Merge branch 'pr/40'

pull/69/head
FongMi 2 years ago
commit 4dad5a865e
  1. 145
      app/src/main/java/com/github/catvod/spider/Douban.java
  2. 253
      app/src/main/java/com/github/catvod/spider/Xb6v.java

@ -0,0 +1,145 @@
package com.github.catvod.spider;
import android.content.Context;
import com.github.catvod.bean.Class;
import com.github.catvod.bean.Result;
import com.github.catvod.bean.Vod;
import com.github.catvod.crawler.Spider;
import com.github.catvod.crawler.SpiderDebug;
import com.github.catvod.net.OkHttp;
import com.github.catvod.utils.Utils;
import com.google.gson.JsonParser;
import org.json.JSONArray;
import org.json.JSONObject;
import java.net.URLEncoder;
import java.util.*;
public class Douban extends Spider {
private final String hostURL = "https://frodo.douban.com/api/v2";
private final String apikey = "?apikey=0ac44ae016490db2204ce0a042db2916";
private String extend;
private Map<String, String> getHeader() {
Map<String, String> header = new HashMap<>();
header.put("Host", "frodo.douban.com");
header.put("Connection", "Keep-Alive");
header.put("Referer", "https://servicewechat.com/wx2f9b06c1de1ccfca/84/page-frame.html");
header.put("User-Agent", "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/53.0.2785.143 Safari/537.36 MicroMessenger/7.0.9.501 NetType/WIFI MiniProgramEnv/Windows WindowsWechat");
return header;
}
@Override
public void init(Context context, String extend) throws Exception {
super.init(context, extend);
this.extend = extend;
}
@Override
public String homeContent(boolean filter) throws Exception {
List<Class> classes = new ArrayList<>();
List<String> typeIds = Arrays.asList("hot_gaia", "tv_hot", "show_hot", "movie", "tv", "rank_list_movie", "rank_list_tv");
List<String> typeNames = Arrays.asList("热门电影", "热播剧集", "热播综艺", "电影筛选", "电视筛选", "电影榜单", "电视剧榜单");
for (int i = 0; i < typeIds.size(); i++) classes.add(new Class(typeIds.get(i), typeNames.get(i)));
String recommendURL = "http://api.douban.com/api/v2/subject_collection/subject_real_time_hotest/items" + apikey;
JSONObject jsonObject = new JSONObject(OkHttp.string(recommendURL, getHeader()));
JSONArray items = jsonObject.optJSONArray("subject_collection_items");
return Result.string(classes, parseVodListFromJSONArray(items), filter ? JsonParser.parseString(OkHttp.string(extend)) : null);
}
@Override
public String categoryContent(String tid, String pg, boolean filter, HashMap<String, String> extend) throws Exception {
HashMap<String, String> ext = new HashMap<>();
if (extend != null && extend.size() > 0) {
ext.putAll(extend);
}
String sort = ext.get("sort") == null ? "T" : ext.get("sort");
String tags = URLEncoder.encode(getTags(ext));
int start = (Integer.parseInt(pg) - 1) * 20;
String cateURL;
String itemKey = "items";
switch (tid) {
case "hot_gaia":
sort = ext.get("sort") == null ? "recommend" : ext.get("sort");
String area = ext.get("area") == null ? "全部" : ext.get("area");
sort = sort + "&area=" + URLEncoder.encode(area);
cateURL = hostURL + "/movie/hot_gaia" + apikey + "&sort=" + sort + "&start=" + start + "&count=20";
break;
case "tv_hot":
String type = ext.get("type") == null ? "tv_hot" : ext.get("type");
cateURL = hostURL + "/subject_collection/" + type + "/items" + apikey + "&start=" + start + "&count=20";
itemKey = "subject_collection_items";
break;
case "show_hot":
String showType = ext.get("type") == null ? "show_hot" : ext.get("type");
cateURL = hostURL + "/subject_collection/" + showType + "/items" + apikey + "&start=" + start + "&count=20";
itemKey = "subject_collection_items";
break;
case "tv":
