package com.github.catvod.spider; import android.content.Context; import android.text.TextUtils; import android.util.Base64; import android.widget.Switch; import java.net.URLDecoder; import com.github.catvod.crawler.Spider; import com.github.catvod.crawler.SpiderDebug; import com.github.catvod.utils.okhttp.OkHttpUtil; import org.json.JSONArray; import org.json.JSONException; import org.json.JSONObject; import org.jsoup.Jsoup; import org.jsoup.nodes.Document; import org.jsoup.nodes.Element; import org.jsoup.select.Elements; import java.net.URLEncoder; import java.util.ArrayList; import java.util.Comparator; import java.util.HashMap; import java.util.Iterator; import java.util.List; import java.util.Map; import java.util.TreeMap; import java.util.regex.Matcher; import java.util.regex.Pattern; public class Tangrenjie extends Spider { private static final String siteUrl = "https://www.tangrenjie.tv"; private static final String siteHost = "tangrenjie.tv"; /** * 播放源配置 */ private JSONObject playerConfig; /** * 筛选配置 */ private JSONObject filterConfig; private Pattern regexCategory = Pattern.compile("/vod/type/id/(\\d+).html"); private Pattern regexVid = Pattern.compile("/vod/detail/id/(\\d+).html"); private Pattern regexPlay = Pattern.compile("/vod/play/id/(\\d+)/sid/(\\d+)/nid/(\\d+).html"); private Pattern regexPage = Pattern.compile("\\S+/page/(\\d+)\\S+"); @Override public void init(Context context) { super.init(context); try { playerConfig = new JSONObject("{\"if101\":{\"show\":\"蓝光一\",\"des\":\"\",\"ps\":\"0\",\"parse\":\"\",\"or\":999}}"); filterConfig = new JSONObject("{\"1\":[{\"key\":\"cateId\",\"name\":\"分类\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"动作片\",\"v\":\"6\"},{\"n\":\"喜剧片\",\"v\":\"7\"},{\"n\":\"爱情片\",\"v\":\"8\"},{\"n\":\"科幻片\",\"v\":\"9\"},{\"n\":\"恐怖片\",\"v\":\"10\"},{\"n\":\"剧情片\",\"v\":\"11\"},{\"n\":\"战争片\",\"v\":\"12\"}]},{\"key\":\"area\",\"name\":\"地区\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"大陆\",\"v\":\"大陆\"},{\"n\":\"香港\",\"v\":\"香港\"},{\"n\":\"台湾\",\"v\":\"台湾\"},{\"n\":\"美国\",\"v\":\"美国\"},{\"n\":\"法国\",\"v\":\"法国\"},{\"n\":\"英国\",\"v\":\"英国\"},{\"n\":\"日本\",\"v\":\"日本\"},{\"n\":\"韩国\",\"v\":\"韩国\"},{\"n\":\"德国\",\"v\":\"德国\"},{\"n\":\"泰国\",\"v\":\"泰国\"},{\"n\":\"印度\",\"v\":\"印度\"},{\"n\":\"其他\",\"v\":\"其他\"}]},{\"key\":\"year\",\"name\":\"年份\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"2022\",\"v\":\"2022\"},{\"n\":\"2021\",\"v\":\"2021\"},{\"n\":\"2020\",\"v\":\"2020\"},{\"n\":\"2019\",\"v\":\"2019\"},{\"n\":\"2018\",\"v\":\"2018\"},{\"n\":\"2017\",\"v\":\"2017\"},{\"n\":\"2016\",\"v\":\"2016\"},{\"n\":\"2008\",\"v\":\"2008\"},{\"n\":\"2000\",\"v\":\"2000\"},{\"n\":\"1997\",\"v\":\"1997\"},{\"n\":\"1980\",\"v\":\"1980\"}]},{\"key\":\"by\",\"name\":\"排序\",\"value\":[{\"n\":\"时间\",\"v\":\"time\"},{\"n\":\"人气\",\"v\":\"hits\"},{\"n\":\"评分\",\"v\":\"score\"}]}],\"2\":[{\"key\":\"cateId\",\"name\":\"分类\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"国产剧\",\"v\":\"13\"},{\"n\":\"港台剧\",\"v\":\"14\"},{\"n\":\"日韩剧\",\"v\":\"15\"},{\"n\":\"欧美剧\",\"v\":\"16\"},{\"n\":\"海外剧\",\"v\":\"27\"}]},{\"key\":\"area\",\"name\":\"地区\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"大陆\",\"v\":\"大陆\"},{\"n\":\"香港\",\"v\":\"香港\"},{\"n\":\"台湾\",\"v\":\"台湾\"},{\"n\":\"日本\",\"v\":\"日本\"},{\"n\":\"欧美\",\"v\":\"欧美\"},{\"n\":\"韩国\",\"v\":\"韩国\"}]},{\"key\":\"year\",\"name\":\"年份\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"2022\",\"v\":\"2022\"},{\"n\":\"2021\",\"v\":\"2021\"},{\"n\":\"2020\",\"v\":\"2020\"},{\"n\":\"2019\",\"v\":\"2019\"},{\"n\":\"2018\",\"v\":\"2018\"},{\"n\":\"2017\",\"v\":\"2017\"},{\"n\":\"2016\",\"v\":\"2016