", encoding="utf-8") as fh:
+          fh.write(f"testing_ref={ref}\n")
+    
+        print(f"Using testing SDK branch: {ref}")
+        PYTHON
+      env:
+        PR_BODY: ${{ github.event.pull_request.body }}
     - name: Checkout Language SDK (this PR)
       uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8  # v5.0.0
       with:
         path: language-sdk
 
+    - name: Parse testing SDK branch from PR body
+      id: parse
+      run: python ops/parse_sdk_branch.py
+      env:
+        PR_BODY: ${{ github.event.pull_request.body }}
+
     - name: Checkout Testing SDK
       uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8  # v5.0.0
       with:
@@ -69,20 +112,17 @@ jobs:
       AWS_REGION: us-west-2
 
     steps:
-    - name: Parse testing SDK branch from PR body
-      id: parse
-      run: |
-        # Look for a line like: TESTING_SDK_BRANCH: feature/foo
-        REF=$(printf '%s\n' '${{ github.event.pull_request.body }}' | sed -n 's/^TESTING_SDK_BRANCH:[[:space:]]*//p' | head -n1)
-        if [ -z "$REF" ]; then REF="main"; fi
-        echo "testing_ref=$REF" >> "$GITHUB_OUTPUT"
-        echo "Using testing SDK branch: $REF"
-
     - name: Checkout Language SDK (this PR)
       uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8  # v5.0.0
       with:
         path: language-sdk
 
+    - name: Parse testing SDK branch from PR body
+      id: parse
+      run: python ops/parse_sdk_branch.py
+      env:
+        PR_BODY: ${{ github.event.pull_request.body }}
+
     - name: Checkout Testing SDK
       uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8  # v5.0.0
       with:

.github/workflows/test-parser.yml

@@ -0,0 +1,21 @@
+name: Test Parser
+
+on:
+  pull_request:
+    paths:
+      - 'ops/parse_sdk_branch.py'
+      - 'ops/__tests__/**'
+  push:
+    branches: [ main ]
+    paths:
+      - 'ops/parse_sdk_branch.py'
+      - 'ops/__tests__/**'
+
+jobs:
+  test-parser:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8  # v5.0.0
+    
+    - name: Run parser tests
+      run: python ops/__tests__/test_parse_sdk_branch.py

ops/__tests__/test_parse_sdk_branch.py

@@ -0,0 +1,83 @@
+#!/usr/bin/env python3
+
+import sys
+import os
+sys.path.insert(0, os.path.dirname(os.path.dirname(__file__)))
+
+from parse_sdk_branch import parse_sdk_branch
+
+
+def test_parse_sdk_branch():
+    test_cases = [
+        # Basic cases
+        ("TESTING_SDK_BRANCH = feature/test", "feature/test"),
+        ("TESTING_SDK_BRANCH: feature/test", "feature/test"),
+        ("TESTING_SDK_BRANCH=feature/test", "feature/test"),
+        ("testing_sdk_branch: feature/test", "feature/test"),
+        
+        # Complex PR body with backticks and contractions
+        ("""Updated the script to safely parse the testing SDK branch from the PR body, handling case insensitivity and whitespace.
+
+The goal here is to fix the usage of backticks such as in `foo`, and contractions that we've been using such as `we've`
+
+```
+plus of course the usage of multiple backticks to include code
+```
+
+TESTING_SDK_BRANCH = main
+
+By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.""", "main"),
+        
+        # Edge cases with markdown and special characters
+        ("""# PR Title
+        
+Some `code` and we've got contractions here.
+
+```python
+def test():
+    return "test"
+```
+
+TESTING_SDK_BRANCH: feature/fix-backticks
+
+More text with `inline code` and don't forget contractions.""", "feature/fix-backticks"),
+        
+        # Multiple occurrences (should take first)
+        ("""TESTING_SDK_BRANCH = first-branch
+        
+Some text here.
+
+TESTING_SDK_BRANCH = second-branch""", "first-branch"),
+        
+        # Whitespace variations
+        ("   TESTING_SDK_BRANCH   =   feature/spaces   ", "feature/spaces"),
+        ("TESTING_SDK_BRANCH:feature/no-space", "feature/no-space"),
+        
+        # Default cases
+        ("No branch specified", "main"),
+        ("", "main"),
+        ("Just some random text", "main"),
+        
+        # Case with backticks in branch name
+        ("TESTING_SDK_BRANCH = feature/fix-`backticks`", "feature/fix-`backticks`"),
+        
+        # Case with contractions in surrounding text
+        ("We've updated this and TESTING_SDK_BRANCH = feature/test and we're done", "feature/test"),
+    ]
+    
+    for i, (input_text, expected) in enumerate(test_cases):
+        result = parse_sdk_branch(input_text)
+        if result != expected:
+            print(f"FAIL Test {i+1}: Expected '{expected}', got '{result}'")
+            print(f"Input: {repr(input_text[:100])}...")
+            return False
+        else:
+            print(f"PASS Test {i+1}: {expected}")
+    
+    print(f"\nAll {len(test_cases)} tests passed!")
+    return True
+
+
+if __name__ == "__main__":
+    success = test_parse_sdk_branch()
+    sys.exit(0 if success else 1)

ops/parse_sdk_branch.py

@@ -0,0 +1,34 @@
+#!/usr/bin/env python3
+
+import os
+import re
+
+
+
+def parse_sdk_branch(pr_body: str, default_ref: str = "main") -> str:
+    """Parse PR body for TESTING_SDK_BRANCH and return the branch reference."""
+    pattern = re.compile(r"(?i)TESTING_SDK_BRANCH\s*[:=]\s*(.+)$", re.MULTILINE)
+
+    match = pattern.search(pr_body)
+    if match:
+        ref = match.group(1).strip()
+        if ref:
+            return ref
+
+    return default_ref
+
+
+def main():
+    pr_body = os.environ.get("PR_BODY", "")
+    ref = parse_sdk_branch(pr_body)
+
+    github_output = os.environ.get("GITHUB_OUTPUT")
+    if github_output:
+        with open(github_output, "a", encoding="utf-8") as f:
+            f.write(f"testing_ref={ref}\n")
+
+    print(f"Using testing SDK branch: {ref}")
+
+
+if __name__ == "__main__":
+    main()