cateURL = hostURL + "/tv/recommend" + apikey + "&sort=" + sort + "&tags=" + tags + "&start=" + start + "&count=20";
break;
case "rank_list_movie":
String rankMovieType = ext.get("榜单") == null ? "movie_real_time_hotest" : ext.get("榜单");
cateURL = hostURL + "/subject_collection/" + rankMovieType + "/items" + apikey + "&start=" + start + "&count=20";
itemKey = "subject_collection_items";
break;
case "rank_list_tv":
String rankTVType = ext.get("榜单") == null ? "tv_real_time_hotest" : ext.get("榜单");
cateURL = hostURL + "/subject_collection/" + rankTVType + "/items" + apikey + "&start=" + start + "&count=20";
itemKey = "subject_collection_items";
break;
default:
cateURL = hostURL + "/movie/recommend" + apikey + "&sort=" + sort + "&tags=" + tags + "&start=" + start + "&count=20";
}
JSONObject jsonObject = new JSONObject(OkHttp.string(cateURL, getHeader()));
JSONArray items = jsonObject.getJSONArray(itemKey);
List<Vod> list = parseVodListFromJSONArray(items);
int page = Integer.parseInt(pg), count = Integer.MAX_VALUE, limit = 20, total = Integer.MAX_VALUE;
return Result.get().vod(list).page(page, count, limit, total).string();
}
private List<Vod> parseVodListFromJSONArray(JSONArray items) throws Exception {
List<Vod> list = new ArrayList<>();
for (int i = 0; i < items.length(); i++) {
JSONObject item = items.getJSONObject(i);
String vodId = "msearch:" + item.optString("id");
String name = item.optString("title");
String pic = getPic(item);
String remark = getRating(item);
list.add(new Vod(vodId, name, pic, remark));
}
return list;
}
private String getRating(JSONObject item) {
try {
return "评分:" + item.getJSONObject("rating").optString("value");
} catch (Exception e) {
SpiderDebug.log(e);
}
return "";
}
private String getPic(JSONObject item) {
try {
return item.getJSONObject("pic").optString("normal") + "@Referer=https://api.douban.com/@User-Agent=" + Utils.CHROME;
} catch (Exception e) {
SpiderDebug.log(e);
}
return "";
}
private String getTags(HashMap<String, String> ext) {
try {
StringBuilder tags = new StringBuilder();
for (String key : ext.keySet()) {
if (key.equals("sort")) continue;
tags.append(ext.get(key)).append(",");
}
return tags.substring(0, tags.lastIndexOf(","));
} catch (Exception e) {
// SpiderDebug.log(e);
}
return "";
}
}

@ -0,0 +1,253 @@
package com.github.catvod.spider;
import android.text.TextUtils;
import com.github.catvod.bean.Class;
import com.github.catvod.bean.Filter;
import com.github.catvod.bean.Result;
import com.github.catvod.bean.Vod;
import com.github.catvod.crawler.Spider;
import com.github.catvod.crawler.SpiderDebug;
import com.github.catvod.net.OkHttp;
import com.github.catvod.utils.Utils;
import okhttp3.FormBody;
import okhttp3.Request;
import okhttp3.RequestBody;
import okhttp3.Response;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;
import java.util.*;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
/**
* @author zhixc
* 新版6V电影网
*/
public class Xb6v extends Spider {
private final String siteURL = "http://www.xb6v.com";
private String nextSearchURLPrefix;
private String nextSearchURLSuffix;
private Map<String, String> getHeader() {
Map<String, String> header = new HashMap<>();
header.put("User-Agent", Utils.CHROME);
header.put("Referer", siteURL + "/");
return header;
}
private Map<String, String> getDetailHeader() {
Map<String, String> header = new HashMap<>();
header.put("User-Agent", Utils.CHROME);
return header;
}
@Override
public String homeContent(boolean filter) throws Exception {
List<Class> classes = new ArrayList<>();
String html = OkHttp.string(siteURL, getHeader());
Document doc = Jsoup.parse(html);