\"},{\"n\":\"2008\",\"v\":\"2008\"},{\"n\":\"2000\",\"v\":\"2000\"},{\"n\":\"1997\",\"v\":\"1997\"},{\"n\":\"1980\",\"v\":\"1980\"}]},{\"key\":\"by\",\"name\":\"排序\",\"value\":[{\"n\":\"时间\",\"v\":\"time\"},{\"n\":\"人气\",\"v\":\"hits\"},{\"n\":\"评分\",\"v\":\"score\"}]}],\"4\":[{\"key\":\"area\",\"name\":\"地区\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"大陆\",\"v\":\"大陆\"},{\"n\":\"香港\",\"v\":\"香港\"},{\"n\":\"台湾\",\"v\":\"台湾\"},{\"n\":\"日本\",\"v\":\"日本\"},{\"n\":\"欧美\",\"v\":\"欧美\"},{\"n\":\"韩国\",\"v\":\"韩国\"}]},{\"key\":\"year\",\"name\":\"年份\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"2022\",\"v\":\"2022\"},{\"n\":\"2021\",\"v\":\"2021\"},{\"n\":\"2020\",\"v\":\"2020\"},{\"n\":\"2019\",\"v\":\"2019\"},{\"n\":\"2018\",\"v\":\"2018\"},{\"n\":\"2017\",\"v\":\"2017\"},{\"n\":\"2016\",\"v\":\"2016\"},{\"n\":\"2008\",\"v\":\"2008\"},{\"n\":\"2000\",\"v\":\"2000\"},{\"n\":\"1997\",\"v\":\"1997\"},{\"n\":\"1980\",\"v\":\"1980\"}]},{\"key\":\"by\",\"name\":\"排序\",\"value\":[{\"n\":\"时间\",\"v\":\"time\"},{\"n\":\"人气\",\"v\":\"hits\"},{\"n\":\"评分\",\"v\":\"score\"}]}],\"3\":[{\"key\":\"cateId\",\"name\":\"分类\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"纪录片\",\"v\":\"28\"}]},{\"key\":\"area\",\"name\":\"地区\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"大陆\",\"v\":\"大陆\"},{\"n\":\"香港\",\"v\":\"香港\"},{\"n\":\"台湾\",\"v\":\"台湾\"},{\"n\":\"美国\",\"v\":\"美国\"},{\"n\":\"法国\",\"v\":\"法国\"},{\"n\":\"英国\",\"v\":\"英国\"},{\"n\":\"日本\",\"v\":\"日本\"},{\"n\":\"韩国\",\"v\":\"韩国\"},{\"n\":\"德国\",\"v\":\"德国\"},{\"n\":\"泰国\",\"v\":\"泰国\"},{\"n\":\"印度\",\"v\":\"印度\"},{\"n\":\"其他\",\"v\":\"其他\"}]},{\"key\":\"year\",\"name\":\"年份\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"2022\",\"v\":\"2022\"},{\"n\":\"2021\",\"v\":\"2021\"},{\"n\":\"2020\",\"v\":\"2020\"},{\"n\":\"2019\",\"v\":\"2019\"},{\"n\":\"2018\",\"v\":\"2018\"},{\"n\":\"2017\",\"v\":\"2017\"},{\"n\":\"2016\",\"v\":\"2016\"},{\"n\":\"2008\",\"v\":\"2008\"},{\"n\":\"2000\",\"v\":\"2000\"},{\"n\":\"1997\",\"v\":\"1997\"},{\"n\":\"1980\",\"v\":\"1980\"}]},{\"key\":\"by\",\"name\":\"排序\",\"value\":[{\"n\":\"时间\",\"v\":\"time\"},{\"n\":\"人气\",\"v\":\"hits\"},{\"n\":\"评分\",\"v\":\"score\"}]}]}"); } catch (JSONException e) { SpiderDebug.log(e); } } /** * 爬虫headers * * @param url * @return */ protected HashMap getHeaders(String url) { HashMap headers = new HashMap<>(); headers.put("User-Agent", "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.114 Safari/537.36"); headers.put("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8"); headers.put("Accept-Language", "zh-CN,zh;q=0.8,zh-TW;q=0.7,zh-HK;q=0.5,en-US;q=0.3,en;q=0.2"); return headers; } /** * 获取分类数据 + 首页最近更新视频列表数据 * * @param filter 是否开启筛选 关联的是 软件设置中 首页数据源里的筛选开关 * @return */ @Override public String homeContent(boolean filter) { try { Document doc = Jsoup.parse(OkHttpUtil.string(siteUrl, getHeaders(siteUrl))); // 分类节点 Elements elements = doc.select("ul.top_nav li a"); JSONArray classes = new JSONArray(); for (Element ele : elements) { Elements b = ele.select("b"); String name = b.text(); boolean show = name.equals("电影") || name.equals("电视剧") || name.equals("动漫") || name.equals("综艺") ; if (show) { Matcher mather = regexCategory.matcher(ele.attr("href")); if (!mather.find()) continue; // 把分类的id和名称取出来加到列表里 String id = mather.group(1).trim(); JSONObject jsonObject = new JSONObject(); jsonObject.put("type_id", id); jsonObject.put("type_name", name); classes.put(jsonObject); } } JSONObject result = new JSONObject(); if (filter) { result.put("filters", filterConfig); } result.put("class", classes); try { // 取首页推荐视频列表 Element homeList = doc.select("div.cbox1").get(1); Elements