| import json | |
| from pydantic import BaseModel | |
| class Sample(BaseModel): | |
| code: str | |
| size: int | |
| license: str | |
| def is_chinese_char(char): | |
| """Check if the char is a Chinese character.""" | |
| return char >= "\u4e00" and char <= "\u9fff" | |
| def is_chinese(text: str): | |
| """Check if the text is Chinese. We consider a text as Chinese if >20% of | |
| the text is Chinese characters.""" | |
| cn_count = 0 | |
| for char in text: | |
| if is_chinese_char(char): | |
| cn_count += 1 | |
| return cn_count * 5 > len(text) | |
| with open("train.jsonl", "wb") as cn: | |
| with open("/Users/diegorojas/Desktop/markdown.jsonl", "r") as samples: | |
| for sample in samples: | |
| try: | |
| data = json.loads(sample) | |
| except json.JSONDecodeError as e: | |
| print(f"Error: {e}") | |
| continue | |
| code, size, license = ( | |
| data["code"], | |
| data["size"], | |
| data["license"], | |
| ) | |
| if license not in ["", "MIT", "Apache-2.0"]: | |
| continue | |
| encoded = Sample( | |
| code=code, | |
| size=size, | |
| license=license, | |
| ).model_dump_json() | |
| language = is_chinese(code) | |
| if language == "cn": | |
| cn.write(encoded.encode("utf-8") + b"\n") | |