fix:优化数据来源

crawler/parser.py

@@ -0,0 +1,52 @@
+# -*- coding: utf-8 -*-
+"""新闻分类与严重度判定"""
+import re
+from typing import Literal
+
+Category = Literal["deployment", "alert", "intel", "diplomatic", "other"]
+Severity = Literal["low", "medium", "high", "critical"]
+
+# 分类关键词
+CAT_DEPLOYMENT = ["deploy", "carrier", "航母", "military build", "troop", "forces"]
+CAT_ALERT = ["strike", "attack", "fire", "blast", "hit", "爆炸", "袭击", "打击"]
+CAT_INTEL = ["satellite", "intel", "image", "surveillance", "卫星", "情报"]
+CAT_DIPLOMATIC = ["talk", "negotiation", "diplomat", "sanction", "谈判", "制裁"]
+
+
+def _match(text: str, words: list[str]) -> bool:
+    t = (text or "").lower()
+    for w in words:
+        if w.lower() in t:
+            return True
+    return False
+
+
+def classify(text: str) -> Category:
+    if _match(text, CAT_ALERT):
+        return "alert"
+    if _match(text, CAT_DEPLOYMENT):
+        return "deployment"
+    if _match(text, CAT_INTEL):
+        return "intel"
+    if _match(text, CAT_DIPLOMATIC):
+        return "diplomatic"
+    return "other"
+
+
+def severity(text: str, category: Category) -> Severity:
+    t = (text or "").lower()
+    critical = [
+        "nuclear", "核", "strike", "attack", "killed", "dead", "casualty",
+        "war", "invasion", "袭击", "打击", "死亡",
+    ]
+    high = [
+        "missile", "drone", "bomb", "explosion", "blasted", "fire",
+        "导弹", "无人机", "爆炸", "轰炸",
+    ]
+    if _match(t, critical):
+        return "critical"
+    if _match(t, high) or category == "alert":
+        return "high"
+    if category == "deployment":
+        return "medium"
+    return "low"