/**
* 摘要
*
* @param list
* @param dom
* @param component
* @param strings
*/
@Override
public void parseBrief(List<WeixinData> list, Node dom, Component component, String... args) {
if (args[0] == null || args[0] == "") return;
List<String> results = StringUtil.regMatches(args[0], "content168>", "/content168", true);
for (int i = 0; i < results.size(); i++) {
String tmp = results.get(i);
String result = StringUtil.regMatcher(tmp, "CDATA\\[", "\\]");
list.get(i).setBrief(result);
}
}
@Override
public void parseTitle(List<WeixinData> list, Node dom, Component component, String... args) {
if (args[0] == null || args[0] == "") return;
List<String> results = StringUtil.regMatches(args[0], "title>", "/title", true);
for (int i = 0; i < results.size(); i++) {
String tmp = results.get(i);
String result = StringUtil.regMatcher(tmp, "CDATA\\[", "\\]");
WeixinData vd = new WeixinData();
vd.setTitle(result);
list.add(vd);
}
}
@Override
public void parseSource(WeixinData data, Node dom, Component component, String... strings) {
if (component == null) return;
NodeList nl = commonList(component.getXpath(), dom);
if (nl == null) return;
if (nl.item(0) != null) data.setSource(StringUtil.format(nl.item(0).getTextContent()));
}
/**
* 来源
*
* @param list
* @param dom
* @param component
* @param strings
*/
@Override
public void parseSource(List<WeixinData> list, Node dom, Component component, String... strings) {
if (component == null) return;
NodeList nl = head(component.getXpath(), dom, list.size(), component.getName());
if (nl == null) return;
for (int i = 0; i < nl.getLength(); i++) {
list.get(i).setSource(StringUtil.format(nl.item(i).getTextContent()));
}
}
@Override
public void parseImgUrl(WeixinData data, Node dom, Component component, String... args) {
if (component == null) return;
NodeList nl = commonList(component.getXpath(), dom);
if (nl == null) return;
String imgs = "";
for (int i = 0; i < nl.getLength(); i++) {
imgs += StringUtil.format(nl.item(i).getTextContent()) + ";";
}
data.setImgUrl(imgs);
}
@Override
public void parseSource(NewsData data, Node dom, Component component, String... args) {
String str = "";
if (component == null) return;
NodeList nl = commonList(component.getXpath(), dom);
if (nl == null) return;
if (nl.item(0) != null) str = StringUtil.format(nl.item(0).getTextContent());
str = str.replace("来源：", "");
str = str.equals("") ? "慧聪橡胶网" : str;
str = str.length() > 20 ? "慧聪橡胶网" : str;
data.setSource(str);
}
@Override
public void parseAuthor(NewsData data, Node dom, Component component, String... args) {
String str = "";
if (component == null) {
return;
}
NodeList nl = commonList(component.getXpath(), dom);
if (nl == null) {
return;
}
if (nl.item(0) != null) str = StringUtil.format(nl.item(0).getTextContent());
str = str.replace("●", "").trim();
data.setSource(str);
}
@Override
public void parseUrl(List<WeixinData> list, Node dom, Component component, String... args) {
if (args[0] == null || args[0] == "" || args[1] == null || args[1] == "") return;
String cookie = args[1];
// String referer = args[1];
List<String> results = StringUtil.regMatches(args[0], "<url>", "/url", true);
for (int i = 0; i < results.size(); i++) {
String tmpUrl = results.get(i);
tmpUrl =
"http://weixin.sogou.com"
+ tmpUrl.substring(tmpUrl.indexOf("CDATA[") + 6, tmpUrl.lastIndexOf("]]>"));
String loc = null;
try {
HttpURLConnection conn = (HttpURLConnection) new URL(tmpUrl).openConnection();
conn.addRequestProperty(
"User-Agent", "Mozilla/5.0 (Windows NT 6.1; rv:38.0) Gecko/20100101 Firefox/38.0");
conn.setRequestProperty("Cookie", cookie);
// conn.setRequestProperty("Referer", referer);
HttpURLConnection.setFollowRedirects(false);
conn.setFollowRedirects(false);
conn.connect();
loc = conn.getHeaderField("Location");
if (loc != null) Systemconfig.sysLog.log(conn.getResponseMessage());
Systemconfig.sysLog.log("real url: " + loc);
