修复货币数字错误和没有货币种类问题，修复UTC和本地时间混用，改用UTC

clean_to_structured.py

@@ -119,7 +119,7 @@ def load_mysql_config() -> dict:
 
 
 def connect_mysql(cfg: dict):
-    return pymysql.connect(
+    conn = pymysql.connect(
         host=cfg["host"],
         port=cfg["port"],
         user=cfg["user"],
@@ -128,6 +128,9 @@ def connect_mysql(cfg: dict):
         charset=cfg["charset"],
         autocommit=True,
     )
+    with conn.cursor() as cur:
+        cur.execute("SET time_zone = '+00:00'")
+    return conn
 
 
 def init_target_db(conn):
@@ -221,16 +224,62 @@ def clean_company_name(s: str | None) -> str | None:
     return s or None
 
 
-def parse_salary(raw: str | None) -> tuple[str | None, int | None, int | None, str | None]:
+def infer_salary_currency(text: str) -> str | None:
+    low = text.lower()
+    if any(k in low for k in ["usd", "us$", "dollar"]) or "$" in text:
+        return "USD"
+    if any(k in text for k in ["¥", "￥", "人民币"]) or "cny" in low:
+        return "CNY"
+    if "元" in text or "万" in text:
+        return "CNY"
+    if "k" in low and any(k in low for k in ["month", "year", "day"]):
+        return "USD"
+    if "eur" in low or "€" in text:
+        return "EUR"
+    if "hkd" in low or "hk$" in low:
+        return "HKD"
+    if "sgd" in low or "s$" in low:
+        return "SGD"
+
+    # No explicit marker: infer by language.
+    if re.search(r"[\u4e00-\u9fff]", text):
+        return "CNY"
+    return "USD"
+
+
+def parse_salary(
+    raw: str | None,
+) -> tuple[str | None, str | None, int | None, int | None, str | None]:
     if not raw:
-        return None, None, None, None
+        return None, None, None, None, None
 
     text = clean_md_text(raw)
     lower = text.lower()
+    currency = infer_salary_currency(text)
 
-    nums = re.findall(r"\d+(?:\.\d+)?", text.replace(",", ""))
-    salary_min = int(float(nums[0])) if len(nums) >= 1 else None
-    salary_max = int(float(nums[1])) if len(nums) >= 2 else None
+    num_tokens = re.findall(r"(\d+(?:\.\d+)?)\s*([kKwW万]?)", text.replace(",", ""))
+    salary_min = None
+    salary_max = None
+    if num_tokens:
+        vals: list[tuple[float, str]] = [(float(n), u) for n, u in num_tokens]
+        if len(vals) >= 2:
+            u1 = vals[0][1]
+            u2 = vals[1][1]
+            if not u1 and u2:
+                vals[0] = (vals[0][0], u2)
+            if not u2 and u1:
+                vals[1] = (vals[1][0], u1)
+
+        def scaled(v: float, unit: str) -> int:
+            m = 1
+            if unit in ("k", "K"):
+                m = 1000
+            elif unit in ("w", "W", "万"):
+                m = 10000
+            return int(v * m)
+
+        salary_min = scaled(vals[0][0], vals[0][1]) if len(vals) >= 1 else None
+        salary_max = scaled(vals[1][0], vals[1][1]) if len(vals) >= 2 else None
 
     period = None
     if "month" in lower or "每月" in text or "月" in text:
@@ -240,7 +289,7 @@ def parse_salary(raw: str | None) -> tuple[str | None, int | None, int | None, s
     elif "day" in lower or "日" in text:
         period = "day"
 
-    return text, salary_min, salary_max, period
+    return text, currency, salary_min, salary_max, period
 
 
 def strip_meta_lines(content: str) -> str:
@@ -570,8 +619,9 @@ def parse_dejob_official(
             salary_line = ln
             break
 
-    salary_raw, salary_min, salary_max, salary_period = parse_salary(salary_line)
-    salary_currency = "USD" if salary_raw and "$" in salary_raw else None
+    salary_raw, salary_currency, salary_min, salary_max, salary_period = parse_salary(
+        salary_line
+    )
 
     responsibilities = extract_list_section(body_text, "岗位职责")
     requirements = extract_list_section(body_text, "岗位要求")
@@ -643,8 +693,9 @@ def parse_generic(
             salary_line = ln
             break
 
-    salary_raw, salary_min, salary_max, salary_period = parse_salary(salary_line)
-    salary_currency = "USD" if salary_raw and "$" in salary_raw else None
+    salary_raw, salary_currency, salary_min, salary_max, salary_period = parse_salary(
+        salary_line
+    )
 
     job_type = "招聘" if ("招聘" in body_text or "recruit" in body_text.lower()) else None
     (
@@ -770,8 +821,9 @@ def parse_dejob_global(
         if "薪酬" in ln or "salary" in ln.lower():
             salary_line = ln
             break
-    salary_raw, salary_min, salary_max, salary_period = parse_salary(salary_line)
-    salary_currency = "USD" if salary_raw and "$" in salary_raw else None
+    salary_raw, salary_currency, salary_min, salary_max, salary_period = parse_salary(
+        salary_line
+    )
 
     responsibilities = extract_list_section(body_text, "岗位职责") or extract_list_section(
         body_text, "Responsibilities"
@@ -849,8 +901,9 @@ def parse_remote_cn(
         if ln.startswith("摘要:"):
             summary_line = ln
             break
-    salary_raw, salary_min, salary_max, salary_period = parse_salary(summary_line)
-    salary_currency = "USD" if salary_raw and "$" in salary_raw else None
+    salary_raw, salary_currency, salary_min, salary_max, salary_period = parse_salary(
+        summary_line
+    )
 
     urls = extract_urls(body_text)
     apply_email = extract_apply_email(body_text)
@@ -935,8 +988,9 @@ def parse_cryptojobslist_source(
         if any(k in ln.lower() for k in ("salary", "$", "usd")):
             salary_line = ln
             break
-    salary_raw, salary_min, salary_max, salary_period = parse_salary(salary_line)
-    salary_currency = "USD" if salary_raw and "$" in salary_raw else None
+    salary_raw, salary_currency, salary_min, salary_max, salary_period = parse_salary(
+        salary_line
+    )
 
     apply_email = extract_apply_email(body_text)
     apply_tg = extract_apply_telegram(body_text)