work on patching support!

config.yaml

@@ -2,6 +2,8 @@
 # ========================================
 #
 
+dedup_index: redis://localhost:6379/0/c:{coll}:cdxj
+
 collections:
     all: $all
     pywb:

pywb/apps/frontendapp.py

@@ -10,6 +10,8 @@ from wsgiprox.wsgiprox import WSGIProxMiddleware
 
 from pywb.recorder.multifilewarcwriter import MultiFileWARCWriter
 from pywb.recorder.recorderapp import RecorderApp
+from pywb.recorder.redisindexer import WritableRedisIndexer
+from pywb.recorder.filters import SkipDupePolicy, WriteRevisitDupePolicy, WriteDupePolicy
 
 from pywb.utils.loaders import load_yaml_config
 from pywb.utils.geventserver import GeventServer
@@ -207,8 +209,25 @@ class FrontEndApp(object):
         else:
             recorder_coll = recorder_config['source_coll']
 
-        # TODO: support dedup
         dedup_index = None
+        if self.warcserver.dedup_index:
+            policy = self.warcserver.config.get('dedup_policy')
+            if policy == 'skip':
+                dedup_policy = SkipDupePolicy()
+            elif policy == 'revisit':
+                dedup_policy = WriteRevisitDupePolicy()
+            elif policy == 'keep':
+                dedup_policy = WriteDupePolicy()
+            else:
+                dedup_policy = WriteRevisitDupePolicy()
+
+            print('Recorder Dedup: {0} policy via dedup index {1}'.format(policy, self.warcserver.dedup_index))
+
+            dedup_index = WritableRedisIndexer(
+                            redis_url=self.warcserver.dedup_index,
+                            rel_path_template=self.warcserver.archive_paths,
+                            dupe_policy=dedup_policy)
+
         warc_writer = MultiFileWARCWriter(self.warcserver.archive_paths,
                                           max_size=int(recorder_config.get('rollover_size', 1000000000)),
                                           max_idle_secs=int(recorder_config.get('rollover_idle_secs', 600)),
@@ -393,8 +412,8 @@ class FrontEndApp(object):
         :return: WbResponse containing the contents of the record/URL
         :rtype: WbResponse
         """
-        if coll in self.warcserver.list_fixed_routes():
+        #if coll in self.warcserver.list_fixed_routes():
-            return WbResponse.text_response('Error: Can Not Record Into Custom Collection "{0}"'.format(coll))
+        #    return WbResponse.text_response('Error: Can Not Record Into Custom Collection "{0}"'.format(coll))
 
         return self.serve_content(environ, coll, url, record=True)
 
@@ -625,8 +644,8 @@ class FrontEndApp(object):
 
         if proxy_config.get('recording'):
             logging.info('Proxy recording into collection "{0}"'.format(proxy_coll))
-            if proxy_coll in self.warcserver.list_fixed_routes():
+            #if proxy_coll in self.warcserver.list_fixed_routes():
-                raise Exception('Can not record into fixed collection')
+            #raise Exception('Can not record into fixed collection')
 
             proxy_route = proxy_coll + self.RECORD_ROUTE
             if not config.get('recorder'):

pywb/indexer/cdxindexer.py

@@ -167,9 +167,11 @@ class SortedCDXWriter(BaseCDXWriter):
         super(SortedCDXWriter, self).write(entry, filename)
         line = self.out.getvalue()
         if line:
-            insort(self.sortlist, line)
+            self.sortlist.append(line)
+            #insort(self.sortlist, line)
 
     def __exit__(self, *args):
+        self.sortlist.sort()
         self.actual_out.write(''.join(self.sortlist))
         return False
 

pywb/recorder/recorderapp.py

@@ -195,6 +195,8 @@ class RecorderApp(object):
 
         req_is_wrapped = False
 
+        print('PARAMS', params)
+
         if not skipping:
             req_stream = ReqWrapper(input_buff,
                                     headers,

pywb/recorder/redisindexer.py

@@ -48,14 +48,18 @@ class WritableRedisIndexer(RedisIndexSource):
         return base_name
 
     def add_warc_file(self, full_filename, params):
+        print('PARAMS W', params)
         base_filename = self._get_rel_or_base_name(full_filename, params)
         file_key = res_template(self.file_key_template, params)
+        if not file_key:
+            return
 
         full_load_path = self.full_warc_prefix + full_filename
 
         self.redis.hset(file_key, base_filename, full_load_path)
 
     def add_urls_to_index(self, stream, params, filename, length):
+        print('PARAMS U', params)
         base_filename = self._get_rel_or_base_name(filename, params)
 
         cdxout = BytesIO()
@@ -64,6 +68,7 @@ class WritableRedisIndexer(RedisIndexSource):
                         writer_cls=params.get('writer_cls'))
 
         z_key = res_template(self.redis_key_template, params)
+        print('KEY', z_key, self.redis_key_template, params)
 
         cdx_list = cdxout.getvalue().rstrip().split(b'\n')
 

pywb/warcserver/warcserver.py

@@ -60,6 +60,7 @@ class WarcServer(BaseWarcServer):
 
         super(WarcServer, self).__init__(debug=config.get('debug', False))
         self.config = config
+        self.dedup_index = self.config.get('dedup_index')
 
         self.root_dir = self.config.get('collections_root', '')
         self.index_paths = self.init_paths('index_paths')
@@ -113,7 +114,14 @@ class WarcServer(BaseWarcServer):
         access_checker = AccessChecker(CacheDirectoryAccessSource(self.acl_paths),
                                        self.default_access)
 
-        return DefaultResourceHandler(dir_source, self.archive_paths,
+        if self.dedup_index:
+            source = SimpleAggregator({'dedup': RedisMultiKeyIndexSource(self.dedup_index),
+                                       'dir': dir_source})
+
+        else:
+            source = dir_source
+
+        return DefaultResourceHandler(source, self.archive_paths,
                                       rules_file=self.rules_file,
                                       access_checker=access_checker)
 
@@ -243,6 +251,7 @@ def init_index_source(value, source_list=None):
                 return source
 
     else:
+        print(value)
         raise Exception('Source config must be string or dict')
 
     raise Exception('No Index Source Found for: ' + str(value))