run_format.py

fromtypingimportList

importos.path
importre
importblack

suffixes= ["md", "py", "java", "c", "cpp", "go", "php", "cs", "rs", "js", "ts", "sql"]

code_blocks= [
"python",
"java",
"cpp",
"c",
"go",
"ts",
"js",
"php",
"cs",
"rust",
"sql",
]

functions_to_replace= [
"ABS",
"ACOS",
"ADDDATE",
"ADDTIME",
"AES_DECRYPT",
"AES_ENCRYPT",
"ASCII",
"ASIN",
"ATAN",
"AVG",
"BIN",
"BIT_COUNT",
"CEIL",
"CHAR",
"CHAR_LENGTH",
"CHARACTER_LENGTH",
"CONCAT",
"CONCAT_WS",
"CONNECTION_ID",
"CONV",
"CONVERT",
"COS",
"COT",
"COUNT",
"CRC32",
"CURDATE",
"CURRENT_DATE",
"CURRENT_TIME",
"CURRENT_TIMESTAMP",
"CURTIME",
"DATABASE",
"DATE",
"DATEDIFF",
"DATE_ADD",
"DATE_FORMAT",
"DATE_SUB",
"DAY",
"DAYNAME",
"DAYOFMONTH",
"DAYOFWEEK",
"DAYOFYEAR",
"DECODE",
"DEFAULT",
"DEGREES",
"DES_DECRYPT",
"DES_ENCRYPT",
"ELT",
"ENCODE",
"ENCRYPT",
"EXP",
"EXPORT_SET",
"EXTRACT",
"FIELD",
"FIND_IN_SET",
"FLOOR",
"FORMAT",
"FOUND_ROWS",
"FROM_DAYS",
"FROM_UNIXTIME",
"GET_FORMAT",
"GET_LOCK",
"GREATEST",
"GROUP_CONCAT",
"HEX",
"HOUR",
"IF",
"IFNULL",
"IN",
"INET_ATON",
"INET_NTOA",
"INSERT",
"INSTR",
"INTERVAL",
"ISNULL",
"LAST_INSERT_ID",
"LCASE",
"LEAST",
"LEFT",
"LENGTH",
"LN",
"LOAD_FILE",
"LOCALTIME",
"LOCALTIMESTAMP",
"LOCATE",
"LOG",
"LOG10",
"LOG2",
"LOWER",
"LPAD",
"LTRIM",
"MAKE_SET",
"MAKEDATE",
"MAKETIME",
"MATCH",
"MAX",
"MD5",
"MICROSECOND",
"MID",
"MIN",
"MINUTE",
"MOD",
"MONTH",
"MONTHNAME",
"NAME_CONST",
"NOW",
"NULLIF",
"OCT",
"OCTET_LENGTH",
"ORD",
"PASSWORD",
"PERIOD_ADD",
"PERIOD_DIFF",
"PI",
"POSITION",
"POW",
"POWER",
"PROCEDURE ANALYSE",
"QUARTER",
"QUOTE",
"RADIANS",
"RAND",
"RELEASE_LOCK",
"REPEAT",
"REPLACE",
"REVERSE",
"RIGHT",
"ROUND",
"ROW_COUNT",
"RPAD",
"RTRIM",
"SCHEMA",
"SEC_TO_TIME",
"SECOND",
"SESSION_USER",
"SHA1",
"SHA",
"SIGN",
"SIN",
"SLEEP",
"SOUNDEX",
"SPACE",
"SQRT",
"STR_TO_DATE",
"STRCMP",
"SUBDATE",
"SUBSTR",
"SUBSTRING",
"SUBSTRING_INDEX",
"SUBTIME",
"SUM",
"SYSDATE",
"SYSTEM_USER",
"TAN",
"TIME",
"TIMEDIFF",
"TIMESTAMP",
"TIMESTAMPADD",
"TIMESTAMPDIFF",
"TIME_FORMAT",
"TIME_TO_SEC",
"TO_DAYS",
"TRIM",
"TRUNCATE",
"UCASE",
"UNCOMPRESS",
"UNCOMPRESSED_LENGTH",
"UNHEX",
"UNIX_TIMESTAMP",
"UPPER",
"USER",
"UTC_DATE",
"UTC_TIME",
"UTC_TIMESTAMP",
"UUID",
"VAR_POP",
"VAR_SAMP",
"VARIANCE",
"VERSION",
"WEEK",
"WEEKDAY",
"WEEKOFYEAR",
"XOR",
"YEAR",
"YEARWEEK",
"ROW_NUMBER",
"RANK",
"DENSE_RANK",
"NTILE",
"LAG",
"LEAD",
"FIRST_VALUE",
"LAST_VALUE",
"CUME_DIST",
"PERCENT_RANK",
"PERCENTILE_CONT",
"PERCENTILE_DISC",
]


