From a07c95003fcacaaf2674e3a9cdb243f88fa01f12 Mon Sep 17 00:00:00 2001
From: Daniel Grana <dangra@gmail.com>
Date: Fri, 23 Jan 2009 03:42:05 +0000
Subject: [PATCH] duplicatesfilter: first version of configurable duplicate
 requests filtering middleware

--HG--
extra : convert_revision : svn%3Ab85faa78-f9eb-468e-a121-7cced6da292c%40763
---
 .../conf/project_template/scrapy_settings.py  |  3 +
 .../spidermiddleware/duplicatesfilter.py      | 65 +++++++++++++++++++
 2 files changed, 68 insertions(+)
 create mode 100644 scrapy/trunk/scrapy/contrib/spidermiddleware/duplicatesfilter.py

diff --git a/scrapy/trunk/scrapy/conf/project_template/scrapy_settings.py b/scrapy/trunk/scrapy/conf/project_template/scrapy_settings.py
index f885d9e42..c0e60df36 100644
--- a/scrapy/trunk/scrapy/conf/project_template/scrapy_settings.py
+++ b/scrapy/trunk/scrapy/conf/project_template/scrapy_settings.py
@@ -69,6 +69,7 @@ SPIDER_MIDDLEWARES = (
     'scrapy.contrib.spidermiddleware.urllength.UrlLengthMiddleware',
     'scrapy.contrib.spidermiddleware.depth.DepthMiddleware',
     'scrapy.contrib.spidermiddleware.urlfilter.UrlFilterMiddleware',
+    'scrapy.contrib.spidermiddleware.duplicatesfilter.DuplicatesFilterMiddleware',
     # Spider side
 )
 
@@ -96,3 +97,5 @@ URLLENGTH_LIMIT = 2083
 WS_ENABLED = 0
 
 SPIDERPROFILER_ENABLED = 0
+
+#DUPLICATESFILTER_FILTERCLASS = 'scrapy.contrib.spidermiddleware.duplicatesfilter.SimplePerDomainFilter'
diff --git a/scrapy/trunk/scrapy/contrib/spidermiddleware/duplicatesfilter.py b/scrapy/trunk/scrapy/contrib/spidermiddleware/duplicatesfilter.py
new file mode 100644
index 000000000..a421a1472
--- /dev/null
+++ b/scrapy/trunk/scrapy/contrib/spidermiddleware/duplicatesfilter.py
@@ -0,0 +1,65 @@
+"""
+DuplicatesFilterMiddleware: Filter out already visited urls
+"""
+
+from pydispatch import dispatcher
+
+from scrapy.core import signals
+from scrapy.http import Request
+from scrapy.core.exceptions import NotConfigured
+from scrapy.utils.request import request_fingerprint
+from scrapy.utils.misc import load_class
+from scrapy.conf import settings
+from scrapy import log
+
+
+class DuplicatesFilterMiddleware(object):
+    """Filter out duplicate requests to avoid visiting same page more than once"""
+
+    def __init__(self):
+        clspath = settings.get('DUPLICATESFILTER_FILTERCLASS')
+        if not clspath:
+            raise NotConfigured
+
+        self.filter = load_class(clspath)()
+        dispatcher.connect(self.filter.open, signals.domain_open)
+        dispatcher.connect(self.filter.close, signals.domain_closed)
+
+    def process_spider_output(self, response, result, spider):
+        domain = spider.domain_name
+
+        for req in result:
+            if isinstance(req, Request):
+                added = self.filter.add(domain, req)
+                if not (added or req.dont_filter):
+                    log.msg('Skipped (already visited): %s' % req, log.DEBUG, domain=domain)
+                    continue
+            yield req
+
+
+class SimplePerDomainFilter(dict):
+    """Filter out a request if already seen for same domain"""
+
+    def open(self, domain):
+        """Initialize the resources needed for filtering for this domain"""
+        self[domain] = set()
+
+    def close(self, domain):
+        """Remove the resources reserved for filtering for this domain"""
+        del self[domain]
+
+    def add(self, domain, request):
+        """Add the fingerprint of a request to the domain set if a equivalent fingerprint has not been added.
+        This method will return true if the fingerprint was added and false otherwise.
+        """
+        fp = request_fingerprint(request)
+        if fp not in self[domain]:
+            self[domain].add(fp)
+            return True
+        return False
+
+    def has(self, domain, request):
+        """Check if a request was already seen for a domain"""
+        return request_fingerprint(request) in self[domain]
+
+