Elements elements = doc.select("#menus > li > a");
LinkedHashMap<String, List<Filter>> filters = new LinkedHashMap<>();
for (int i = 0; i < elements.size(); i++) {
if (i < 2 || i == elements.size() - 1) continue;
Element e = elements.get(i);
String typeId = e.attr("href");
String typeName = e.text();
if (typeName.equals("电视剧")) {
List<Filter.Value> values = new ArrayList<>();
values.add(new Filter.Value("不限", ""));
for (Element a : e.nextElementSibling().select("a")) {
values.add(new Filter.Value(a.text(), a.attr("href").replaceAll(typeId, "")));
}
List<Filter> filterList = new ArrayList<>();
filterList.add(new Filter("cateId", "类型", values));
filters.put(typeId, filterList);
}
classes.add(new Class(typeId, typeName));
}
return Result.string(classes, parseVodListFromDoc(doc), filters);
}
private List<Vod> parseVodListFromDoc(Document doc) {
Elements items = doc.select("#post_container .post_hover");
List<Vod> list = new ArrayList<>();
for (Element item : items) {
Element element = item.select("[class=zoom]").get(0);
String vodId = element.attr("href");
String name = element.attr("title").replaceAll("</?[^>]+>", "");
String pic = element.select("img").attr("src");
String remark = item.select("[rel=category tag]").text();
list.add(new Vod(vodId, name, pic, remark));
}
return list;
}
@Override
public String categoryContent(String tid, String pg, boolean filter, HashMap<String, String> extend) throws Exception {
HashMap<String, String> ext = new HashMap<>();
if (extend != null && extend.size() > 0) {
ext.putAll(extend);
}
String cateId = ext.get("cateId") == null ? "" : ext.get("cateId");
String cateURL = siteURL + tid + cateId;
if (!pg.equals("1")) cateURL += "index_" + pg + ".html";
String html = OkHttp.string(cateURL, getHeader());
Document doc = Jsoup.parse(html);
String href = doc.select(".pagination > a").last().attr("href");
int page = Integer.parseInt(pg);
int count = Integer.parseInt(getStrByRegex(Pattern.compile("index_(.*?).html"), href));
int limit = 18;
Elements items = doc.select("#post_container .post_hover");
int total = page == count ? (page - 1) * limit + items.size() : count * limit;
return Result.get().vod(parseVodListFromDoc(doc)).page(page, count, limit, total).string();
}
@Override
public String detailContent(List<String> ids) throws Exception {
String vodId = ids.get(0);
String detailURL = siteURL + vodId;
String html = OkHttp.string(detailURL, getDetailHeader());
Document doc = Jsoup.parse(html);
Elements sourceList = doc.select("#post_content");
String circuitName = "磁力线路";
Map<String, String> playMap = new LinkedHashMap<>();
int i = 0;
for (Element source : sourceList) {
Elements aList = source.select("table a");
List<String> vodItems = new ArrayList<>();
for (Element a : aList) {
String episodeURL = a.attr("href");
String episodeName = a.text();
if (!episodeURL.toLowerCase().startsWith("magnet")) continue;
vodItems.add(episodeName + "$" + episodeURL);
}
if (vodItems.size() > 0) {
i++;
playMap.put(circuitName + i, TextUtils.join("#", vodItems));
}
}
String partHTML = doc.select(".context").html();
String name = doc.select(".article_container > h1").text();
String pic = doc.select("#post_content img").attr("src");
String typeName = getStrByRegex(Pattern.compile("◎类   (.*?)<br>"), partHTML);
if (typeName.equals("")) typeName = doc.select("[rel=category tag]").text();
String year = getStrByRegex(Pattern.compile("◎年   (.*?)<br>"), partHTML);
if (year.equals("")) year = getStrByRegex(Pattern.compile("首播:(.*?)<br>"), partHTML);
String area = getStrByRegex(Pattern.compile("◎产   (.*?)<br>"), partHTML);
