parse image results without re (WIP)

searx/engines/yandex.py

@@ -5,8 +5,10 @@
 import re
 from urllib.parse import urlencode, urlparse, parse_qs
 from lxml import html
+from searx.utils import humanize_bytes
 from html import unescape
 from searx import logger
+from searx import utils
 from searx.exceptions import SearxEngineCaptchaException
 from datetime import datetime
 
@@ -31,6 +33,10 @@ base_url_videos = 'https://yandex.com/video/search'
 
 url_extension = 'tmpl_version=releases%2Ffrontend%2Fvideo%2Fv1.1168.0%238d942de0f4ebc4eb6b8f3c24ffbd1f8dbc5bbe63'
 
+url_xpath = "//a[@class='b-serp-item__title-link']/@href"
+title_xpath = "//h3[@class='b-serp-item__title']/a[@class='b-serp-item__title-link']/span"
+content_xpath = "//div[@class='b-serp-item__content']//div[@class='b-serp-item__text']"
+
 images_request_block = '{"blocks":[{"block":"extra-content","params":{},"version":2},{"block":"i-global__params:ajax","params":{},"version":2},{"block":"search2:ajax","params":{},"version":2},{"block":"preview__isWallpaper","params":{},"version":2},{"block":"content_type_search","params":{},"version":2},{"block":"serp-controller","params":{},"version":2},{"block":"cookies_ajax","params":{},"version":2},{"block":"advanced-search-block","params":{},"version":2}],"metadata":{"bundles":{"lb":"AS?(E<X120"},"assets":{"las":"justifier-height=1;justifier-setheight=1;fitimages-height=1;justifier-fitincuts=1;react-with-dom=1;"},"extraContent":{"names":["i-react-ajax-adapter"]}}}'
 
 videos_request_block = '{"blocks":[{"block":"extra-content","params":{},"version":2},{"block":"i-global__params:ajax","params":{},"version":2},{"block":"search2:ajax","params":{},"version":2},{"block":"vital-incut","params":{},"version":2},{"block":"content_type_search","params":{},"version":2},{"block":"serp-controller","params":{},"version":2},{"block":"cookies_ajax","params":{},"version":2}],"metadata":{"bundles":{"lb":"^G]!q<X120"},"assets":{"las":"react-with-dom=1;185.0=1;73.0=1;145.0=1;5a502a.0=1;32c342.0=1;b84ac8.0=1"},"extraContent":{"names":["i-react-ajax-adapter"]}}}'
@@ -103,27 +109,29 @@ def get_youtube_iframe_src(url):
         if video_id:
             return 'https://www.youtube-nocookie.com/embed/' + video_id[0]
 
-
 def response(resp):
     if yandex_category == 'web':
         if (resp.url).path.startswith('/showcaptcha'):
             raise SearxEngineCaptchaException()
 
-        html_data = html.fromstring(resp.text)
+        dom = html.fromstring(resp.text)
-        text = html.tostring(html_data, encoding='unicode')
+        results_dom = dom.xpath('//li[contains(@class, "serp-item")]')
 
-        urls = re.findall(r'title-link" href="(.*?)" target="_blank"', text)
+        results = []
-        titles = re.findall(r'tabindex="4"><span>(.*?)</span></a></h3>', text)
+        for result_dom in results_dom:
-        contents = re.findall(r'"b-serp-item__text">(.*?)</div>', text)
+            urls = result_dom.xpath(url_xpath)
+            titles = result_dom.xpath(title_xpath)
+            contents = result_dom.xpath(content_xpath)
 
-        results = [
+            title_texts = [title.xpath("normalize-space(.)") for title in titles]
-            {
+            content_texts = [content.xpath("normalize-space(.)") for content in contents]
-                "url": url,
+
-                "title": title,
+            for url, title_text, content_text in zip(urls, title_texts, content_texts):
-                "content": content,
+                results.append({
-            }
+                    "url": url,
-            for url, title, content in zip(urls, titles, contents)
+                    "title": title_text,
-        ]
+                    "content": content_text,
+                })
 
         return results
 
@@ -132,21 +140,49 @@ def response(resp):
             raise SearxEngineCaptchaException()
 
         html_data = html.fromstring(resp.text)
-        text = unescape(html.tostring(html_data, encoding='unicode'))
+        html_sample = unescape(html.tostring(html_data, encoding='unicode'))
 
-        urls = re.findall(r'"img_href":"(.*?)"', text)
+        start_tag = 'data-state="'
-        titles = re.findall(r'"alt":"(.*?)"', text)
+        end_tag = '"advRsyaSearchColumn":null}}"'
 
-        results = [
+        start_pos = html_sample.find(start_tag)
-            {
+        start_pos += len(start_tag)
+
+        end_pos = html_sample.find(end_tag, start_pos)
+        end_pos += len(end_tag) - 1
+
+        content_between_tags = html_sample[start_pos:end_pos]
+
+        json_resp = utils.js_variable_to_python(content_between_tags)
+
+        # save to a file
+        #with open('/path/to/yandexdump.txt', 'w') as f:
+        #sys.stdout = f
+        #print(json_resp)
+
+        results = []
+        for item_id, item_data in json_resp['initialState']['serpList']['items']['entities'].items():
+            title = item_data['snippet']['title']
+            source = item_data['snippet']['url']
+            thumb = item_data['image']
+#            fullsize_image = item_data['origUrl']
+            fullsize_image = item_data['viewerData']['dups'][0]['url']
+#            height = item_data['height']
+#            width = item_data['width']
+            height = item_data['viewerData']['dups'][0]['h']
+            width = item_data['viewerData']['dups'][0]['w']
+            filesize = item_data['viewerData']['dups'][0]['fileSizeInBytes']
+            humanized_filesize = humanize_bytes(filesize)
+
+            results.append({
                 "title": title,
-                "url": url,
+                "url": source,
-                "img_src": url,
+                "img_src": fullsize_image,
-                "thumbnail_src": url,
+                "filesize": humanized_filesize,
+                "thumbnail_src": thumb,
                 "template": "images.html",
-            }
+                "resolution": f'{width} x {height}'
-            for url, title in zip(urls, titles)
+            })
-        ]
 
         return results