package com.github.catvod.spider; import android.content.Context; import android.text.TextUtils; 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.LinkedHashMap; import java.util.regex.Matcher; import java.util.regex.Pattern; import okhttp3.Headers; public class Acg extends Spider { private static final String siteUrl = "http://www.acgnya.com"; private static final String siteHost = "www.acgnya.com"; private String cookie=""; private String referer=""; /** * 播放源配置 */ private JSONObject playerConfig; /** * 筛选配置 */ private JSONObject filterConfig; private Pattern regexCategory = Pattern.compile("/vodtype/(\\d+)/"); private Pattern regexVid = Pattern.compile("/voddetail/(\\d+)/"); private Pattern regexPlay = Pattern.compile("/vodplay/(\\d+)-(\\d+)-(\\d+)/"); private Pattern regexPage = Pattern.compile("/vodshow/(\\d+)/"); @Override public void init(Context context) { super.init(context); try { playerConfig = new JSONObject("{\"buka\":{\"sh\":\"不卡推荐\",\"pu\":\"https://pro.ascepan.top/player/?url=\",\"sn\":1,\"or\":999}}"); filterConfig = new JSONObject("{\"20\":[{\"key\":3,\"name\":\"剧情\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"科幻\",\"v\":\"科幻\"},{\"n\":\"奇幻\",\"v\":\"奇幻\"},{\"n\":\"历史\",\"v\":\"历史\"},{\"n\":\"异世界\",\"v\":\"异世界\"},{\"n\":\"穿越\",\"v\":\"穿越\"},{\"n\":\"古装\",\"v\":\"古装\"},{\"n\":\"搞笑\",\"v\":\"搞笑\"},{\"n\":\"动作\",\"v\":\"动作\"}]},{\"key\":11,\"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\":\"2015\",\"v\":\"2015\"},{\"n\":\"2014\",\"v\":\"2014\"},{\"n\":\"2013\",\"v\":\"2013\"},{\"n\":\"2012\",\"v\":\"2012\"},{\"n\":\"2011\",\"v\":\"2011\"},{\"n\":\"2010\",\"v\":\"2010\"}]},{\"key\":4,\"name\":\"语言\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"普通话\",\"v\":\"普通话\"}]},{\"key\":5,\"name\":\"字母\",\"value\":[{\"n\":\"全部\",\"v\":\"\"},{\"n\":\"A\",\"v\":\"A\"},{\"n\":\"B\",\"v\":\"B\"},{\"n\":\"C\",\"v\":\"C\"},{\"n\":\"D\",\"v\":\"D\"},{\"n\":\"E\",\"v\":\"E\"},{\"n\":\"F\",\"v\":\"F\"},{\"n\":\"G\",\"v\":\"G\"},{\"n\":\"H\",\"v\":\"H\"},{\"n\":\"I\",\"v\":\"I\"},{\"n\":\"J\",\"v\":\"J\"},{\"n\":\"K\",\"v\":\"K\"},{\"n\":\"L\",\"v\":\"L\"},{\"n\":\"M\",\"v\":\"M\"},{\"n\":\"N\",\"v\":\"N\"},{\"n\":\"O\",\"v\":\"O\"},{\"n\":\"P\",\"v\":\"P\"},{\"n\":\"Q\",\"v\":\"Q\"},{\"n\":\"R\",\"v\":\"R\"},{\"n\":\"S\",\"v\":\"S\"},{\"n\":\"T\",\"v\":\"T\"},{\"n\":\"U\",\"v\":\"U\"},{\"n\":\"V\",\"v\":\"V\"},{\"n\":\"W\",\"v\":\"W\"},{\"n\":\"X\",\"v\":\"X\"},{\"n\":\"Y\",\"v\":\"Y\"},{\"n\":\"Z\",\"v\":\"Z\"},{\"n\":\"0-9\",\"v\":\"0-9\"}]},{\"key\":2,\"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); } } protected static HashMap Headers() { HashMap headers = new HashMap<>(); headers.put("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/102.0.5005.62 Safari/537.36"); headers.put("Referer", siteUrl); return headers; } /** * 爬虫headers * * @param url * @return */ protected HashMap getHeaders(String url) { HashMap headers = new HashMap<>(); headers.put("method", "GET"); headers.put("Host", siteHost); headers.put("Upgrade-Insecure-Requests", "1"); headers.put("DNT", "1"); 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, getHeaders2(siteUrl,referer))); // 分类节点 Elements elements = doc.select("ul.navbar-items>li.navbar-item>a"); System.out.println("shouye" + elements); JSONArray classes = new JSONArray(); for (Element ele : elements) { String name = ele.text(); boolean show = name.equals("连载动画") || name.equals("国漫") || name.equals("完结动画"); 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 Elements list = doc.select("div[class='module-items module-poster-items-base'] >a"); JSONArray videos = new JSONArray(); for (int i = 0; i < list.size(); i++) { Element vod = list.get(i); Matcher