defadd_header(path: str):
"""Add header to php and go files"""
print(f"[add header] path: {path}")
withopen(path, "r", encoding="utf-8") asf:
content=f.read()
ifpath.endswith(".php"):
content="<?php\n"+content
elifpath.endswith(".go") and"sorting"notinpath:
content="package main\n"+content
elifpath.endswith(".sql"):
forfuncinfunctions_to_replace:
pattern=r"\b{}\s*\(".format(func)
content=re.sub(pattern, f"{func.upper()}(", content, flags=re.IGNORECASE)
else:
return
withopen(path, "w", encoding="utf-8") asf:
f.write(content)


defremove_header(path: str):
"""Remove header from php and go files"""
print(f"[remove header] path: {path}")
withopen(path, "r", encoding="utf-8") asf:
content=f.read()
ifpath.endswith(".php"):
content=content.rstrip()
content=content.replace("<?php\n", "")
elifpath.endswith(".go"):
content=content.rstrip()
if"sorting"notinpath:
content=content.replace("package main\n\n", "").replace(
"package main\n", ""
 )
else:
return
withopen(path, "w", encoding="utf-8") asf:
f.write(content)


deffind_all_paths() ->List[str]:
"""Find all paths of files with suffixes"""
paths= []
forroot, _, filesinos.walk(os.getcwd()):
forfileinfiles:
path=root+"/"+file
if"node_modules"inpathor"__pycache__"inpathor".git"inpath:
continue
ifany(path.endswith(f".{suf}") forsufinsuffixes):
paths.append(path)
returnpaths


defformat_inline_code(path: str):
"""Format inline code in .md file"""
ifnotpath.endswith(".md"):
return
withopen(path, "r", encoding="utf-8") asf:
content=f.read()
root=path[: path.rfind("/")]
forsufincode_blocks:
res=re.findall(f"```{suf}\n(.*?)```", content, re.S)
forblockinresor []:
# skip empty code block
ifnotblockornotblock.strip():
continue
ifsufin ["c", "cpp", "java", "go"]:
file=f"{root}/Solution2.{suf}"
withopen(file, "w", encoding="utf-8") asf:
f.write(block)
ifsuf=="go":
add_header(file)
os.system(f'gofmt -w "{file}"')
remove_header(file)
else:
os.system(f'npx clang-format -i --style=file "{file}"')
withopen(file, "r", encoding="utf-8") asf:
new_block=f.read()
ifnotnew_block.endswith("\n"):
new_block+="\n"
content=content.replace(block, new_block)
os.remove(file)
elifsuf=="python":
new_block=black.format_str(
block, mode=black.FileMode(string_normalization=False)
 )
content=content.replace(block, new_block)
elifsuf=="sql":
forfuncinfunctions_to_replace:
pattern=r"\b{}\s*\(".format(func)
new_block=re.sub(
pattern, f"{func.upper()}(", block, flags=re.IGNORECASE
 )
content=content.replace(block, new_block)
block=new_block

withopen(path, "w", encoding="utf-8") asf:
f.write(content)


defrun():
"""Start formatting"""
paths=find_all_paths()

forpathinpaths:
add_header(path)
ifany(path.endswith(suf) forsufin ["c", "cpp", "java"]):
# format with clang-format
os.system(f'npx clang-format -i --style=file "{path}"')

# format with prettier
os.system('npx prettier --write "**/*.{js,ts,php,sql,rs,md}"')

# format with gofmt
os.system("gofmt -w .")

forpathinpaths:
remove_header(path)
forpathinpaths:
format_inline_code(path)


if__name__=="__main__":
run()