#include "de_mrbesen_youtubecrawler_CrawlerThread.h"

#include <iostream>
#include <set>
#include <string>
#include <sstream>
#include <regex>
#include <vector>

#include <curl/curl.h>


static const std::string YOUTUBEBASE = "https://youtube.com/watch?v=";
static const std::regex YOUTUBELINKPATTERN("watch\\?v=([-_a-zA-Z0-9]{11})");

std::vector<CURL*> curls;

static size_t WriteCallback(void* contents, size_t size, size_t nmemb, void* userp) {
	*((std::ostringstream*) userp) << std::string((char*) contents, size * nmemb);
	return size * nmemb;
}

JNIEXPORT void JNICALL Java_de_mrbesen_youtubecrawler_CrawlerThread_initLib(JNIEnv* env, jclass clazz, jint threadcount) {
	curls.reserve(threadcount);
	for(int i = 0; i < threadcount; ++i) {
		CURL* curl = curl_easy_init();
		curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1);
		curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, WriteCallback);
		curls.push_back(curl);
	}
}

JNIEXPORT void JNICALL Java_de_mrbesen_youtubecrawler_CrawlerThread_deinitLib(JNIEnv* env, jclass) {
	for(int i = 0; i < curls.size(); ++i) {
		CURL* curl = curls.at(i);
		curl_easy_cleanup(curl);
	}
	curls.clear();
}

std::string download(CURL* curl, const std::string& url) {
	curl_easy_setopt(curl, CURLOPT_URL, url.c_str());

	std::ostringstream out;
	curl_easy_setopt(curl, CURLOPT_WRITEDATA, &out);

	CURLcode res = curl_easy_perform(curl);
	long responsecode = 404;
	curl_easy_getinfo(curl, CURLINFO_RESPONSE_CODE, &responsecode);
	if(responsecode != 200) {
		std::cout << "Curl error: got " << responsecode << std::endl;
		return "";
	}
	if(res != CURLE_OK) {
		std::cout << "Curl error: " << res << std::endl;
		return "";
	}

	return out.str();
}

JNIEXPORT jboolean JNICALL Java_de_mrbesen_youtubecrawler_CrawlerThread_crawl(JNIEnv* env, jobject that, jstring videoid, jint threadid) {
	// get videoid argument
	jboolean myfalseval = false;
	const char* cvideid = env->GetStringUTFChars(videoid, &myfalseval);
	const std::string svideoid(cvideid); // makes a copy
	env->ReleaseStringUTFChars(videoid, cvideid);

	// std::cout << "crawl: " << svideoid << std::endl;

	// use curl to get the website
	CURL* curl = curls.at(threadid);
	std::string webcontent = download(curl, YOUTUBEBASE + svideoid);
	if(webcontent.empty()) {
		std::cout << "webcontent is empty" << std::endl;
		return JNI_FALSE;
	}

	jclass jclass_ct = env->FindClass("de/mrbesen/youtubecrawler/CrawlerThread"); // class of CrawlerThread
	jfieldID fid_ctfound = env->GetFieldID(jclass_ct, "found" , "Ljava/util/List;"); // fieldID of CrawlerThread.found
	jclass jclass_ll = env->FindClass("java/util/List"); // Linked List Class
	jmethodID mid_add = env->GetMethodID(jclass_ll, "add", "(Ljava/lang/Object;)Z"); // add() method of LinkedList
	jobject ll_found = env->GetObjectField(that, fid_ctfound); // linked list to store the results to

	// match regex
	std::sregex_iterator it = std::sregex_iterator(webcontent.begin(), webcontent.end(), YOUTUBELINKPATTERN);
	const std::sregex_iterator itend = std::sregex_iterator();

	// this set is to find duplicated strings it does not catch everything, because it is only valid during this crawl, but it should filter a few with very little effort
	std::set<std::string> known;
	known.insert(svideoid); // do not "find" the same ytid

	env->MonitorEnter(that); // syncronized(this) {
	for( ; it != itend; ++it) {
		const std::string ytid = (*it)[1].str(); // get the String from the first sub-group match

		// only if the ytid is not known
		if(known.find(ytid) == known.end()) {
			// add to the found list
			// std::cout << ytid << std::endl;

			jstring jytid = env->NewStringUTF(ytid.c_str()); // create a java string object
			jboolean b = env->CallBooleanMethod(ll_found, mid_add, jytid); // call add() on the LinkedList object with the String

			known.insert(ytid);
		}
	}
	env->MonitorExit(that); // end of syncronized block

	return JNI_TRUE;
}