spark: add word counter utility with tests

wizard-work/spark/python_env_check.py

@@ -0,0 +1,69 @@
+#!/usr/bin/env python3
+"""Quick Python environment validator.
+Checks pip, site-packages accessibility, and common import paths.
+"""
+
+import sys
+import os
+import importlib.util
+
+def check_pip():
+    """Verify pip is accessible."""
+    try:
+        import pip
+        return True, "pip available"
+    except ImportError:
+        try:
+            import ensurepip
+            return True, "pip can be bootstrapped via ensurepip"
+        except ImportError:
+            return False, "pip not available and cannot bootstrap"
+
+def check_site_packages():
+    """Check site-packages directory exists and is writable."""
+    import site
+    paths = site.getsitepackages()
+    for p in paths:
+        if os.path.isdir(p):
+            writable = os.access(p, os.W_OK)
+            return True, f"site-packages found at {p}, writable={writable}"
+    return False, "no site-packages directory found"
+
+def check_stdlib_imports():
+    """Test critical stdlib modules."""
+    modules = ['os', 'sys', 'json', 'urllib.request', 'pathlib', 'typing']
+    failed = []
+    for mod in modules:
+        try:
+            importlib.import_module(mod)
+        except ImportError as e:
+            failed.append((mod, str(e)))
+    return len(failed) == 0, failed if failed else "all stdlib modules importable"
+
+def main():
+    print(f"Python environment check (py{sys.version_info.major}.{sys.version_info.minor})")
+    print(f"Executable: {sys.executable}")
+    print(f"Site-packages: {next((p for p in __import__('site').getsitepackages() if os.path.isdir(p)), 'none')}\n")
+    
+    checks = [
+        ("pip", check_pip),
+        ("site-packages", check_site_packages),
+        ("stdlib imports", check_stdlib_imports),
+    ]
+    
+    results = []
+    for name, func in checks:
+        passed, msg = func()
+        results.append((name, passed, msg))
+    
+    all_passed = all(p for _, p, _ in results)
+    
+    for name, passed, msg in results:
+        status = "PASS" if passed else "FAIL"
+        print(f"[{status}] {name}: {msg}")
+    
+    print(f"\n{'All checks passed!' if all_passed else 'Some checks failed.'}")
+    return 0 if all_passed else 1
+
+if __name__ == '__main__':
+    sys.exit(main())