[fix] engine deviantart: review of the result-scrapper

The deviantart site changed and hence deviantart is currently unusable.

searx/engines/deviantart.py

@@ -1,12 +1,14 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
 # lint: pylint
-"""
- Deviantart (Images)
+"""Deviantart (Images)
+
 """
 
-from urllib.parse import urlencode
+import urllib.parse
 from lxml import html
 
+from searx.utils import extract_text, eval_xpath, eval_xpath_list
+
 # about
 about = {
     "website": 'https://www.deviantart.com/',
@@ -20,31 +22,30 @@ about = {
 # engine dependent config
 categories = ['images']
 paging = True
-time_range_support = True
-
-time_range_dict = {
-    'day': 'popular-24-hours',
-    'week': 'popular-1-week',
-    'month': 'popular-1-month',
-    'year': 'most-recent',
-}
 
 # search-url
 base_url = 'https://www.deviantart.com'
 
+results_xpath = '//div[@class="_2pZkk"]/div/div/a'
+url_xpath = './@href'
+thumbnail_src_xpath = './div/img/@src'
+img_src_xpath = './div/img/@srcset'
+title_xpath = './@aria-label'
+premium_xpath = '../div/div/div/text()'
+premium_keytext = 'Watch the artist to view this deviation'
+cursor_xpath = '(//a[@class="_1OGeq"]/@href)[last()]'
+
 
 def request(query, params):
 
-    # https://www.deviantart.com/search/deviations?page=5&q=foo
+    # https://www.deviantart.com/search?q=foo
 
-    query = {
-        'page': params['pageno'],
-        'q': query,
-    }
-    if params['time_range'] in time_range_dict:
-        query['order'] = time_range_dict[params['time_range']]
-
-    params['url'] = base_url + '/search/deviations?' + urlencode(query)
+    nextpage_url = params['engine_data'].get('nextpage')
+    # don't use nextpage when user selected to jump back to page 1
+    if params['pageno'] > 1 and nextpage_url is not None:
+        params['url'] = nextpage_url
+    else:
+        params['url'] = f"{base_url}/search?{urllib.parse.urlencode({'q': query})}"
 
     return params
 
@@ -52,30 +53,36 @@ def request(query, params):
 def response(resp):
 
     results = []
-
     dom = html.fromstring(resp.text)
 
-    for row in dom.xpath('//div[contains(@data-hook, "content_row")]'):
-        for result in row.xpath('./div'):
+    for result in eval_xpath_list(dom, results_xpath):
+        # skip images that are blurred
+        _text = extract_text(eval_xpath(result, premium_xpath))
+        if _text and premium_keytext in _text:
+            continue
+        img_src = extract_text(eval_xpath(result, img_src_xpath))
+        if img_src:
+            img_src = img_src.split(' ')[0]
+            parsed_url = urllib.parse.urlparse(img_src)
+            img_src = parsed_url._replace(path=parsed_url.path.split('/v1')[0]).geturl()
 
-            a_tag = result.xpath('.//a[@data-hook="deviation_link"]')[0]
-            noscript_tag = a_tag.xpath('.//noscript')
+        results.append(
+            {
+                'template': 'images.html',
+                'url': extract_text(eval_xpath(result, url_xpath)),
+                'img_src': img_src,
+                'thumbnail_src': extract_text(eval_xpath(result, thumbnail_src_xpath)),
+                'title': extract_text(eval_xpath(result, title_xpath)),
+            }
+        )
 
-            if noscript_tag:
-                img_tag = noscript_tag[0].xpath('.//img')
-            else:
-                img_tag = a_tag.xpath('.//img')
-            if not img_tag:
-                continue
-            img_tag = img_tag[0]
-
-            results.append(
-                {
-                    'template': 'images.html',
-                    'url': a_tag.attrib.get('href'),
-                    'img_src': img_tag.attrib.get('src'),
-                    'title': img_tag.attrib.get('alt'),
-                }
-            )
+    nextpage_url = extract_text(eval_xpath(dom, cursor_xpath))
+    if nextpage_url:
+        results.append(
+            {
+                'engine_data': nextpage_url.replace("http://", "https://"),
+                'key': 'nextpage',
+            }
+        )
 
     return results