list = homeList.select("ul.vodlist li.vodlist_item"); JSONArray videos = new JSONArray(); for (int i = 0; i < list.size(); i++) { Element vod = list.get(i); String title = vod.selectFirst("a.vodlist_thumb").attr("title"); String cover = siteUrl + vod.selectFirst("a.vodlist_thumb").attr("data-original"); String remark = vod.select("span.pic_text b").text(); Matcher matcher = regexVid.matcher(vod.selectFirst("a.vodlist_thumb").attr("href")); if (!matcher.find()) continue; String id = "1="+ matcher.group(1); JSONObject v = new JSONObject(); v.put("vod_id", id); v.put("vod_name", title); v.put("vod_pic", cover); v.put("vod_remarks", remark); videos.put(v); } result.put("list", videos); } catch (Exception e) { SpiderDebug.log(e); } return result.toString(); } catch (Exception e) { SpiderDebug.log(e); } return ""; } /** * 获取分类信息数据 * * @param tid 分类id * @param pg 页数 * @param filter 同homeContent方法中的filter * @param extend 筛选参数{k:v, k1:v1} * @return */ @Override public String categoryContent(String tid, String pg, boolean filter, HashMap extend) { try { String url = siteUrl + "/vod/show/id/"; if (extend != null && extend.size() > 0 && extend.containsKey("tid") && extend.get("tid").length() > 0) { url += extend.get("tid"); } else { url += tid; } if (extend != null && extend.size() > 0) { for (Iterator it = extend.keySet().iterator(); it.hasNext(); ) { String key = it.next(); String value = extend.get(key); if (value.length() > 0) { url += "/" + key + "/" + URLEncoder.encode(value); } } } url += "/page/" + pg + ".html"; String html = OkHttpUtil.string(url, getHeaders(url)); Document doc = Jsoup.parse(html); JSONObject result = new JSONObject(); int pageCount = 0; int page = -1; // 取页码相关信息 Elements pageInfo = doc.select("ul.page"); if (pageInfo.size() == 0) { page = Integer.parseInt(pg); pageCount = page; } else { for (int i = 0; i < pageInfo.size(); i++) { Element li = pageInfo.get(i); Element a = li.selectFirst("li a"); if (a == null) continue; String span = pageInfo.select("li.active a").text(); String wy = doc.select("ul.page li a").last().attr("href"); if (page == -1) { page = Integer.parseInt(span); } else { page = 0; } Matcher matcher = regexPage.matcher(wy); if (matcher.find()) { pageCount = Integer.parseInt(matcher.group(1).trim()); } else { pageCount = 0; } break; } } JSONArray videos = new JSONArray(); if (!html.contains("没有找到您想要的结果哦")) { // 取当前分类页的视频列表 Elements list = doc.select("ul.vodlist li.vodlist_item"); for (int i = 0; i < list.size(); i++) { Element vod = list.get(i); String title = vod.selectFirst("a.vodlist_thumb").attr("title"); String cover = siteUrl + vod.selectFirst("a.vodlist_thumb").attr("data-original"); String remark = vod.select("span.pic_text b").text(); Matcher matcher = regexVid.matcher(vod.selectFirst("a.vodlist_thumb").attr("href")); if (!matcher.find()) continue; String id = tid + "=" + matcher.group(1); JSONObject v = new JSONObject(); v.put("vod_id", id); v.put("vod_name", title); v.put("vod_pic", cover); v.put("vod_remarks", remark); videos.put(v); } } result.put("page", page); result.put("pagecount", pageCount); result.put("limit", 20); result.put("total", pageCount <= 1 ? videos.length() : pageCount * 20); result.put("list", videos); return result.toString(); } catch (Exception e) { SpiderDebug.log(e); } return ""; } /** * 视频详情信息 * * @param ids 视频id * @return */ @Override public String detailContent(List ids) { try { // 视频详情url String[] idd = ids.get(0).split("="); String tid =idd[0]; String detail = idd[1]; String url = siteUrl + "/vod/detail/id/" + detail + ".html"; String htmltemp = OkHttpUtil.string(url, getHeaders(url)); String