Add simple script to check hex content

Author: xiaozhikang0916

formats/protobuf/commonTest/src/hex_proto.py

@@ -0,0 +1,76 @@
+#!/usr/bin/env python3
+
+"""
+This script is used to add human-readable comments of hex strings in protoscope syntext.
+Before you start, make sure to install [protoscope](https://github.com/protocolbuffers/protoscope).
+
+NO WARRANTY. Double check before commit.
+
+Usage: hex_proto.py <path_to_kotlin_file>
+"""
+
+import re
+import subprocess
+import sys
+
+# Check input args
+if len(sys.argv) != 2:
+    print("Usage: python hex_proto.py <path_to_kotlin_file>")
+    sys.exit(1)
+
+file_path = sys.argv[1]
+
+# Regexp for hex string
+hex_string_pattern = re.compile(r'\"(0x)?([0-9a-fA-F]+)\"')
+
+# Regexp for existing comment
+existing_comment_pattern = re.compile(r'/\*\*.*?\*/', re.DOTALL)
+
+with open(file_path, 'r') as file:
+    lines = file.readlines()
+
+# Lines to be written back
+updated_lines = []
+
+inside_comment_block = False
+potential_comment_block = []
+
+for i, line in enumerate(lines):
+    # Check if we are inside a comment block
+    if line.strip().startswith("/*"):
+        inside_comment_block = True
+        potential_comment_block = [line]
+        continue
+    elif inside_comment_block:
+        potential_comment_block.append(line)
+        # Check if we are at the end of the comment block
+        if line.strip().endswith("*/"):
+            inside_comment_block = False
+            # Check if the next line is a hex string
+            if i + 1 < len(lines) and not hex_string_pattern.search(lines[i + 1]):
+                updated_lines.extend(potential_comment_block)
+            potential_comment_block = []
+        continue
+
+    hex_match = hex_string_pattern.search(line)
+    if hex_match:
+        indent = len(line) - len(line.lstrip())
+        indent_space = ' ' * indent
+
+        # Hex to protoscope
+        hex_string = hex_match.group(2)
+        binary_data = bytes.fromhex(hex_string)
+        result = subprocess.run(['protoscope', '-explicit-wire-types'], input=binary_data, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        if result.returncode == 0:
+            # Build comment string
+            comment = result.stdout.decode('utf-8').strip().replace('\n', f'\n{indent_space} * ')
+            new_comment_block = f'{indent_space}/**\n{indent_space} * {comment}\n{indent_space} */\n'
+            updated_lines.append(new_comment_block)
+        updated_lines.append(line)
+    else:
+        updated_lines.append(line)
+
+with open(file_path, 'w') as file:
+    file.writelines(updated_lines)
+
+print("Hex strings have been processed and comments have been added or replaced in the Kotlin file.")