static void ParseArticle(rssDatabase *db, const char *articleTitle, const char *articleURL)
{
url u;
urlconnection urlconn;
streamtokenizer st;
int articleID;
URLNewAbsolute(&u, articleURL);
rssNewsArticle newsArticle = { articleTitle, u.serverName, u.fullName };
pthread_mutex_t *articlesLock = &(db->locks.articlesVectorLock);
pthread_mutex_lock(articlesLock);
if (VectorSearch(&db->previouslySeenArticles, &newsArticle, NewsArticleCompare, 0, false) >= 0) {
pthread_mutex_unlock(articlesLock);
printf("[Ignoring \"%s\": we've seen it before.]\n", articleTitle);
URLDispose(&u);
return;
}
pthread_mutex_unlock(articlesLock);
lockConnection(db,u.serverName);
URLConnectionNew(&urlconn, &u);
switch (urlconn.responseCode) {
case 0: printf("Unable to connect to \"%s\". Domain name or IP address is nonexistent.\n", articleURL);
break;
case 200: //printf("[%s] Ready to Index \"%s\"\n", u.serverName, articleTitle);
pthread_mutex_lock(articlesLock);
printf("[%s] Indexing \"%s\"\n", u.serverName, articleTitle);
NewsArticleClone(&newsArticle, articleTitle, u.serverName, u.fullName);
VectorAppend(&db->previouslySeenArticles, &newsArticle);
articleID = VectorLength(&db->previouslySeenArticles) - 1;
pthread_mutex_unlock(articlesLock);
STNew(&st, urlconn.dataStream, kTextDelimiters, false);
ScanArticle(&st, articleID, &db->indices, &db->stopWords,
&(db->locks.indicesHashSetLock),&(db->locks.stopWordsHashSetLock) );
STDispose(&st);
break;
case 301:
case 302:{ // just pretend we have the redirected URL all along, though index using the new URL and not the old one...
int newURLLength = strlen(urlconn.newUrl)+1;
char newURLBuffer[newURLLength];
strcpy(newURLBuffer, urlconn.newUrl);
URLConnectionDispose(&urlconn);
unlockConnection(db,u.serverName);
URLDispose(&u);
ParseArticle(db, articleTitle, newURLBuffer);
return;
} default: printf("Unable to pull \"%s\" from \"%s\". [Response code: %d] Punting...\n", articleTitle, u.serverName, urlconn.responseCode);
break;
}
URLConnectionDispose(&urlconn);
unlockConnection(db,u.serverName);
URLDispose(&u);
}
static void ParseArticle(void *userData)
{
rssFeedData *data = userData;
rssFeedItem *item = &data->item;
url u;
urlconnection urlconn;
streamtokenizer st;
URLNewAbsolute(&u, item->url);
URLConnectionNew(&urlconn, &u);
switch (urlconn.responseCode) {
case 0: printf("Unable to connect to \"%s\". Domain name or IP address is nonexistent.\n", item->url);
break;
case 200: printf("[%s] Indexing \"%s\"\n", u.serverName, item->title);
STNew(&st, urlconn.dataStream, kTextDelimiters, false);
ScanArticle(&st, data);
STDispose(&st);
break;
case 301:
case 302:
case 303: // just pretend we have the redirected URL, though index using the new URL and not the old one...
strcpy(item->url,urlconn.newUrl);
ParseArticle(data);
break;
default: printf("Unable to pull \"%s\" from \"%s\". [Response code: %d] Punting...\n", item->title, u.fullName, urlconn.responseCode);
break;
}
URLConnectionDispose(&urlconn);
URLDispose(&u);
}
static void* PthreadParseArticle(void * threadData){
threadArguments *arg = threadData;
ParseArticle(arg->db ,arg->title,arg->URL);
pthread_exit(NULL);
}
static void ParseArticle(const char *articleTitle, const char *articleDescription, const char *articleURL,
hashset *stopWords, hashset *wordCounts)
{
url u;
urlconnection urlconn;
streamtokenizer st;
URLNewAbsolute(&u, articleURL);
URLConnectionNew(&urlconn, &u);
switch (urlconn.responseCode) {
case 0: printf("Unable to connect to \"%s\". Domain name or IP address is nonexistent.\n", articleURL);
break;
case 200: printf("Scanning \"%s\" from \"http://%s\"\n", articleTitle, u.serverName);
STNew(&st, urlconn.dataStream, kTextDelimiters, false);
ScanArticle(&st, articleTitle, articleDescription, articleURL, stopWords, wordCounts);
STDispose(&st);
break;
case 301:
case 302: // just pretend we have the redirected URL all along, though index using the new URL and not the old one...
