import unittest from importlib.machinery import SourceFileLoader from pathlib import Path from unittest.mock import patch ROOT = Path(__file__).resolve().parents[1] MODULE = SourceFileLoader("erxiao_spider", str(ROOT / "二小.py")).load_module() Spider = MODULE.Spider class TestErXiaoSpider(unittest.TestCase): def setUp(self): Spider._instance = None self.spider = Spider() self.spider.init() def test_home_content_exposes_all_categories(self): content = self.spider.homeContent(False) self.assertEqual( [(item["type_id"], item["type_name"]) for item in content["class"]], [ ("1", "二小电影"), ("2", "二小剧集"), ("3", "二小动漫"), ("17", "二小综艺"), ], ) def test_home_video_content_returns_empty_list(self): self.assertEqual(self.spider.homeVideoContent(), {"list": []}) def test_build_url_and_detect_pan_type(self): self.assertEqual( self.spider._build_url("/index.php/vod/detail/id/1.html"), f"{self.spider.host}/index.php/vod/detail/id/1.html", ) self.assertEqual(self.spider._detect_pan_type("https://pan.baidu.com/s/demo"), ("baidu", "百度资源")) self.assertEqual(self.spider._detect_pan_type("https://pan.quark.cn/s/demo"), ("quark", "夸克资源")) self.assertEqual(self.spider._detect_pan_type("https://example.com/video"), ("", "")) def test_parse_cards_extracts_short_path_ids(self): html = """
示例影片
HD
2025
""" self.assertEqual( self.spider._parse_cards(html), [ { "vod_id": "/index.php/vod/detail/id/123.html", "vod_name": "示例影片", "vod_pic": f"{self.spider.host}/poster.jpg", "vod_remarks": "HD", "vod_year": "2025", } ], ) def test_parse_cards_ignores_module_items_wrapper_and_keeps_each_card_separate(self): html = """
影片一
HD
影片二
更新至2集
""" cards = self.spider._parse_cards(html) self.assertEqual( cards, [ { "vod_id": "/index.php/vod/detail/id/111.html", "vod_name": "影片一", "vod_pic": f"{self.spider.host}/poster-1.jpg", "vod_remarks": "HD", "vod_year": "", }, { "vod_id": "/index.php/vod/detail/id/222.html", "vod_name": "影片二", "vod_pic": f"{self.spider.host}/poster-2.jpg", "vod_remarks": "更新至2集", "vod_year": "", }, ], ) @patch.object(Spider, "_request_html") def test_category_content_builds_reference_url_and_returns_page_payload(self, mock_request_html): mock_request_html.return_value = """
分类影片
更新至10集
2024
""" result = self.spider.categoryContent("2", "3", False, {}) self.assertEqual( mock_request_html.call_args.args[0], f"{self.spider.host}/index.php/vod/show/id/2/page/3.html", ) self.assertEqual(result["page"], 3) self.assertEqual(result["limit"], 1) self.assertEqual(result["list"][0]["vod_name"], "分类影片") self.assertNotIn("pagecount", result) @patch.object(Spider, "_request_html") def test_search_content_builds_reference_search_url_and_parses_results(self, mock_request_html): mock_request_html.return_value = """
抢先版
搜索影片
""" result = self.spider.searchContent("繁花", False, "2") self.assertEqual( mock_request_html.call_args.args[0], f"{self.spider.host}/index.php/vod/search/page/2/wd/%E7%B9%81%E8%8A%B1.html", ) self.assertEqual( result["list"][0], { "vod_id": "/index.php/vod/detail/id/789.html", "vod_name": "搜索影片", "vod_pic": f"{self.spider.host}/search.jpg", "vod_remarks": "抢先版", }, ) def test_search_content_returns_empty_list_for_blank_keyword(self): self.assertEqual(self.spider.searchContent("", False, "1"), {"page": 1, "total": 0, "list": []}) def test_build_pan_lines_deduplicates_and_sorts_supported_links(self): detail = { "pan_urls": [ "https://pan.quark.cn/s/q1", "https://pan.baidu.com/s/b1", "https://pan.baidu.com/s/b1", "https://example.com/ignored", ] } self.assertEqual( self.spider._build_pan_lines(detail), [ ("baidu#二小", "百度资源$https://pan.baidu.com/s/b1"), ("quark#二小", "夸克资源$https://pan.quark.cn/s/q1"), ], ) def test_parse_detail_page_extracts_meta_content_and_pan_urls(self): html = """
示例剧
年代
2024
导演
导演甲
主演
演员甲演员乙
剧情

一段剧情简介

https://pan.quark.cn/s/q1

https://pan.baidu.com/s/b1

""" detail = self.spider._parse_detail_page("/index.php/vod/detail/id/123.html", html) self.assertEqual(detail["vod_name"], "示例剧") self.assertEqual(detail["vod_pic"], f"{self.spider.host}/poster.jpg") self.assertEqual(detail["vod_year"], "2024") self.assertEqual(detail["vod_director"], "导演甲") self.assertEqual(detail["vod_actor"], "演员甲,演员乙") self.assertEqual(detail["vod_content"], "一段剧情简介") self.assertEqual(detail["pan_urls"], ["https://pan.quark.cn/s/q1", "https://pan.baidu.com/s/b1"]) def test_parse_detail_page_uses_only_first_mobile_play_poster(self): html = """
示例剧

https://pan.baidu.com/s/b1

""" detail = self.spider._parse_detail_page("/index.php/vod/detail/id/123.html", html) self.assertEqual(detail["vod_pic"], f"{self.spider.host}/poster-main.jpg") @patch.object(Spider, "_request_html") def test_detail_content_builds_pan_play_fields(self, mock_request_html): mock_request_html.return_value = """
示例剧
年代
2024
导演
导演甲
主演
演员甲
剧情

一段剧情简介

https://pan.quark.cn/s/q1

https://pan.baidu.com/s/b1

""" result = self.spider.detailContent(["/index.php/vod/detail/id/123.html"]) vod = result["list"][0] self.assertEqual(vod["vod_name"], "示例剧") self.assertEqual(vod["vod_play_from"], "baidu#二小$$$quark#二小") self.assertEqual( vod["vod_play_url"], "百度资源$https://pan.baidu.com/s/b1$$$夸克资源$https://pan.quark.cn/s/q1", ) def test_player_content_passthroughs_supported_pan_urls(self): self.assertEqual( self.spider.playerContent("baidu#二小", "https://pan.baidu.com/s/demo", {}), {"parse": 0, "playUrl": "", "url": "https://pan.baidu.com/s/demo"}, ) def test_player_content_rejects_non_pan_url(self): self.assertEqual( self.spider.playerContent("site", "/index.php/vod/play/id/1.html", {}), {"parse": 0, "playUrl": "", "url": ""}, ) if __name__ == "__main__": unittest.main()