if (area.equals("")) area = getStrByRegex(Pattern.compile("地区:(.*?)<br>"), partHTML);
String remark = getStrByRegex(Pattern.compile("◎上映日期 (.*?)<br>"), partHTML);
String actor = getActorOrDirector(Pattern.compile("◎演   (.*?)</p>"), partHTML);
if (actor.equals("")) actor = getActorOrDirector(Pattern.compile("◎主   (.*?)</p>"), partHTML);
if (actor.equals("")) actor = getActorOrDirector(Pattern.compile("主演:(.*?)<br>"), partHTML);
String director = getActorOrDirector(Pattern.compile("◎导   (.*?)<br>"), partHTML);
if (director.equals("")) director = getActorOrDirector(Pattern.compile("导演:(.*?)<br>"), partHTML);
String description = getDescription(Pattern.compile("◎简  介(.*?)<hr>", Pattern.CASE_INSENSITIVE | Pattern.DOTALL), partHTML);
if (description.equals("")) description = getDescription(Pattern.compile("简介(.*?)</p>", Pattern.CASE_INSENSITIVE | Pattern.DOTALL), partHTML);
Vod vod = new Vod();
vod.setVodId(ids.get(0));
vod.setVodName(name);
vod.setVodPic(pic);
vod.setTypeName(typeName);
vod.setVodYear(year);
vod.setVodArea(area);
vod.setVodRemarks(remark);
vod.setVodActor(actor);
vod.setVodDirector(director);
vod.setVodContent(description);
if (playMap.size() > 0) {
vod.setVodPlayFrom(TextUtils.join("$$$", playMap.keySet()));
vod.setVodPlayUrl(TextUtils.join("$$$", playMap.values()));
}
return Result.string(vod);
}
private String getStrByRegex(Pattern pattern, String str) {
try {
Matcher matcher = pattern.matcher(str);
if (matcher.find()) {
return matcher.group(1).trim();
}
} catch (Exception e) {
SpiderDebug.log(e);
}
return "";
}
private String getActorOrDirector(Pattern pattern, String str) {
return getStrByRegex(pattern, str)
.replaceAll("<br>", "")
.replaceAll("&nbsp;", "")
.replaceAll("&amp;", "")
.replaceAll("middot;", "・")
.replaceAll("     ", ",")
.replaceAll("      ", ",");
}
private String getDescription(Pattern pattern, String str) {
return getStrByRegex(pattern, str)
.replaceAll("</?[^>]+>", "")
.replaceAll("\n", "")
.replaceAll("    ", "")
.replaceAll("&amp;", "")
.replaceAll("middot;", "・")
.replaceAll("ldquo;", "【")
.replaceAll("rdquo;", "】");
}
@Override
public String searchContent(String key, boolean quick) throws Exception {
return searchContent(key, quick, "1");
}
@Override
public String searchContent(String key, boolean quick, String pg) throws Exception {
String searchURL = siteURL + "/e/search/index.php";
String html;
if (pg.equals("1")) {
RequestBody formBody = new FormBody.Builder()
.add("show", "title")
.add("tempid", "1")
.add("tbname", "article")
.add("mid", "1")
.add("dopost", "search")
.add("submit", "")
.addEncoded("keyboard", key)
.build();
Request request = new Request.Builder()
.url(searchURL)
.addHeader("User-Agent", Utils.CHROME)
.addHeader("Origin", siteURL)
.addHeader("Referer", siteURL + "/")
.post(formBody)
.build();
Response response = OkHttp.client().newCall(request).execute();
if (response.body() == null) return "";
initNextSearchURL(response);
html = response.body().string();
response.close();
} else {
int page = Integer.parseInt(pg) - 1;
searchURL = nextSearchURLPrefix + page + nextSearchURLSuffix;
html = OkHttp.string(searchURL, getHeader());
}
return Result.string(parseVodListFromDoc(Jsoup.parse(html)));
}
private void initNextSearchURL(Response response) {
String[] split = String.valueOf(response.request().url()).split("\\?searchid=");
nextSearchURLPrefix = split[0] + "index.php?page=";
nextSearchURLSuffix = "&searchid=" + split[1];
}
@Override
public String playerContent(String flag, String id, List<String> vipFlags) throws Exception {
return Result.get().url(id).string();
}
}
Loading…
Cancel
Save