yeyinly/TVBoxSE
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

first commit

Browse Source
This commit is contained in:
T00700
2022-12-02 13:36:22 +08:00
commit 236efd9a7f
1515 changed files with 326020 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
TVBox_Java/spider/Lib.java
+516
View File
@@ -0,0 +1,516 @@
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.regex.Matcher;
import java.util.regex.Pattern;
public class Lib extends Spider {
private static final String siteUrl = "https://www.libvio.fun";
private static final String siteHost = "www.libvio.fun";
/**
* 播放源配置
*/
private JSONObject playerConfig;
/**
* 筛选配置
*/
private JSONObject filterConfig;
private Pattern regexCategory = Pattern.compile("/type/(\\d+).html");
private Pattern regexVid = Pattern.compile("/detail/(\\d+).html");
private Pattern regexPlay = Pattern.compile("/play/(\\d+)-(\\d+)-(\\d+).html");
private Pattern regexPage = Pattern.compile("/show/(\\S+).html");
protected String ext = null;
@Override
public void init(Context context) {
super.init(context);
try {
String DataExt = OkHttpUtil.string("http://box.qinglin.tk/live/Lib.json", null);
JSONObject Config = new JSONObject(DataExt);
playerConfig = Config.getJSONObject("player");
filterConfig = Config.getJSONObject("filter");
} catch (JSONException e) {
SpiderDebug.log(e);
}
}
protected static HashMap<String, String> Headers() {
HashMap<String, String> 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<String, String> getHeaders(String url) {
HashMap<String, String> headers = new HashMap<>();
headers.put("method", "GET");
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, getHeaders(siteUrl)));
// 分类节点
Elements elements = doc.select("ul.stui-header__menu li a");
JSONArray classes = new JSONArray();
for (Element ele : elements) {
String name = ele.text();
boolean show = name.equals("电影") ||
name.equals("剧集") ||
name.equals("动漫") ||
name.equals("日韩剧") ||
name.equals("欧美剧");
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("ul.stui-vodlist").get(0);
Elements list = homeList.select("div.stui-vodlist__box");
JSONArray videos = new JSONArray();
for (int i = 0; i < list.size(); i++) {
Element vod = list.get(i);
String title = vod.select("a").attr("title");
String cover = vod.select("a").attr("data-original");
String remark = vod.select("a .pic-text").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("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<String, String> extend) {
try {
String[] urlParams = new String[]{"", "", "", "", "", "", "", "", "", "", "", ""};
urlParams[0] = tid;
urlParams[8] = pg;
if (extend != null && extend.size() > 0) {
for (Iterator<String> 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 + "/show/" + TextUtils.join("-", urlParams) + ".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.stui-page__item li");
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 name = a.text();
if (page == -1 && li.hasClass("active")) {
Matcher matcher = regexPage.matcher(a.attr("href"));
if (matcher.find()) {
//System.out.println("哈哈"+matcher.group(1));
page = Integer.parseInt(matcher.group(1).split("-")[8]);
} else {
page = 0;
}
}
if (name.equals("尾页")) {
Matcher matcher = regexPage.matcher(a.attr("href"));
if (matcher.find()) {
//System.out.println("尾页" + matcher.group(1));
pageCount = Integer.parseInt(matcher.group(1).split("-")[8]);
} 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.stui-vodlist__box");
for (int i = 0; i < list.size(); i++) {
Element vod = list.get(i);
String title = vod.select("a").attr("title");
String cover = vod.select("a").attr("data-original");
String remark = vod.select("a .pic-text").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 "";
}
private static String doReplaceRegex(Pattern pattern, String content) {
if (pattern == null) {
return content;
}
try {
Matcher matcher = pattern.matcher(content);
if (matcher.find()) {
return matcher.group(1).trim();
}
} catch (Exception e) {
SpiderDebug.log(e);
}
return content;
}
/**
* 视频详情信息
*
* @param ids 视频id
* @return
*/
@Override
public String detailContent(List<String> ids) {
try {
// 视频详情url
String url = siteUrl + "/detail/" + ids.get(0) + ".html";
Document doc = Jsoup.parse(OkHttpUtil.string(url, getHeaders(url)));
JSONObject result = new JSONObject();