ParseArticle(articleTitle, articleDescription, urlconn.newUrl, stopWords, wordCounts);
break;
default: printf("Unable to pull \"%s\" from \"%s\". [Response code: %d] Punting...\n", articleTitle, u.serverName, urlconn.responseCode);
break;
}
URLConnectionDispose(&urlconn);
URLDispose(&u);
}
static void ProcessEndTag(void *userData, const char *name)
{
rssFeedData *data = userData;
rssFeedItem *item = &data->item;
item->activeField = NULL;
if (strcasecmp(name, "item") == 0)
ParseArticle(userData);
}
static void ParseArticle(const char *articleTitle,const char *articleDescription,
const char *articleURL,hashset* stopWords,
hashset* wordHash,hashset *articlesSeen)
{
url u;
urlconnection urlconn;
streamtokenizer st;
URLNewAbsolute(&u, articleURL);
URLConnectionNew(&urlconn, &u);
article currArt;
currArt.server = strdup(u.serverName);
currArt.title = strdup(articleTitle);
currArt.url = strdup(articleURL);
currArt.numOccurrences = 0;
switch (urlconn.responseCode) {
case 0: printf("Unable to connect to \"%s\". Domain name or IP address is nonexistent.\n", articleURL);
ArticleFree(&currArt);
break;
case 200:
if(HashSetLookup(articlesSeen,&currArt)== NULL){ //if we haven't seen this article before
printf("[%s] Indexing \"%s\"\n", u.serverName,articleTitle);
HashSetEnter(articlesSeen, &currArt);
STNew(&st, urlconn.dataStream, kTextDelimiters, false);
ScanArticle(&st, &currArt, stopWords, wordHash, articlesSeen);
STDispose(&st);
break;
}
else { //if we have seen it before
printf("[Ignoring \"%s\": we've seen it before.]\n",
articleTitle);
ArticleFree(&currArt);
break;
}
case 301:
case 302: // just pretend we have the redirected URL all along,though index using the new URL and not the old one...
ParseArticle(articleTitle, articleDescription,
urlconn.newUrl, stopWords, wordHash,articlesSeen);
ArticleFree(&currArt);
break;
default: printf("Unable to pull \"%s\" from \"%s\". [Response code: %d] Punting...\n", articleTitle, u.serverName, urlconn.responseCode);
ArticleFree(&currArt);
break;
}
URLConnectionDispose(&urlconn);
URLDispose(&u);
}
static void ParseArticle(const char *articleURL, const char *articleTitle, rssData *allData)
{
url u;
urlconnection urlconn;
streamtokenizer st;
int articleIndex;
URLNewAbsolute(&u, articleURL);
/* check to see if we've previously scanned the article. If the article we're processing
* has already been scanned release the url and return
*/
article a = {articleURL, articleTitle, u.serverName};
if(VectorSearch(&allData->explored, &a, ArticleCmp, 0, false) >= 0) {
printf("[Pass. article already indexed: \"%s\"]\n", articleTitle);
URLDispose(&u);
return;
}
URLConnectionNew(&urlconn, &u);
switch (urlconn.responseCode) {
case 0: printf("Unable to connect to \"%s\". Domain name or IP address is nonexistent.\n", articleURL);
break;
case 200: printf("Scanning \"%s\" from \"http://%s\"\n", articleTitle, u.serverName);
STNew(&st, urlconn.dataStream, kTextDelimiters, false);
PersistArticle(&a, articleURL, articleTitle, u.serverName);
VectorAppend(&allData->explored, &a);
articleIndex = VectorLength(&allData->explored)-1;
ScanArticle(&st, &a, articleIndex, allData);
STDispose(&st);
break;
case 301:
case 302: // just pretend we have the redirected URL all along, though index using the new URL and not the old one...