htmlsrc =""; if(tid.equals("1")){ htmlsrc = htmltemp.substring(0,25000); }else{ htmlsrc = htmltemp; } Document doc = Jsoup.parse(htmlsrc); JSONObject result = new JSONObject(); JSONObject vodList = new JSONObject(); // 取基本数据 String vid = doc.selectFirst("div.content_thumb a").attr("href"); String cover = siteUrl + doc.selectFirst("div.content_thumb a").attr("data-original"); String title = doc.selectFirst("div.content_thumb a").attr("title"); String desc = doc.selectFirst("div.content > section > font").text(); String category = "", area = "", year = "", director = "", actor = "", remark = ""; remark = doc.select("span.data_style b").text(); Elements span_text_muted = doc.select("li.data span"); for (int i = 0; i < span_text_muted.size() - 2; i++) { Element text = span_text_muted.get(i); String info = text.text(); if (info.contains("类型")) { try { String cate = text.parent().text(); category =cate.substring(3); } catch (Exception e) { category = ""; } } else if (info.contains("年份")) { try { year = text.parent().select("a").get(0).text(); } catch (Exception e) { year = ""; } } else if (info.contains("地区")) { try { area = text.parent().select("a").get(0).text(); } catch (Exception e) { area = ""; } } } List actors = new ArrayList<>(); Elements aa = doc.select("li.data").get(0).select("a"); for (int j = 0; j < aa.size(); j++) { actors.add(aa.get(j).text()); } actor = TextUtils.join(",", actors); List directors = new ArrayList<>(); Elements bb = doc.select("li.data").get(1).select("a"); for (int j = 0; j < bb.size(); j++) { directors.add(bb.get(j).text()); } director = TextUtils.join(",", directors); vodList.put("vod_id", vid); vodList.put("vod_name", title); vodList.put("vod_pic", cover); vodList.put("type_name", category); vodList.put("vod_year", year); vodList.put("vod_area", area); vodList.put("vod_remarks", remark); vodList.put("vod_actor", actor); vodList.put("vod_director", director); vodList.put("vod_content", desc); Map vod_play = new TreeMap<>(new Comparator() { @Override public int compare(String o1, String o2) { try { int sort1 = playerConfig.getJSONObject(o1).getInt("or"); int sort2 = playerConfig.getJSONObject(o2).getInt("or"); if (sort1 == sort2) { return 1; } return sort1 - sort2 > 0 ? 1 : -1; } catch (JSONException e) { SpiderDebug.log(e); } return 1; } }); // 取播放列表数据 Elements sources = doc.select("div.play_source_tab a"); Elements sourceList = doc.select("div.playlist_full"); for (int i = 0; i < sources.size(); i++) { Element source = sources.get(i); String sourceName = source.attr("alt"); sourceName = sourceName.replaceAll("[^\u4E00-\u9FA5]", ""); boolean found = false; for (Iterator it = playerConfig.keys(); it.hasNext(); ) { String flag = it.next(); if (playerConfig.getJSONObject(flag).getString("show").equals(sourceName)) { sourceName = flag; found = true; break; } } if (!found) continue; String playList = ""; Elements playListA = sourceList.get(i).select("ul>li>a"); List vodItems = new ArrayList<>(); for (int j = 0; j < playListA.size(); j++) { Element vod = playListA.get(j); Matcher matcher = regexPlay.matcher(vod.attr("href")); if (!matcher.find()) continue; String playURL = matcher.group(1) + "/sid/" + matcher.group(2) + "/nid/" + matcher.group(3); vodItems.add( vod.text() + "$" + playURL + "="+ tid); } if (vodItems.size() > 0) playList = TextUtils.join("#", vodItems); if (playList.length() == 0) continue; vod_play.put(sourceName, playList); } if (vod_play.size() > 0) { String vod_play_from = TextUtils.join("$$$", vod_play.keySet()); String vod_play_url = TextUtils.join("$$$", vod_play.values()); vodList.put("vod_play_from", vod_play_from); vodList.put("vod_play_url", vod_play_url); } JSONArray