Fix extracting cached URLs from google search results

src/calibre/ebooks/metadata/sources/search_engines.py

@@ -9,10 +9,10 @@ import re
 import time
 from collections import defaultdict, namedtuple
 try:
-    from urllib.parse import parse_qs, quote_plus, urlencode
+    from urllib.parse import parse_qs, quote_plus, urlencode, unquote
 except ImportError:
     from urlparse import parse_qs
-    from urllib import quote_plus, urlencode
+    from urllib import quote_plus, urlencode, unquote
 
 from lxml import etree
 
@@ -20,7 +20,7 @@ from calibre import browser as _browser, prints, random_user_agent
 from calibre.utils.monotonic import monotonic
 from calibre.utils.random_ua import accept_header_for_ua
 
-current_version = (1, 0, 7)
+current_version = (1, 0, 8)
 minimum_calibre_version = (2, 80, 0)
 
 
@@ -60,7 +60,7 @@ def parse_html(raw):
         return parse(raw)
 
 
-def query(br, url, key, dump_raw=None, limit=1, parser=parse_html, timeout=60):
+def query(br, url, key, dump_raw=None, limit=1, parser=parse_html, timeout=60, save_raw=None):
     delta = monotonic() - last_visited[key]
     if delta < limit and delta > 0:
         time.sleep(delta)
@@ -71,6 +71,8 @@ def query(br, url, key, dump_raw=None, limit=1, parser=parse_html, timeout=60):
     if dump_raw is not None:
         with open(dump_raw, 'wb') as f:
             f.write(raw)
+    if save_raw is not None:
+        save_raw(raw)
     return parser(raw)
 
 
@@ -221,15 +223,34 @@ def google_url_processor(url):
     return url
 
 
-def google_search(terms, site=None, br=None, log=prints, safe_search=False, dump_raw=None, timeout=60):
-    terms = [quote_term(google_term(t)) for t in terms]
-    if site is not None:
-        terms.append(quote_term(('site:' + site)))
-    q = '+'.join(terms)
-    url = 'https://www.google.com/search?q={q}'.format(q=q)
-    log('Making google query: ' + url)
-    br = br or browser()
-    root = query(br, url, 'google', dump_raw, timeout=timeout)
+def google_extract_cache_urls(raw):
+    if isinstance(raw, bytes):
+        raw = raw.decode('utf-8', 'replace')
+    pat = re.compile(r'\\x22(https://webcache\.googleusercontent\.com/.+?)\\x22')
+    upat = re.compile(r'\\\\u([0-9a-fA-F]{4})')
+    cache_pat = re.compile('cache:([^:]+):(.+)')
+
+    def urepl(m):
+        return chr(int(m.group(1), 16))
+
+    seen = set()
+    ans = {}
+    for m in pat.finditer(raw):
+        cache_url = upat.sub(urepl, m.group(1))
+        m = cache_pat.search(cache_url)
+        cache_id, src_url = m.group(1), m.group(2)
+        if cache_id in seen:
+            continue
+        seen.add(cache_id)
+        src_url = src_url.split('+')[0]
+        src_url = unquote(src_url)
+        ans[src_url] = cache_url
+    return ans
+
+
+def google_parse_results(root, raw, log=prints):
+    cache_url_map = google_extract_cache_urls(raw)
+    # print('\n'.join(cache_url_map))
     ans = []
     for div in root.xpath('//*[@id="search"]//*[@id="rso"]//*[@class="g"]'):
         try:
@@ -238,22 +259,45 @@ def google_search(terms, site=None, br=None, log=prints, safe_search=False, dump
             log('Ignoring div with no main result link')
             continue
         title = tostring(a)
-        try:
-            c = div.xpath('descendant::*[@role="menuitem"]//a[@class="fl"]')[0]
-        except IndexError:
-            log('Ignoring {!r} as it has no cached page'.format(title))
-            continue
-        cached_url = c.get('href')
+        src_url = a.get('href')
+        if src_url in cache_url_map:
+            cached_url = cache_url_map[src_url]
+        else:
+            try:
+                c = div.xpath('descendant::*[@role="menuitem"]//a[@class="fl"]')[0]
+            except IndexError:
+                log('Ignoring {!r} as it has no cached page'.format(title))
+                continue
+            cached_url = c.get('href')
         ans.append(Result(a.get('href'), title, cached_url))
     if not ans:
         title = ' '.join(root.xpath('//title/text()'))
         log('Failed to find any results on results page, with title:', title)
-    return ans, url
+    return ans
 
 
-def google_develop(search_terms='1423146786'):
-    br = browser()
-    for result in google_search(search_terms.split(), 'www.amazon.com', dump_raw='/t/raw.html', br=br)[0]:
+def google_search(terms, site=None, br=None, log=prints, safe_search=False, dump_raw=None, timeout=60):
+    terms = [quote_term(google_term(t)) for t in terms]
+    if site is not None:
+        terms.append(quote_term(('site:' + site)))
+    q = '+'.join(terms)
+    url = 'https://www.google.com/search?q={q}'.format(q=q)
+    log('Making google query: ' + url)
+    br = br or browser()
+    r = []
+    root = query(br, url, 'google', dump_raw, timeout=timeout, save_raw=r.append)
+    return google_parse_results(root, r[0], log=log), url
+
+
+def google_develop(search_terms='1423146786', raw_from=''):
+    if raw_from:
+        with open(raw_from, 'rb') as f:
+            raw = f.read()
+        results = google_parse_results(parse_html(raw), raw)
+    else:
+        br = browser()
+        results = google_search(search_terms.split(), 'www.amazon.com', dump_raw='/t/raw.html', br=br)[0]
+    for result in results:
         if '/dp/' in result.url:
             print(result.title)
             print(' ', result.url)