代码片段-牛津3000

import re
import sys

def groupByLevel(path):
    patterns = [f"{chr(x)}{y}" for x in range(ord('A'), ord('D')) for y in range(1,3)]
    results = results = {key: set() for key in patterns}

    with open(path, "r") as file:
        for line in file:
            r = re.match(r"^- \[[x ]\] (.*)", line)
            if r:
                line = r.groups()[0]
            for pattern in patterns:
                if line.find(" "+pattern) != -1:
                    results[pattern].add(line)
    
    with open("groupByLevel.txt", "w") as outfile:
        sum = 0
        for key in results:
            length = len(results[key])
            sum += length
            print(key, length)
            outfile.write("".join(["\n", "## ", key, "\n\n"]))
            for value in sorted(results[key]):
                outfile.write("".join(["- [ ] ", value]))
        print("sum", sum, "\n")

def groupByPart(path):
    patterns = [
        r'(?<=\s|/)n\.',
        r'(?<=\s|/)v\.',
        r'(?<=\s|/)adj\.',
        r'(?<=\s|/)adv\.',
        r'(?<=\s|/)prep\.',
        r'(?<=\s|/)conj\.',
        r'(?<=\s|/)pron\.',
        r'(?<=\s|/)det\.',
        r'(?<=\s|/)number',
    ]
    default_pattern = "*"
    results = {key: set() for key in patterns}
    results[default_pattern] = set()
    
    with open(path, "r") as file:
        for line in file:
            r = re.match(r"^- \[[x ]\] (.*)", line)
            if r:
                line = r.groups()[0]
            matched = False
            for pattern in patterns:
                if re.search(pattern, line):
                    matched = True
                    results[pattern].add(line)
            if not matched:
                print(line)
                results[default_pattern].add(line)

    with open("groupByPart.txt", "w") as outfile:
        sum = 0
        for key in results:
            length = len(results[key])
            sum += length
            key2 = key.split(")")[1].replace('\\', '') if key != default_pattern else key
            print(key2, length)
            outfile.write("".join(["\n", "## ", key2, "\n\n"]))
            for value in sorted(results[key]):
                outfile.write("".join(["- [ ] ", value]))
        print("sum", sum, "\n")

def filterChecked(path):
    res = set()
    with open(path, "r") as file:
        for line in file:
            r = re.match(r"^- \[x\] (\w+)(.*)", line)
            if r:
                line = r.groups()[0].strip()
                res.add(line)
    for line in sorted(res):
        print(line)

if __name__ == "__main__":
    filepath = sys.argv[1]
    groupByPart(filepath)
    groupByLevel(filepath)
    #filterChecked(filepath)