/** Implementation method */
public List<CCCAMEntity> getLines() {
List<CCCAMEntity> clines = new ArrayList<CCCAMEntity>();
try {
Response res =
Jsoup.connect(BASE_URL)
.data("user", "RDS580" + System.currentTimeMillis())
.data("pass", "RDS580")
.data("submit", "Activate!")
.userAgent(
"Mozilla/5.0 (Windows NT 10.0; WOW64; rv:43.0) Gecko/20100101 Firefox/43.0")
.referrer(BASE_URL)
.method(Method.POST)
.execute();
final String linesweb = res.body();
String lineSearch1 = "C: ";
String lineSearch2 = " :|: and it will ";
String line =
linesweb.substring(
linesweb.indexOf(lineSearch1) + lineSearch1.length(),
linesweb.indexOf(lineSearch2, linesweb.indexOf(lineSearch1) + lineSearch1.length()));
line = line.trim();
final String[] tokens = line.split(" ");
final String host = tokens[0].trim();
final String port = tokens[1].trim();
final String user = tokens[2].trim();
final String pass = tokens[3].trim();
clines.add(new CCCAMEntity(host, port, user, pass, default_hops));
} catch (Exception e) {
System.out.println("Error en " + BASE_URL + ". " + e.getMessage());
}
return clines;
}
/**
* Parse cn page and write in hbase
*
* @param symbol
*/
public static void parseCNSymbols(String symbol) {
if (!Hbase.getData(symbol).equals("")) {
// System.out.println(symbol + " Exists!");
return;
}
String url = "http://xueqiu.com/S/" + symbol + "/historical.csv";
Response rs = null;
// System.out.println(url);
try {
Connection con = getConnection(url, "historyHttp");
con.header("Referer", " http://xueqiu.com/S/" + symbol);
rs = con.execute();
// System.out.println(rs.body());
} catch (IOException e1) {
if (handleError) {
System.out.println(symbol + " http error");
errors.add(symbol);
} else {
WriteError(symbol);
System.out.println(symbol + " http error");
}
return;
}
try {
BufferedReader reader = new BufferedReader(new StringReader(rs.body())); // 换成你的文件名
reader.readLine(); // 第一行信息，为标题信息，不用，如果需要，注释掉
String line = null;
JSONArray HistoricalData = new JSONArray();
List<JSONArray> jsonLists = new ArrayList<JSONArray>();
while ((line = reader.readLine()) != null) {
String item[] = line.split(","); // CSV格式文件为逗号分隔符文件，这里根据逗号切分
// System.out.println(item[0]);
JSONArray DailyData = new JSONArray();
for (int i = 1; i < item.length; i++) {
item[i] = item[i].replace("\"", "");
DailyData.put(item[i]);
}
if (Double.valueOf(DailyData.getString(2)) != 0) {
jsonLists.add(DailyData);
}
}
for (int i = (jsonLists.size() - 1); i >= 0; i--) {
HistoricalData.put(jsonLists.get(i));
}
Hbase.addData(symbol, type, HistoricalData.toString());
// System.out.println(symbol + " done");
// System.out.println(jsonLists);
} catch (Exception e) {
if (handleError) {
System.out.println(symbol + " parsing error");
errors.add(symbol);
} else {
WriteError(symbol);
System.out.println(symbol + " parsing error");
}
}
}
public static Map<String, String> login(
HttpServletRequest req, HttpServletResponse res, String username, String password)
throws Exception {
String url = ReadProperties.getByName("login.ip") + "/login";
Map<String, String> datas = new HashMap<String, String>();
Map<String, String> cookies = new HashMap<String, String>();
Connection con = Jsoup.connect(url).timeout(120000); // 获取连接
con.header(
"User-Agent",
"Mozilla/5.0 (Windows NT 6.1; WOW64; rv:29.0) Gecko/20100101 Firefox/29.0"); // 配置模拟浏览器
Response rs;
rs = con.execute();
cookies = rs.cookies();
