utils.py

"""
utils.py - 通用工具函数
包含：序列清洗、FASTA 解析、安全浮点转换、网络会话创建
"""

import re
import math
import requests
from typing import Optional, Dict

AA_ALLOWED = set("ACDEFGHIKLMNPQRSTVWYX")

def clean_sequence(seq: str) -> str:
    """清洗序列：去除非字母字符，将非标准氨基酸替换为 X，转为大写"""
    return "".join((ch if ch in AA_ALLOWED else "X") for ch in seq.upper() if ch.isalpha())

def parse_fasta(text: str) -> Dict[str, str]:
    """
    解析 FASTA 文本，返回 {header: sequence} 字典。
    自动处理 BOM、Windows/Mac 换行符，忽略空序列。
    """
    # 移除 UTF-8 BOM
    if text.startswith('\ufeff'):
        text = text[1:]

    records = {}
    header = None
    seq_chunks = []

    # 统一换行符为 \n
    text = text.replace('\r\n', '\n').replace('\r', '\n')
    for line in text.splitlines():
        line = line.strip()
        if not line:
            continue
        if line.startswith(">"):
            if header is not None:
                seq = clean_sequence("".join(seq_chunks))
                if seq:  # 忽略空序列
                    records[header] = seq
            header = line[1:].strip()
            seq_chunks = []
        else:
            seq_chunks.append(line)
    # 处理最后一条
    if header is not None:
        seq = clean_sequence("".join(seq_chunks))
        if seq:
            records[header] = seq
    return records

def safe_float(s: str) -> Optional[float]:
    try:
        val = float(str(s).strip())
        if math.isnan(val) or math.isinf(val):
            return None
        return val
    except (ValueError, TypeError):
        return None

def create_session(timeout: int = 30) -> requests.Session:
    sess = requests.Session()
    sess.trust_env = False
    sess.headers.update({"User-Agent": "Mozilla/5.0 (compatible; LLPS-Predictor/2.0)"})
    sess.timeout = timeout
    return sess