feat: add letu player parsing

2026-04-20 09:36:04 +08:00
parent 5aa57f5100
commit 15873e87dc
2 changed files with 96 additions and 1 deletions
@@ -1,6 +1,8 @@
 import unittest
+import base64
 from importlib.machinery import SourceFileLoader
 from pathlib import Path
+from urllib.parse import quote
 from unittest.mock import patch


@@ -135,6 +137,46 @@ class TestLeTuSpider(unittest.TestCase):
        self.spider.detailContent(["detail/demo"])
        self.assertEqual(mock_request_html.call_args.args[0], "https://www.letu.me/detail/demo.html")

+    @patch.object(Spider, "_request_html")
+    def test_player_content_returns_direct_json_url(self, mock_request_html):
+        mock_request_html.return_value = '{"code":200,"url":"https://video.example/direct.m3u8"}'
+        result = self.spider.playerContent("线路A", "play/123-1-1", {})
+        self.assertEqual(result["parse"], 0)
+        self.assertEqual(result["jx"], 0)
+        self.assertEqual(result["url"], "https://video.example/direct.m3u8")
+
+    @patch.object(Spider, "_request_html")
+    def test_player_content_decodes_rose_base64_url(self, mock_request_html):
+        encoded = quote(base64.b64encode(b"https://video.example/rose.m3u8").decode("utf-8"))
+        mock_request_html.return_value = '{"code":200,"url":"rose_' + encoded + '"}'
+        result = self.spider.playerContent("线路A", "play/123-1-1", {})
+        self.assertEqual(result["parse"], 0)
+        self.assertEqual(result["url"], "https://video.example/rose.m3u8")
+
+    def test_extract_player_data_reads_player_json(self):
+        html = '<script>var player_aaaa={"url":"https%3A%2F%2Fvideo.example%2Fenc1.m3u8","encrypt":"1"};</script>'
+        self.assertEqual(self.spider._extract_player_data(html)["encrypt"], "1")
+
+    @patch.object(Spider, "_request_html")
+    def test_player_content_supports_encrypt_1_and_encrypt_2(self, mock_request_html):
+        encoded = quote(base64.b64encode(b"https://video.example/enc2.m3u8").decode("utf-8"))
+        mock_request_html.side_effect = [
+            '<script>var player_aaaa={"url":"https%3A%2F%2Fvideo.example%2Fenc1.m3u8","encrypt":"1"};</script>',
+            '<script>var player_aaaa={"url":"' + encoded + '","encrypt":"2"};</script>',
+        ]
+        encrypt_1 = self.spider.playerContent("线路A", "play/123-1-1", {})
+        encrypt_2 = self.spider.playerContent("线路A", "play/123-1-2", {})
+        self.assertEqual(encrypt_1["url"], "https://video.example/enc1.m3u8")
+        self.assertEqual(encrypt_2["url"], "https://video.example/enc2.m3u8")
+
+    @patch.object(Spider, "_request_html")
+    def test_player_content_falls_back_to_system_parse(self, mock_request_html):
+        mock_request_html.return_value = "<html></html>"
+        result = self.spider.playerContent("线路A", "play/123-1-1", {})
+        self.assertEqual(result["parse"], 1)
+        self.assertEqual(result["jx"], 1)
+        self.assertEqual(result["url"], "https://www.letu.me/play/123-1-1.html")
+

 if __name__ == "__main__":
    unittest.main()
@@ -1,7 +1,9 @@
 # coding=utf-8
+import base64
+import json
 import re
 import sys
-from urllib.parse import quote, urljoin
+from urllib.parse import quote, unquote, urljoin

 from base.spider import Spider as BaseSpider

@@ -203,3 +205,54 @@ class Spider(BaseSpider):
            parsed = self._parse_detail_page(self._request_html(detail_url), vod_id)
            result["list"].extend(parsed.get("list", []))
        return result
+
+    def _base64_decode(self, value):
+        text = re.sub(r"\s+", "", str(value or ""))
+        text += "=" * ((4 - len(text) % 4) % 4)
+        try:
+            return base64.b64decode(text).decode("utf-8")
+        except Exception:
+            return ""
+
+    def _decode_rose_url(self, value):
+        raw = str(value or "")[5:]
+        decoded = self._base64_decode(unquote(raw))
+        if decoded:
+            return decoded
+        return self._base64_decode(raw)
+
+    def _extract_player_data(self, html):
+        matched = re.search(r"player_[a-z0-9]+\s*=\s*(\{[\s\S]*?\})", str(html or ""), re.I)
+        if not matched:
+            return None
+        try:
+            return json.loads(matched.group(1))
+        except Exception:
+            return None
+
+    def playerContent(self, flag, id, vipFlags):
+        play_url = self._decode_play_id(id) or self._build_url(id)
+        body = self._request_html(play_url)
+        try:
+            payload = json.loads(body)
+        except Exception:
+            payload = None
+        if isinstance(payload, dict) and str(payload.get("code")) == "200" and payload.get("url"):
+            video_url = str(payload.get("url"))
+            if video_url.startswith("rose_"):
+                video_url = self._decode_rose_url(video_url)
+            elif video_url.startswith("/"):
+                video_url = self._build_url(video_url)
+            if video_url:
+                return {"parse": 0, "jx": 0, "url": video_url, "header": dict(self.headers)}
+        player = self._extract_player_data(body)
+        if player:
+            video_url = str(player.get("url") or "")
+            encrypt = str(player.get("encrypt") or "")
+            if encrypt == "1":
+                video_url = unquote(video_url)
+            elif encrypt == "2":
+                video_url = self._base64_decode(unquote(video_url))
+            if video_url:
+                return {"parse": 0, "jx": 0, "url": video_url, "header": dict(self.headers)}
+        return {"parse": 1, "jx": 1, "url": play_url, "header": dict(self.headers)}