ParseArticle(urlconn.newUrl, articleTitle, allData );
break;
default: printf("Unable to pull \"%s\" from \"%s\". [Response code: %d] Punting...\n", articleTitle, u.serverName, urlconn.responseCode);
break;
}
URLConnectionDispose(&urlconn);
URLDispose(&u);
}
static void ProcessSingleNewsItem(streamtokenizer *st, hashset *stopWords, hashset *prevSeenArticles, hashset *wordCounts)
{
char htmlTag[1024];
char articleTitle[1024];
char articleDescription[1024];
char articleURL[1024];
articleTitle[0] = articleDescription[0] = articleURL[0] = '\0';
while (GetNextTag(st, htmlTag, sizeof(htmlTag)) && (strcasecmp(htmlTag, kItemEndTag) != 0)) {
if (strncasecmp(htmlTag, kTitleTagPrefix, strlen(kTitleTagPrefix)) == 0) ExtractElement(st, htmlTag, articleTitle, sizeof(articleTitle));
if (strncasecmp(htmlTag, kDescriptionTagPrefix, strlen(kDescriptionTagPrefix)) == 0) ExtractElement(st, htmlTag, articleDescription, sizeof(articleDescription));
if (strncasecmp(htmlTag, kLinkTagPrefix, strlen(kLinkTagPrefix)) == 0) ExtractElement(st, htmlTag, articleURL, sizeof(articleURL));
}
if (strncmp(articleURL, "", sizeof(articleURL)) == 0) return; // punt, since it's not going to take us anywhere
if (isNewArticle(prevSeenArticles, articleTitle, articleURL)) {
ParseArticle(articleTitle, articleDescription, articleURL, stopWords, wordCounts);
}
else {
printf("Skipping previously seen article: \"%s\"\n\tfrom \"%s\"\n", articleTitle, articleURL);
}
}
static void ProcessSingleNewsItem(streamtokenizer *st, hashset *stopWords,
hashset * wordHash, hashset *articlesSeen)
{
char htmlTag[1024];
char articleTitle[1024];
char articleDescription[1024];
char articleURL[1024];
articleTitle[0] = articleDescription[0] = articleURL[0] = '\0';
while (GetNextTag(st, htmlTag, sizeof(htmlTag)) && (strcasecmp(htmlTag, kItemEndTag) != 0)) {
if (strncasecmp(htmlTag, kTitleTagPrefix, strlen(kTitleTagPrefix)) == 0)
ExtractElement(st, htmlTag, articleTitle, sizeof(articleTitle));
if (strncasecmp(htmlTag, kDescriptionTagPrefix,strlen(kDescriptionTagPrefix)) == 0)
ExtractElement(st, htmlTag, articleDescription, sizeof(articleDescription));
if (strncasecmp(htmlTag, kLinkTagPrefix, strlen(kLinkTagPrefix)) == 0)
ExtractElement(st, htmlTag, articleURL, sizeof(articleURL));
}
if (strncmp(articleURL, "", sizeof(articleURL)) == 0) return; // punt, since it's not going to take us anywhere
ParseArticle(articleTitle,articleDescription,articleURL,stopWords,wordHash,articlesSeen);
}
static void ProcessSingleNewsItem(streamtokenizer *st, rssData *allData)
{
char htmlTag[1024];
char articleTitle[1024];
char articleDescription[1024];
char articleURL[1024];
articleTitle[0] = articleDescription[0] = articleURL[0] = '\0';
int count = 0;
while (GetNextTag(st, htmlTag, sizeof(htmlTag)) && (strcasecmp(htmlTag, kItemEndTag) != 0)) {
if (strncasecmp(htmlTag, kTitleTagPrefix, strlen(kTitleTagPrefix)) == 0) ExtractElement(st, htmlTag, articleTitle, sizeof(articleTitle));
if (strncasecmp(htmlTag, kDescriptionTagPrefix, strlen(kDescriptionTagPrefix)) == 0) ExtractElement(st, htmlTag, articleDescription, sizeof(articleDescription));
if (strncasecmp(htmlTag, kLinkTagPrefix, strlen(kLinkTagPrefix)) == 0) ExtractElement(st, htmlTag, articleURL, sizeof(articleURL));
count++;
if (count == 5 ) break;
}
if (strncmp(articleURL, "", sizeof(articleURL)) == 0) return; // punt, since it's not going to take us anywhere
ParseArticle(articleURL, articleTitle, allData);
}