aboutsummaryrefslogtreecommitdiff
path: root/src/Scraping/Scraping.c
diff options
context:
space:
mode:
Diffstat (limited to 'src/Scraping/Scraping.c')
-rw-r--r--src/Scraping/Scraping.c131
1 files changed, 98 insertions, 33 deletions
diff --git a/src/Scraping/Scraping.c b/src/Scraping/Scraping.c
index b81c216..6d94450 100644
--- a/src/Scraping/Scraping.c
+++ b/src/Scraping/Scraping.c
@@ -1,9 +1,11 @@
#include "Scraping.h"
#include "../Cache/Cache.h"
#include "../Proxy/Proxy.h"
+#include "../Utility/XmlHelper.h"
#include "Config.h"
#include <curl/curl.h>
#include <libxml/HTMLparser.h>
+#include <libxml/parser.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
@@ -24,7 +26,8 @@ static int response_is_startpage_captcha(const ScrapeJob *job,
return response_contains(response, "<title>Startpage Captcha</title>") ||
response_contains(response, "Startpage Captcha") ||
- response_contains(response, "/static-pages-assets/page-data/captcha/") ||
+ response_contains(response,
+ "/static-pages-assets/page-data/captcha/") ||
response_contains(response, ">Startpage Blocked</title>");
}
@@ -39,7 +42,8 @@ static int response_looks_like_results_page(const ScrapeJob *job,
}
if (strcmp(job->engine->name, "Startpage") == 0) {
- return response_contains(response, "<title>Startpage Search Results</title>") ||
+ return response_contains(response,
+ "<title>Startpage Search Results</title>") ||
response_contains(response, "class=\"w-gl") ||
response_contains(response, "data-testid=\"gl-title-link\"");
}
@@ -55,6 +59,11 @@ static int response_looks_like_results_page(const ScrapeJob *job,
response_contains(response, "Mojeek Search");
}
+ if (strcmp(job->engine->name, "YaCy") == 0) {
+ return response_contains(response, "<item") ||
+ response_contains(response, "<rss");
+ }
+
return 0;
}
@@ -72,18 +81,24 @@ static void classify_job_response(ScrapeJob *job, const char *response,
return;
}
- xmlDocPtr doc = htmlReadMemory(response, response_size, NULL, NULL,
- HTML_PARSE_RECOVER | HTML_PARSE_NOERROR |
- HTML_PARSE_NOWARNING);
+ xmlDocPtr doc;
+ if (job->engine->is_xml) {
+ doc = xmlReadMemory(response, response_size, NULL, NULL,
+ XML_PARSE_RECOVER | XML_PARSE_NOERROR |
+ XML_PARSE_NOWARNING);
+ } else {
+ doc = htmlReadMemory(response, response_size, NULL, NULL,
+ HTML_PARSE_RECOVER | HTML_PARSE_NOERROR |
+ HTML_PARSE_NOWARNING);
+ }
if (!doc) {
job->status = SCRAPE_STATUS_FETCH_ERROR;
return;
}
- job->results_count =
- job->engine->parser(job->engine->name, doc, job->out_results,
- job->max_results);
+ job->results_count = job->engine->parser(job->engine->name, doc,
+ job->out_results, job->max_results);
xmlFreeDoc(doc);
if (job->results_count > 0) {
@@ -156,14 +171,26 @@ void parse_and_cache_response(ScrapeJob *job) {
}
}
-void cleanup_job_handle(ScrapeJob *job, CURL *handle) {
+static void cleanup_job_results(ScrapeJob *job) {
+ if (!job || !job->out_results)
+ return;
+
+ xml_result_free(*job->out_results, job->results_count);
+ *job->out_results = NULL;
+ job->results_count = 0;
+}
+
+static void cleanup_job_handle(ScrapeJob *job, CURL *handle) {
struct curl_slist *headers = NULL;
- curl_easy_getinfo(handle, CURLINFO_PRIVATE, &headers);
+ if (handle)
+ curl_easy_getinfo(handle, CURLINFO_PRIVATE, &headers);
if (headers)
curl_slist_free_all(headers);
free(job->response.memory);
job->response.memory = NULL;
+ job->response.size = 0;
+ job->response.capacity = 0;
}
void process_response(ScrapeJob *job, CURL *handle, CURLMsg *msg) {
@@ -180,19 +207,25 @@ void process_response(ScrapeJob *job, CURL *handle, CURLMsg *msg) {
}
int setup_job(ScrapeJob *job, CURLM *multi_handle) {
- if (job->handle)
+ if (job->handle) {
+ cleanup_job_handle(job, job->handle);
curl_easy_cleanup(job->handle);
- if (job->response.memory)
+ job->handle = NULL;
+ } else if (job->response.memory) {
free(job->response.memory);
+ job->response.memory = NULL;