Document doc = Jsoup.parse(rs.body()); // 转换为Dom树
List<Element> et = doc.select("form"); // 获取form表单，可以通过查看页面源码代码得知
for (Element e : et.get(0).getAllElements()) {
if (e.attr("name").equals("username")) {
e.attr("value", username); // 设置用户名
}
if (e.attr("name").equals("password")) {
e.attr("value", password); // 设置用户密码
}
if (e.attr("name").length() > 0) { // 排除空值表单属性
datas.put(e.attr("name"), e.attr("value"));
}
}
// 设置cookie和post上面的map数据
Response login = null;
login = con.data(datas).cookies(cookies).method(Method.POST).execute();
url = ReadProperties.getByName("common.ip") + req.getContextPath() + "/user/getUser";
con =
Jsoup.connect(url)
.cookies(login.cookies())
.ignoreContentType(true)
.method(Method.GET); // 获取连接
rs = con.execute();
for (Entry<String, String> entry : rs.cookies().entrySet()) {
Cookie cookie = new Cookie(entry.getKey(), entry.getValue());
cookie.setPath(req.getContextPath() + "/");
res.addCookie(cookie);
}
return JsonUtil.jsonToObject(rs.body(), Map.class);
}
public boolean loginByEmailAndPwd() {
loginCookies.clear();
Scanner sc = new Scanner(System.in);
getCaptchaImgAndCookies(0);
log.info("请输入账号:");
email = sc.nextLine();
log.info("请输入密码");
password = sc.nextLine();
log.info("查看验证码并输入");
captcha = sc.nextLine();
Connection con = JsoupUtil.getPostCon("https://www.zhihu.com/login/email");
Response rs = null;
try {
rs =
con.data("_xsrf", xsrf)
.data("email", email)
.data("password", password)
.data("remember_me", remeberMe)
.data("captcha", captcha)
.cookies(captchaCookies)
.ignoreContentType(true)
.execute();
} catch (IOException e) {
e.printStackTrace();
log.info("通过账号密码登录发生异常");
return false;
}
JSONObject jsonObject = new JSONObject(rs.body());
String result = jsonObject.get("r").toString();
log.info(EzraPoundUtil.unicode2Character(jsonObject.get("msg").toString()));
Response rs2 = null;
try {
rs2 = JsoupUtil.getGetCon("https://www.zhihu.com").cookies(rs.cookies()).execute();
} catch (IOException e) {
e.printStackTrace();
}
if (checkLogin(Jsoup.parse(rs2.body()))) {
loginCookies.putAll(rs.cookies());
saveCookies(EzraPoundUtil.LOGIN_COOKIES_DIR, loginCookies);
return true;
}
return false;
}
private AccessToken getAccessToken() {
try {
Response response =
Jsoup.connect(GET_URL).ignoreContentType(true).method(Method.GET).execute();
String json = response.body();
Gson gson = new Gson();
return gson.fromJson(json, new TypeToken<AccessToken>() {}.getType());
} catch (Exception e) {
log.error("[GetAccessTokenTask] getAccessToken Error:", e);
return null;
}
}
public static void main(String[] args) throws Exception {
int index = 0;
long timeout = 5000;
String pattern = "MMdd_HH_mm_ss";
String dataDir = "./data/" + YhdPriceMonitor.class.getSimpleName().toLowerCase();
disableSSLCertCheck();
String dateChars = DateFormatUtils.format(new Date(), pattern);
int lastErrorCount = 0;
while (true) {
index++;
String url =
"http://gps.yhd.com/restful/detail?mcsite=1&provinceId=1&pmId=41909728&callback=jQuery111304328004347221549_1447325832073&_="
+ System.currentTimeMillis();
long start = System.currentTimeMillis();
Response resp = doRequest(url, 3);
long cost = System.currentTimeMillis() - start;
if (resp == null) {
log.warn("error,index:" + index + ",status:null,cost:" + cost + ",url:" + url);
} else {
log.info(
"done,index:"
+ index
+ ",status:"
+ resp.statusCode()
+ ",cost:"
+ cost
+ ",url:"