matcher = regexVid.matcher(vod.selectFirst(".module-poster-item").attr("href")); if (!matcher.find()) continue; String title = vod.selectFirst(".module-poster-item").attr("title"); String cover = vod.selectFirst("div.module-item-cover div.module-item-pic img").attr("data-original"); String remark = vod.selectFirst("div.module-item-cover div.module-item-note").text(); if (!TextUtils.isEmpty(cover) && !cover.startsWith("http")) { cover = siteUrl + cover; } String id = 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[] urlParams = new String[]{"", "", "", "", "", "", "", "", "", "", "", ""}; urlParams[0] = tid; urlParams[8] = pg; if (extend != null && extend.size() > 0) { for (Iterator it = extend.keySet().iterator(); it.hasNext(); ) { String key = it.next(); String value = extend.get(key); urlParams[Integer.parseInt(key)] = URLEncoder.encode(value); } } // 获取分类数据的url String url = siteUrl + "/vodshow/" + TextUtils.join("-", urlParams) + "/"; 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("div[id='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("a"); if (a == null) continue; String span = pageInfo.select("span.page-current").text(); String wy = doc.select("div[id='page'] 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("div.module-items.module-poster-items-base a "); for (int i = 0; i < list.size(); i++) { Element vod = list.get(i); String title = vod.select("a").attr("title"); String cover = vod.select("img.lazy.lazyload").attr("data-original"); String remark = vod.select("div.module-item-note").text(); Matcher matcher = regexVid.matcher(vod.select("a").attr("href")); if (!matcher.find()) continue; String id = 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", 48); result.put("total", pageCount <= 1 ? videos.length() : pageCount * 48); 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 url = siteUrl + "/voddetail/" + ids.get(0) + "/"; Document doc = Jsoup.parse(OkHttpUtil.string(url,getHeaders2(url,referer))); JSONObject result = new JSONObject(); JSONObject vodList = new JSONObject(); // 取基本数据 String cover = doc.selectFirst("div.module-item-cover div.module-item-pic > img").attr("data-original"); String title = doc.selectFirst("div.module-info-main div.module-info-heading > h1").text(); String desc = doc.selectFirst("div.module-info-item div.module-info-introduction-content > p").text().trim(); String category = "", area = "", year = "", director = "", actor = ""; category = doc.select("div.module-info-main div.module-info-heading div.module-info-tag-link").get(2).text(); area = doc.select("div.module-info-main div.module-info-heading div.module-info-tag-link").get(1).text(); year = doc.select("div.module-info-main div.module-info-heading div.module-info-tag-link").get(0).text(); Elements span_text_muted = doc.select("div.module-info-content div.module-info-items div.module-info-item"); for (int i = 0; i < span_text_muted.size(); i++) { Element text = span_text_muted.get(i); String info = text.select("span").text(); if (info.contains("导演")) { try { director = text.select("div > a").text(); } catch (Exception e) { director = ""; } } else if (info.contains("主演")) { try { actor = text.select("div > a").text(); } catch (Exception e) { actor = ""; } } } vodList.put("vod_id", ids.get(0)); 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_actor", actor); vodList.put("vod_director", director); vodList.put("vod_content", desc); Map vod_play = new LinkedHashMap<>(); /* 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.module-tab-items-box>div>span"); System.out.print("sor++" + sources); Elements sourceList = doc.select("div.module-list>div.module-play-list"); System.out.print("sor1++" + sourceList); for (int i = 0; i < sources.size(); i++) { Element source = sources.get(i); String sourceName = source.text(); // boolean found = false; boolean found = true; // for (Iterator it = playerConfig.keys(); it.hasNext(); ) { // String flag = it.next(); // if (playerConfig.getJSONObject(flag).getString("sh").equals(sourceName)) { // sourceName = playerConfig.getJSONObject(flag).getString("sh"); // found = true; // break; // } // } if (!found) continue; String playList = ""; Elements playListA = sourceList.get(i).select("div.module-play-list-content>a"); System.out.print("pl++" + playList); 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) + "-" + matcher.group(2) + "-" + matcher.group(3); vodItems.add(vod.text() + "$" + playURL); } 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 { //定义播放用的headers JSONObject headers = new JSONObject(); //headers.put("Host", " cokemv.co"); headers.put("origin", " http://www.acgnya.com"); 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", " */*"); headers.put("Accept-Language", " zh-CN,zh;q=0.9,en-US;q=0.3,en;q=0.7"); headers.put("Accept-Encoding", " gzip, deflate"); // 播放页 url String url = siteUrl + "/vodplay/" + id + "/"; Document doc = Jsoup.parse(OkHttpUtil.string(url, getHeaders(url))); Elements allScript = doc.select("script"); JSONObject result = new JSONObject(); for (int i = 0; i < allScript.size(); i++) { String scContent = allScript.get(i).html().trim(); if (scContent.startsWith("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); if (playerConfig.has(player.getString("from"))) { JSONObject pCfg = playerConfig.getJSONObject(player.getString("from")); String videoUrl = player.getString("url"); String playUrl = pCfg.getString("pu"); result.put("parse", pCfg.getInt("sn")); result.put("playUrl", playUrl); result.put("url", videoUrl); result.put("header", headers.toString()); } break; } } return result.toString(); } catch (Exception e) { SpiderDebug.log(e); } return ""; } protected HashMap getHeaders2(String url,String ref) { 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"); if(!ref.equals("google")){ headers.put("Authority", "www.acgnya.com"); if(ref.length()>0){ if(ref.equals("origin")){ headers.put("Origin", "http://www.acgnya.com"); } else { headers.put("Referer", ref); } } if(cookie.length()>0){ headers.put("Cookie", cookie); } } 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; } protected void getCookie(){ cookie=""; String cookieurl="http://www.acgnya.com/zzzzz"; Map> cookies = new HashMap<>(); OkHttpUtil.string(cookieurl,getHeaders2(cookieurl,""),cookies); for( Map.Entry> entry : cookies.entrySet() ){ if(entry.getKey().equals("set-cookie")){ cookie = TextUtils.join(";",entry.getValue()); break; } } } @Override public String searchContent(String key, boolean quick) { try { long currentTime = System.currentTimeMillis(); String url = siteUrl + "/index.php/ajax/suggest?mid=1&wd=" + URLEncoder.encode(key) + "&limit=35×tamp=" + currentTime; JSONObject searchResult = new JSONObject(OkHttpUtil.string(url, getHeaders(url))); JSONObject result = new JSONObject(); JSONArray videos = new JSONArray(); if (searchResult.getInt("total") > 0) { JSONArray lists = new JSONArray(searchResult.getString("list")); for (int i = 0; i < lists.length(); i++) { JSONObject vod = lists.getJSONObject(i); String id = vod.getString("id"); String title = vod.getString("name"); String cover = vod.getString("pic"); JSONObject v = new JSONObject(); v.put("vod_id", id); v.put("vod_name", title); v.put("vod_pic", cover); v.put("vod_remarks", ""); videos.put(v); } } result.put("list", videos); return result.toString(); } catch (Exception e) { SpiderDebug.log(e); } return ""; } }