# 为oss内的文件批量生成可访问链接
import os
import re
import csv
import oss2
import pandas as pd
import configparser
import oss2

BUCKET = "lsh-oss-sns-offline-data"
PREFIX = "user/dawo/ai_search/songping_0302top300_model_903"
EXPIRE = 8640000

cfg_path = os.path.expanduser("~/.ossutilconfig")
cfg = configparser.RawConfigParser()
cfg.read(cfg_path, encoding="utf-8")

AK = cfg.get("Credentials", "accessKeyID")
SK = cfg.get("Credentials", "accessKeySecret")
endpoint = cfg.get("Credentials", "endpoint")
ENDPOINT = endpoint if endpoint.startswith("http") else "https://" + endpoint

auth = oss2.Auth(AK, SK)
bucket = oss2.Bucket(auth, ENDPOINT, BUCKET)

def leading_index(filename: str):
    """
    从文件名提取前缀编号：如 '0_xxx.png' -> 0
    提取不到返回 None
    """
    m = re.match(r"^(\d+)", filename)
    return int(m.group(1)) if m else None

rows = []
for obj in oss2.ObjectIterator(bucket, prefix=PREFIX):
    if obj.key.endswith("/"):
        continue
    filename = obj.key.split("/")[-1]
    idx = leading_index(filename)
    url = bucket.sign_url("GET", obj.key, EXPIRE)
    rows.append({"idx": idx, "filename": filename, "key": obj.key, "signed_url": url})

# 先按编号排序（从 0 开始），无法解析编号的放最后，再按文件名排
df = pd.DataFrame(rows)
df = df.sort_values(
    by=["idx", "filename"],
    key=lambda s: s.fillna(10**18) if s.name == "idx" else s,
    kind="mergesort"  # 稳定排序
).reset_index(drop=True)

# 可选：检查是否从0开始连续（不连续只提示，不中断）
idx_list = df["idx"].dropna().astype(int).tolist()
if idx_list:
    if idx_list[0] != 0:
        print(f"Warning: 最小编号不是0，而是 {idx_list[0]}")
    # 连续性检查
    missing = sorted(set(range(idx_list[0], idx_list[-1] + 1)) - set(idx_list))
    if missing:
        print(f"Warning: 编号不连续，缺失: {missing[:50]}" + ("..." if len(missing) > 50 else ""))

# 导出 CSV（按你要的列顺序）
out_csv = "signed_urls.csv"
df[["idx", "filename", "key", "signed_url"]].to_csv(out_csv, index=False, encoding="utf-8-sig")

print(f"done -> {out_csv}, rows={len(df)}")