重构：模块化 RRPP Lite 并更新入口

2026-09-16 23:13:28 +08:00
parent c37d8af821
commit 02d36c72f7
35 changed files with 2286 additions and 1079 deletions
@@ -0,0 +1,81 @@
+"""PAC 路由引擎使用的纯算法，不依赖文件、网络或 Qt。"""
+
+import ipaddress
+import re
+
+
+def merge_ip_ranges(cidrs, version):
+    """将指定 IP 版本的 CIDR 合并为整数闭区间，供二分查找。"""
+    ranges = []
+    for cidr in cidrs:
+        network = ipaddress.ip_network(str(cidr).strip(), strict=False)
+        if network.version == version:
+            ranges.append((int(network.network_address), int(network.broadcast_address)))
+    ranges.sort()
+    merged = []
+    for start, end in ranges:
+        if merged and start <= merged[-1][1] + 1:
+            if end > merged[-1][1]:
+                merged[-1] = (merged[-1][0], end)
+        else:
+            merged.append((start, end))
+    return merged
+
+
+def build_suffix_trie(suffixes):
+    """按反向域名标签建立 Trie，适合大规模后缀规则的 O(标签数) 匹配。"""
+    root = {}
+    for suffix in suffixes:
+        node = root
+        for label in reversed(suffix.split(".")):
+            node = node.setdefault(label, {})
+        node[""] = True
+    return root
+
+
+def suffix_trie_matches(host: str, trie) -> bool:
+    """判断主机名是否命中一个后缀 Trie。"""
+    node = trie
+    for label in reversed(host.rstrip(".").lower().split(".")):
+        node = node.get(label)
+        if node is None:
+            return False
+        if node.get(""):
+            return True
+    return bool(node and node.get(""))
+
+
+def parse_domain_rules(raw_rules, seed_suffixes=()):
+    """解析后缀、``full:`` 与 ``regexp:`` 三种 PAC 域名规则。
+
+    返回 ``(suffixes, exact, regexes, invalid_regex_count)``。调用方决定
+    这些规则是“强制代理”还是“明确直连”，从而避免解析层携带业务策略。
+    """
+    suffixes = set(seed_suffixes)
+    exact = set()
+    regexes = []
+    invalid_regexes = 0
+    for raw in raw_rules or ():
+        raw_rule = str(raw).strip()
+        rule = raw_rule.lower()
+        if not raw_rule:
+            continue
+        if rule.startswith("full:"):
+            host = rule[len("full:"):].rstrip(".")
+            if host:
+                exact.add(host)
+            continue
+        if rule.startswith("regexp:"):
+            pattern = raw_rule[len("regexp:"):]
+            if not pattern:
+                invalid_regexes += 1
+                continue
+            try:
+                regexes.append(re.compile(pattern, re.IGNORECASE))
+            except re.error:
+                invalid_regexes += 1
+            continue
+        suffix = rule.lstrip("*.").rstrip(".")
+        if suffix:
+            suffixes.add(suffix)
+    return suffixes, frozenset(exact), tuple(regexes), invalid_regexes