int sleepTime = 30 + (int) (Math.random() * 20);
Systemconfig.sysLog.log("sleep..." + sleepTime);
TimeUtil.rest(sleepTime);
} catch (MalformedURLException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
list.get(i).setUrl(loc == null ? "err." : loc);
}
}
@Override
public void parseSource(NewsData data, Node dom, Component component, String... args) {
String str = "";
if (component == null) return;
NodeList nl = commonList(component.getXpath(), dom);
if (nl == null) return;
if (nl.item(0) != null) {
str = nl.item(0).getTextContent();
str = StringUtil.format(str);
}
if (str.contains("来源：")) str = str.substring(str.indexOf("来源：") + 3);
if (str.contains("发布日期")) str = str.substring(0, str.indexOf("发布日期")).replace(" ", "").trim();
if (str == null) str = "中国橡胶工业协会-轮胎分会";
else {
str = str.contains("本站") ? "中国橡胶工业协会-轮胎分会" : str;
}
str = str.length() > 20 ? "中国橡胶工业协会-轮胎分会" : str;
data.setSource(str);
}
public void parseNumber(WeixinData data, Node dom, Component component, String... args) {
// http://mp.weixin.qq.com/s?__biz=MjM5ODE1NTMxMQ==&mid=201653867&idx=1&sn=6f3445a3640eb09ce7cfa5a49509f165&3rd=MzA3MDU4NTYzMw==&scene=6#rd
String biz = "";
String mid = "";
String uin = "";
String key = "";
String fromFile = StringUtil.getContent("config/WeixinKey/WeixinKey.txt");
try {
biz = StringUtil.regMatcher(data.getUrl(), "__biz=", "&");
mid = StringUtil.regMatcher(data.getUrl(), "mid=", "&");
for (String string : fromFile.split("&")) {
if (string.contains("uin")) uin = string.split("=")[1].trim();
if (string.contains("key")) key = string.split("=")[1].trim();
}
} catch (Exception e) {
e.printStackTrace();
}
String url =
"http://mp.weixin.qq.com"
+ "/mp/getappmsgext?"
+ "__biz="
+ biz
+ "&mid="
+ mid
+ "&uin="
+ uin
+ "&key="
+ key
// +
// "&pass_ticket=b3hV91xTLYZxRGKemRNz%2FAi4VKElPnwHYUNtoV8w4dE%3D"
+ "";
HtmlInfo html = new HtmlInfo();
String charSet = "UTF-8";
html.setType("DATA");
html.setEncode(charSet);
html.setOrignUrl(url);
html.setCookie("Set-Cookie: wxuin=20156425; Path=/; Expires=Fri, 02-Jan-1970 00:00:00 GMT");
html.setUa(
"Mozilla/5.0 (iPhone; CPU iPhone OS 8_0 like Mac OS X) AppleWebKit/600.1.3 (KHTML, like Gecko) Version/8.0 Mobile/12A4345d Safari/600.1.4");
SimpleHttpProcess shp = new SimpleHttpProcess();
shp.getContent(html);
String content = html.getContent();
int retry = 0;
while (!content.contains("read_num")) {
if (retry++ > 3) break;
Systemconfig.sysLog.log("请获取key后输入任意内容回车继续...输入c忽略(很可能无法继续采集，不推荐)");
System.err.println("请获取key后输入任意内容回车继续...输入c忽略(很可能无法继续采集，不推荐)");
Scanner input = new Scanner(System.in);
String s = input.next();
if (s.equals("c") || s.equals("C")) break;
fromFile = StringUtil.getContent("config/WeixinKey/WeixinKey.txt");
try {
for (String string : fromFile.split("&")) {
if (string.contains("uin")) uin = string.split("=")[1].trim();
if (string.contains("key")) key = string.split("=")[1].trim();
}
} catch (Exception e) {
e.printStackTrace();
}
url =
"http://mp.weixin.qq.com"
+ "/mp/getappmsgext?"
+ "__biz="
+ biz
+ "&mid="
+ mid
+ "&uin="
+ uin
+ "&key="
+ key;
html = new HtmlInfo();
charSet = "UTF-8";
html.setType("DATA");
html.setEncode(charSet);
html.setOrignUrl(url);
html.setCookie("Set-Cookie: wxuin=20156425; Path=/; Expires=Fri, 02-Jan-1970 00:00:00 GMT");
html.setUa(
"Mozilla/5.0 (iPhone; CPU iPhone OS 8_0 like Mac OS X) AppleWebKit/600.1.3 (KHTML, like Gecko) Version/8.0 Mobile/12A4345d Safari/600.1.4");
shp = new SimpleHttpProcess();
shp.getContent(html);
content = html.getContent();
}
String readNumStr = StringUtil.regMatcher(content, "\"read_num\":", ",");
String praiseNumStr = StringUtil.regMatcher(content, "\"like_num\":", ",");
try {
if (readNumStr != null) data.setReadNum(Integer.parseInt(readNumStr));
if (praiseNumStr != null) data.setPraiseNum(Integer.parseInt(praiseNumStr));
} catch (Exception e) {
e.printStackTrace();
}
}