spark: add word counter utility with tests

wizard-work/spark/json_diff.py

@@ -0,0 +1,72 @@
+#!/usr/bin/env python3
+"""json_diff.py - Compare two JSON files for structural equality.
+
+Usage: python3 json_diff.py <file1.json> <file2.json>
+
+Returns exit code 0 if JSON structures match (ignoring key order in objects),
+1 if they differ, 2 on error. Outputs the diff to stdout.
+"""
+
+import json
+import sys
+from pathlib import Path
+
+
+def normalize_json(data):
+    """Recursively normalize JSON: sort dict keys, convert lists."""
+    if isinstance(data, dict):
+        return {k: normalize_json(data[k]) for k in sorted(data.keys())}
+    elif isinstance(data, list):
+        return [normalize_json(item) for item in data]
+    else:
+        return data
+
+
+def json_equal(file1: str, file2: str) -> tuple:
+    """Compare two JSON files. Returns (equal: bool, diff: str, error: str|None)."""
+    try:
+        data1 = json.loads(Path(file1).read_text())
+        data2 = json.loads(Path(file2).read_text())
+        
+        norm1 = normalize_json(data1)
+        norm2 = normalize_json(data2)
+        
+        if norm1 == norm2:
+            return True, "", None
+        
+        diff_lines = []
+        diff_lines.append(f"Files differ:")
+        diff_lines.append(f"  {file1}: {json.dumps(norm1, indent=2)[:500]}")
+        diff_lines.append(f"  {file2}: {json.dumps(norm2, indent=2)[:500]}")
+        
+        return False, "\n".join(diff_lines), None
+        
+    except FileNotFoundError as e:
+        return False, "", f"File not found: {e.filename}"
+    except json.JSONDecodeError as e:
+        return False, "", f"JSON error in {e.doc}: line {e.lineno}"
+    except Exception as e:
+        return False, "", str(e)
+
+
+def main():
+    if len(sys.argv) != 3:
+        print("Usage: python3 json_diff.py <file1.json> <file2.json>", file=sys.stderr)
+        sys.exit(2)
+    
+    result, diff, error = json_equal(sys.argv[1], sys.argv[2])
+    
+    if error:
+        print(f"Error: {error}", file=sys.stderr)
+        sys.exit(2)
+    
+    if result:
+        print("STRUCTURALLY EQUAL")
+        sys.exit(0)
+    else:
+        print(diff)
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()