fix imdb parser

oxdata/movie/imdbids.py

@@ -21,6 +21,13 @@ DAY = 24 * 60 * 60
 TIMEOUT = 90 * DAY
 DATA_ROOT = os.path.join(settings.MEDIA_ROOT, 'imdb')
 
+DEBUG = False
+
+def debug(*args, **kwargs):
+    if DEBUG:
+        print(*args, **kwargs)
+
+
 def read_url(url, timeout):
     data = ox.web.imdb.read_url(url, unicode=True, timeout=timeout)
     while '>500 Error - IMDb<' in data:
@@ -85,11 +92,9 @@ def update_year(year, film_counts):
             if film_count != film_counts.get(key):
                 print_info(key, film_count, film_counts)
                 film_counts[key] = film_count
-                update_month(year, month, film_counts)
+                update_month(year, month, film_counts, expected=film_count)
     else:
         r = update_ids(year)
-        if r != film_counts[key]:
-            print('%s: count %s, got ids %s' % (key, film_counts[key], r))
         save_film_counts(film_counts)
 
 def update_month(year, month, film_counts):
@@ -106,23 +111,17 @@ def update_month(year, month, film_counts):
                 print_info(key, film_count, film_counts)
                 film_counts[key] = film_count
                 if film_count > MAX_PER_RANGE and film_count < 2*MAX_PER_RANGE:
-                    r = update_ids(year, month, day, sort='alpha')
+                    r = update_ids(year, month, day, sort='alpha', expected=film_count)
-                    if r != film_counts[key]:
-                        print('%s: count %s, got ids %s' % (key, film_counts[key], r))
                     save_film_counts(film_counts)
                 elif film_count > MAX_PER_RANGE:
-                    print(key, '!!!to many per day')
+                    print(key, '!!!to many per day', film_count, key)
                 else:
-                    r = update_ids(year, month, day)
+                    r = update_ids(year, month, day, expected=film_count)
-                    if r != film_counts[key]:
-                        print('%s: count %s, got ids %s' % (key, film_counts[key], r))
                     save_film_counts(film_counts)
         if days_total != month_total:
             print('!! month and days don\'t add up: %s month vs %s days total' % (month_total, days_total))
     else:
-        r = update_ids(year, month)
+        r = update_ids(year, month, expected=film_count)
-        if r != film_counts[key]:
-            print('%s: count %s, got ids %s' % (key, film_counts[key], r))
         save_film_counts(film_counts)
 
 def parse_cast(string):
@@ -139,7 +138,7 @@ def parse_cast(string):
             results[rkey] = cast[key].split(', \n')
     return results
 
-def update_ids(year, month=None, day=None, sort=None):
+def update_ids(year, month=None, day=None, sort=None, expexted=None):
     films = {}
     if day is not None:
         url = get_day(year, month, day)
@@ -157,16 +156,23 @@ def update_ids(year, month=None, day=None, sort=None):
         ]
     else:
         urls = [url]
+
+    if not expexted:
+        expexted = get_film_count(year, month, day)
+
     for url in urls:
         data = read_url(url, timeout=TIMEOUT)
         n = True
         step = 50
         start = 1
         while n:
-            n = re.compile('Next &#187;</a>', re.DOTALL).findall(data)
+            start += step
-            if n:
+            next_link = 'start=%s&ref_=adv_nxt"' % (start)
+            has_next = re.compile(next_link).findall(data)
+            if has_next:
                 n = '%s&start=%s' % (url, start)
-                start += step
+            else:
+                n = False
             doc = lxml.html.fromstring(data)
             article = doc.find_class('article')
             if article:
@@ -228,21 +234,26 @@ def update_ids(year, month=None, day=None, sort=None):
                         films[id]['genre'] = genre
                     if cast:
                         films[id].update(cast)
-
+            if expected and len(films) == expected and sort == 'alpha':
-            #print(key, len(films), 'films')
+                n = False
+            debug('%s: %s of %s films - next: %s' % (key, len(films), expected, n))
             if n:
-                #print(n)
                 data = read_url(n, timeout=TIMEOUT)
     path = get_path('ids/%s.json' % key)
     with open(path, 'w') as fd:
         json.dump(films, fd, indent=4, ensure_ascii=False, sort_keys=True)
-    return len(films)
+    r = len(films)
+    if r != expected:
+        print('%s: got %s, expected %s' % (key, r, expexted))
+    return r
 
 def save_film_counts(film_counts):
     with open(get_path('film_counts.json'), 'w') as fd:
         json.dump(film_counts, fd, indent=4, sort_keys=True)
 
-def update_index():
+def update_index(from_year=None):
+    if from_year is None:
+        from_year = 1874
     film_counts_json = get_path('film_counts.json')
     if os.path.exists(film_counts_json):
         with open(film_counts_json) as fd:
@@ -250,7 +261,7 @@ def update_index():
     else:
         film_counts = {}
 
-    for year in range(1894, datetime.now().year+1):
+    for year in range(from_year, datetime.now().year+1):
         film_count = get_film_count(year)
         key = '%s' % year
         if film_count != film_counts.get(key):

oxdata/movie/models.py

@@ -255,13 +255,13 @@ class Imdb(models.Model):
             j['year'] = int(j['year'])
         return j
 
-def get_new_ids(timeout=-1):
+def get_new_ids(timeout=-1, from_year=None):
     new_ids_cache = '/tmp/missing.json'
     if os.path.exists(new_ids_cache):
         with open(new_ids_cache) as fd:
             new_ids = set(json.load(fd))
     else:
-        update_index()
+        update_index(from_year)
         known_ids = frozenset([i['imdb'] for i in Imdb.objects.all().values('imdb')])
         new_ids = get_unknown_ids(known_ids)
     if new_ids: