update submission validation

.github/workflows/validate-community-submission.yaml

@@ -30,11 +30,17 @@ jobs:
           ${{ github.event.issue.body }}
           ISSUE_BODY_EOF
 
+      - name: Save issue body to file
+        run: |
+          cat << 'ISSUE_BODY_EOF' > /tmp/issue_body.txt
+          ${{ github.event.issue.body }}
+          ISSUE_BODY_EOF
+
       - name: Validate submission
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           GITHUB_REPOSITORY: ${{ github.repository }}
         run: |
           python -m src.contributions.validate_submission \
-            --issue-body "${{ github.event.issue.body }}" \
+            --issue-body-file /tmp/issue_body.txt \
             --issue-number ${{ github.event.issue.number }}

src/contributions/schema.py

@@ -54,7 +54,9 @@ def extract_json_from_issue_body(body: str) -> str | None:
     """
     Extract JSON from GitHub issue body.
     
-    Looks for JSON in the 'Submission JSON' section wrapped in code blocks.
+    Looks for JSON in the 'Submission JSON' section, either:
+    - Wrapped in code blocks (```json ... ``` or ``` ... ```)
+    - Or raw JSON after the header
     
     Args:
         body: The issue body text
@@ -62,13 +64,28 @@ def extract_json_from_issue_body(body: str) -> str | None:
     Returns:
         Extracted JSON string or None if not found
     """
-    # Match JSON in "### Submission JSON" section
+    # Try: JSON in code blocks after "### Submission JSON"
-    pattern = r"### Submission JSON\s*\n\s*```(?:json)?\s*\n([\s\S]*?)\n\s*```"
+    pattern_codeblock = r"### Submission JSON\s*\n\s*```(?:json)?\s*\n([\s\S]*?)\n\s*```"
-    match = re.search(pattern, body)
+    match = re.search(pattern_codeblock, body)
-    
     if match:
         return match.group(1).strip()
     
+    # Try: Raw JSON after "### Submission JSON" until next section or end
+    pattern_raw = r"### Submission JSON\s*\n\s*([\[{][\s\S]*?[\]}])(?=\n###|\n\n###|$)"
+    match = re.search(pattern_raw, body)
+    if match:
+        return match.group(1).strip()
+    
+    # Try: Any JSON object/array in the body (fallback)
+    pattern_any = r"([\[{][\s\S]*?[\]}])"
+    for match in re.finditer(pattern_any, body):
+        candidate = match.group(1).strip()
+        # Validate it looks like JSON
+        if candidate.startswith('{') and candidate.endswith('}'):
+            return candidate
+        if candidate.startswith('[') and candidate.endswith(']'):
+            return candidate
+    
     return None
 
 

src/contributions/validate_submission.py

@@ -7,6 +7,7 @@ submissions when issues are opened or edited.
 
 Usage:
     python -m src.contributions.validate_submission --issue-body "..."
+    python -m src.contributions.validate_submission --issue-body-file /path/to/body.txt
     python -m src.contributions.validate_submission --file submission.json
     echo '{"registration_number": "N12345"}' | python -m src.contributions.validate_submission --stdin
     
@@ -106,6 +107,7 @@ def main():
     parser = argparse.ArgumentParser(description="Validate community submission JSON")
     source_group = parser.add_mutually_exclusive_group(required=True)
     source_group.add_argument("--issue-body", help="Issue body text containing JSON")
+    source_group.add_argument("--issue-body-file", help="File containing issue body text")
     source_group.add_argument("--file", help="JSON file to validate")
     source_group.add_argument("--stdin", action="store_true", help="Read JSON from stdin")
     
@@ -125,6 +127,20 @@ def main():
                     "Please ensure your JSON is in the 'Submission JSON' field wrapped in code blocks."
                 )
             sys.exit(1)
+    elif args.issue_body_file:
+        with open(args.issue_body_file) as f:
+            issue_body = f.read()
+        json_str = extract_json_from_issue_body(issue_body)
+        if not json_str:
+            print("❌ Could not extract JSON from issue body", file=sys.stderr)
+            print(f"Issue body:\n{issue_body}", file=sys.stderr)
+            if args.issue_number:
+                add_issue_comment(
+                    args.issue_number,
+                    "❌ **Validation Failed**\n\nCould not extract JSON from submission. "
+                    "Please ensure your JSON is in the 'Submission JSON' field."
+                )
+            sys.exit(1)
     elif args.file:
         with open(args.file) as f:
             json_str = f.read()