list = new JSONArray(); list.put(vodList); result.put("list", list); return result.toString(); } catch (Exception e) { SpiderDebug.log(e); } return ""; } /** * 获取视频播放信息 * * @param flag 播放源 * @param id 视频id * @param vipFlags 所有可能需要vip解析的源 * @return */ @Override public String playerContent(String flag, String id, List vipFlags) { try { // 播放页 url String[] idd = id.split("="); String tid =idd[1]; String detail = idd[0]; String url = siteUrl + "/vod/play/id/" + detail + ".html"; String htmltemp = OkHttpUtil.string(url, getHeaders(url)); String htmlsrc =""; if(tid.equals("1")){ htmlsrc = htmltemp.substring(0,25000); }else{ htmlsrc = htmltemp; } Document doc = Jsoup.parse(htmlsrc); Elements allScript = doc.select("script"); JSONObject result = new JSONObject(); for (int i = 0; i < allScript.size(); i++) { String scContent = allScript.get(i).toString(); if (scContent.contains("var player_")) { // 取直链 int start = scContent.indexOf('{'); int end = scContent.lastIndexOf('}') + 1; String json = scContent.substring(start, end); JSONObject player = new JSONObject(json); String urlsrc = player.getString("url"); String urlb64 = new String(Base64.decode(urlsrc.getBytes(), Base64.DEFAULT)); String urldec = URLDecoder.decode(urlb64,"UTF-8"); result.put("parse", 0); result.put("playUrl", ""); result.put("url", urldec); result.put("header", ""); break; } } //无直链则嗅探 String uu = result.getString("url"); if (!(uu.length() >0)){ result.put("parse", 1); result.put("url", url); } return result.toString(); } catch (Exception e) { SpiderDebug.log(e); } return ""; } @Override public String searchContent(String key, boolean quick) { try { String url = siteUrl + "/vod/search.html?wd=" + URLEncoder.encode(key) + "&submit="; String html = OkHttpUtil.string(url, getHeaders(url)); Document doc = Jsoup.parse(html); JSONObject result = new JSONObject(); JSONArray videos = new JSONArray(); if (!html.contains("没有找到您想要的结果哦")) { Elements list = doc.select("li.searchlist_item"); for (int i = 0; i < list.size(); i++) { Element vod = list.get(i); String classes = vod.selectFirst("h4.vodlist_title > a >span").text(); String tid=""; switch (classes) { case "电影": tid = "1"; break; case "电视剧": tid = "2"; break; case "综艺": tid = "3"; break; case "动漫": tid = "4"; break; } String title = vod.selectFirst("a.vodlist_thumb").attr("title"); String cover = siteUrl + vod.selectFirst("a.vodlist_thumb").attr("data-original"); String remark = vod.select("span.pic_text").text(); Matcher matcher = regexVid.matcher(vod.selectFirst("a.vodlist_thumb").attr("href")); if (!matcher.find()) continue; String id = tid + "=" + matcher.group(1); JSONObject v = new JSONObject(); v.put("vod_id", id); v.put("vod_name", title); v.put("vod_pic", cover); v.put("vod_remarks", remark); videos.put(v); } result.put("list", videos); return result.toString(); } } catch (Exception e) { SpiderDebug.log(e); } return ""; } private static String doReplaceRegex(Pattern pattern, String src) { if (pattern == null) return src; try { Matcher matcher = pattern.matcher(src); if (matcher.find()) { return matcher.group(1).trim(); } } catch (Exception e) { SpiderDebug.log(e); } return src; } @Override public boolean manualVideoCheck() { return true; } private String[] videoFormatList = new String[]{".m3u8", ".mp4", ".mpeg", ".flv"}; @Override public boolean isVideoFormat(String url) { url = url.toLowerCase(); if (url.contains("=http") || url.contains("=https") || url.contains("=https%3a%2f") || url.contains("=http%3a%2f")) { return false; } for (String format : videoFormatList) { if (url.contains(format)) { return true; } } return false; } }