+ job->response.size = 0;
+ job->response.capacity = 0;
+ }
- job->results_count = 0;
+ cleanup_job_results(job);
job->http_status = 0;
job->status = SCRAPE_STATUS_PENDING;
- if (check_cache_for_job(job)) {
- job->results_count = job->results_count > 0 ? job->results_count : 0;
+ if (check_cache_for_job(job))
return 0;
- }
+
+ cleanup_job_results(job);
char *encoded_query = curl_easy_escape(NULL, job->query, 0);
if (!encoded_query) {
@@ -200,12 +233,17 @@ int setup_job(ScrapeJob *job, CURLM *multi_handle) {
return -1;
}
- for (char *p = encoded_query + strlen(encoded_query) - 3; p >= encoded_query; p--) {
- if (p[0] == '%' && p[1] == '2' && p[2] == '0') {
- *p = '+';
- memmove(p + 1, p + 3, strlen(p + 3) + 1);
+ char *read = encoded_query;
+ char *write = encoded_query;
+ while (*read) {
+ if (read[0] == '%' && read[1] == '2' && read[2] == '0') {
+ *write++ = '+';
+ read += 3;
+ } else {
+ *write++ = *read++;
}
}
+ *write = '\0';
char *full_url =
build_search_url(job->engine->base_url, job->engine->page_param,
@@ -228,6 +266,14 @@ int setup_job(ScrapeJob *job, CURLM *multi_handle) {
job->response.memory = (char *)malloc(INITIAL_BUFFER_SIZE);
job->response.size = 0;
job->response.capacity = INITIAL_BUFFER_SIZE;
+ if (!job->response.memory) {
+ curl_easy_cleanup(job->handle);
+ job->handle = NULL;
+ job->response.capacity = 0;
+ free(full_url);
+ job->status = SCRAPE_STATUS_FETCH_ERROR;
+ return -1;
+ }
struct curl_slist *headers =
build_request_headers(job->engine->host_header, job->engine->referer);
@@ -236,10 +282,33 @@ int setup_job(ScrapeJob *job, CURLM *multi_handle) {
curl_easy_setopt(job->handle, CURLOPT_PRIVATE, headers);
free(full_url);
- curl_multi_add_handle(multi_handle, job->handle);
+ CURLMcode add_result = curl_multi_add_handle(multi_handle, job->handle);
+ if (add_result != CURLM_OK) {
+ cleanup_job_handle(job, job->handle);
+ curl_easy_cleanup(job->handle);
+ job->handle = NULL;
+ job->status = SCRAPE_STATUS_FETCH_ERROR;
+ return -1;
+ }
return 0;
}
+static void cleanup_unfinished_jobs(CURLM *multi_handle, ScrapeJob *jobs,
+ int num_jobs) {
+ for (int i = 0; i < num_jobs; i++) {
+ if (!jobs[i].handle)
+ continue;
+
+ curl_multi_remove_handle(multi_handle, jobs[i].handle);
+ cleanup_job_handle(&jobs[i], jobs[i].handle);
+ curl_easy_cleanup(jobs[i].handle);
+ jobs[i].handle = NULL;
+
+ if (jobs[i].status == SCRAPE_STATUS_PENDING)
+ jobs[i].status = SCRAPE_STATUS_FETCH_ERROR;
+ }
+}
+
int handle_responses(CURLM *multi_handle, ScrapeJob *jobs, int num_jobs) {
CURLMsg *msg;
int msgs_left;
@@ -279,34 +348,30 @@ int should_retry(ScrapeJob *jobs, int num_jobs) {
int scrape_engines_parallel(ScrapeJob *jobs, int num_jobs) {
int retries = 0;
-retry:
- ;
+retry:;
CURLM *multi_handle = curl_multi_init();
if (!multi_handle)
return -1;
for (int i = 0; i < num_jobs; i++) {
- if (setup_job(&jobs[i], multi_handle) != 0 && jobs[i].handle) {
- curl_multi_remove_handle(multi_handle, jobs[i].handle);
- curl_easy_cleanup(jobs[i].handle);
- jobs[i].handle = NULL;
- }
+ setup_job(&jobs[i], multi_handle);
}
http_delay();
int still_running = 0;
- curl_multi_perform(multi_handle, &still_running);
+ CURLMcode mc = curl_multi_perform(multi_handle, &still_running);
- do {
+ while (mc == CURLM_OK && still_running) {
int numfds = 0;
- CURLMcode mc = curl_multi_wait(multi_handle, NULL, 0, 1000, &numfds);
+ mc = curl_multi_wait(multi_handle, NULL, 0, 1000, &numfds);
if (mc != CURLM_OK)
break;
- curl_multi_perform(multi_handle, &still_running);
- } while (still_running);
+ mc = curl_multi_perform(multi_handle, &still_running);
+ }
handle_responses(multi_handle, jobs, num_jobs);
+ cleanup_unfinished_jobs(multi_handle, jobs, num_jobs);
curl_multi_cleanup(multi_handle);
if (retries < max_proxy_retries && should_retry(jobs, num_jobs)) {
generated by cgit v1.3 (git 2.53.0) at 2026-09-07 07:53:28 +0000