JSONObject vodList = new JSONObject();
// 取基本数据
String cover = doc.select("div.stui-content__thumb a img").attr("data-original");
String title = doc.select("div.stui-content__detail h1").text();
String category = "", area = "", year = "", remark = "", director = "", actor = "", desc = "";
Elements data = doc.select("p.data");
Pattern cate = Pattern.compile("类型：(\\S+)");
category = doReplaceRegex(cate, data.get(0).text());
Pattern areas = Pattern.compile("地区：(\\S+)");
area = doReplaceRegex(areas, data.get(0).text());
Pattern years = Pattern.compile("年份：(\\S+)");
year = doReplaceRegex(years, data.get(0).text());
Pattern rms = Pattern.compile("更新：(\\S+)");
remark = doReplaceRegex(rms, data.get(3).text());
Pattern acs = Pattern.compile("主演：(\\S+)");
actor = doReplaceRegex(acs, data.get(1).text());
Pattern dis = Pattern.compile("导演：(\\S+)");
director = doReplaceRegex(dis, data.get(1).text());
desc = doc.selectFirst("span.detail-content").text().trim();
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_remarks", remark);
vodList.put("vod_actor", actor);
vodList.put("vod_director", director);
vodList.put("vod_content", desc);
Map<String, String> vod_play = new TreeMap<>(new Comparator<String>() {
@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.stui-vodlist__head h3");
Elements sourceList = doc.select("ul.stui-content__playlist");
for (int i = 0; i < sources.size(); i++) {
Element source = sources.get(i);
String sourceName = source.text();
boolean found = false;
for (Iterator<String> it = playerConfig.keys(); it.hasNext(); ) {
String flag = it.next();
if (playerConfig.getJSONObject(flag).getString("sh").equals(sourceName)) {
sourceName = flag;
found = true;
break;
}
}
if (!found)
continue;
String playList = "";
Elements playListA = sourceList.get(i).select("li a");
List<String> 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<String> vipFlags) {
try {
//定义播放用的headers
JSONObject headers = new JSONObject();
headers.put("origin", " https://www.libvio.fun");
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");
headers.put("referer", " https://www.libvio.fun");
// 播放页 url
String url = siteUrl + "/play/" + id + ".html";
Document doc = Jsoup.parse(OkHttpUtil.string(url, getHeaders(url)));
Elements allScript = doc.select("script");
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"));
if (player.getString("from").contains("LINE")) {
String videoUrl = pCfg.getString("pu") + player.getString("url");
Document docs = Jsoup.parse(OkHttpUtil.string(videoUrl, Headers()));
Pattern pattern = Pattern.compile("(?<=urls\\s=\\s').*?(?=')");
Elements allScripts = docs.select("body script");
for (int j = 0; j < allScripts.size(); j++) {
String scContents = allScripts.get(j).html().trim();
Matcher matcher = pattern.matcher(scContents);
if (matcher.find()) {
JSONObject result = new JSONObject();
String players = matcher.group(0);
result.put("parse", 0);
result.put("playUrl", "");
result.put("url", players);
return result.toString();
}
}
}
else {
SpiderDebug.log(url);
JSONObject result = new JSONObject();
result.put("parse", 1);
result.put("playUrl", "");
result.put("url", url);
// result.put("header", "{\"Referer\", \"https://www.libvio.me/\"}");
return result.toString();
}
break;
}
}
}
} catch (Exception e) {
SpiderDebug.log(e);
}
return "";
}
// protected static HashMap<String, String> sHeaders(String url) {
// HashMap<String, String> headers = new HashMap<>();
// headers.put("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.0.0 Safari/537.36");
// headers.put("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.9");
// headers.put("Accept-encoding", "gzip, deflate, br");
// headers.put("Accept-language", "zh-SG,zh;q=0.9,en-GB;q=0.8,en;q=0.7,zh-CN;q=0.6");
// return headers;
// }
@Override
public String searchContent(String key, boolean quick) {
try {
if (quick)
return "";
long currentTime = System.currentTimeMillis();
String url = siteUrl + "/index.php/ajax/suggest?mid=1&wd=" + URLEncoder.encode(key) + "&limit=10&timestamp=" + 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");
if (!TextUtils.isEmpty(cover) && !cover.startsWith("http")) {
cover = siteUrl + cover;
}
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 "";
}
}
// 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;
// }
//}