+ url);
}
boolean success = doValidateResponse(resp);
if (success) {
lastErrorCount = 0;
log.info("validate=true.index:" + index + ",cookies:" + JSON.toJSONString(resp.cookies()));
log.info("validate=true.index:" + index + ",headers:" + JSON.toJSONString(resp.headers()));
} else {
lastErrorCount++;
if (resp != null) {
log.warn(
"validate=false.index:" + index + ",cookies:" + JSON.toJSONString(resp.cookies()));
log.warn(
"validate=false.index:" + index + ",cookies:" + JSON.toJSONString(resp.headers()));
}
}
FileUtils.writeStringToFile(
new File(dataDir, dateChars + File.separator + index + ".html"),
resp == null ? "no response" : resp.body());
if (lastErrorCount >= 10) {
break;
}
TimeUnit.MILLISECONDS.sleep(timeout);
}
}
public boolean loginBySavedCookies() {
loginCookies.clear();
readCookies(EzraPoundUtil.LOGIN_COOKIES_DIR, loginCookies);
Connection con = JsoupUtil.getGetCon("https://www.zhihu.com");
Response rs = null;
try {
rs = con.cookies(loginCookies).execute();
} catch (IOException e) {
e.printStackTrace();
log.info("携带cookie登录测试失败");
return false;
}
return checkLogin(Jsoup.parse(rs.body()));
}
private String getContent(Class response, String... strings) throws Exception {
Map<String, String> data = new HashMap<String, String>();
for (int i = 0; i < strings.length; i = i + 2) data.put(strings[i], strings[i + 1]);
Response res = null;
System.out.println("Start connect: " + response.getName() + ", " + new Date());
while (res == null) {
try {
res = Jsoup.connect(url).data(data).timeout(0).method(Method.POST).execute();
} catch (IOException e) {
System.out.println("exception: " + response.getName() + ", " + new Date());
}
}
System.out.println("Finish connect: " + response.getName() + ", " + new Date());
return changeEncoding(res.body());
}
private static boolean doValidateResponse(Response resp) {
if (resp == null) {
return false;
}
String body = resp.body();
List<String> containList = new ArrayList<String>();
containList.add("marketPrice");
containList.add("89");
containList.add("41909728");
for (String sVal : containList) {
if (body.indexOf(sVal) < 0) {
return false;
}
}
return true;
}
public boolean getXsrf(int times) {
if (times > maxRecursiveTimes) return false;
Connection con = JsoupUtil.getGetCon("http://www.zhihu.com");
Response rs = null;
try {
rs = con.execute();
} catch (IOException e) {
e.printStackTrace();
log.info("获取_xsrf第" + times + "次失败");
return getXsrf(++times);
}
Document doc = Jsoup.parse(rs.body());
xsrf = doc.select(".view.view-signin [name=\"_xsrf\"]").attr("value");
log.info("已获得xsrf:" + xsrf);
return true;
}
/** Implementation method */
public List<CCCAMEntity> getLines() {
List<CCCAMEntity> clines = new ArrayList<CCCAMEntity>();
try {
Response res =
Jsoup.connect(BASE_URL)
.timeout(7500)
.userAgent(
"Mozilla/5.0 (Windows NT 10.0; WOW64; rv:43.0) Gecko/20100101 Firefox/43.0")
.method(Method.GET)
.execute();
Document doc = Jsoup.parse(res.body());
final String[] tokens = doc.getElementsByTag("h1").get(0).text().split(" ");
final String host = tokens[1].trim();
final String port = tokens[2].trim();
final String user = tokens[3].trim();
final String pass = tokens[4].trim();
clines.add(new CCCAMEntity(host, port, user, pass, default_hops));
} catch (Exception e) {
System.out.println("Error en " + BASE_URL);
System.out.println("Error: " + e.getMessage());
}
return clines;
}
