zz

mongo/mongocwarler.py

@@ -1,4 +1,5 @@
 import pymongo
+import pymongo.errors
 import trafilatura
 import trafilatura.feeds
 import trafilatura.sitemaps
@@ -71,7 +72,6 @@ def calculate_checksums(text):
 def is_robot_good(link,rules):
     # check robots.txt rules
     if rules is not None and not rules.can_fetch("*", link):
-        print("bad>>>" + link)
         return False
     return True
 
@@ -203,7 +203,7 @@ def index_pages(db,hostname,extracted_pages):
                 nd = checkcol.find_one({"_id":chs})
                 if nd is not None:
                     copysz += paragraph_size
-            if copysz / len(text) > TEXT_TRASH_RATIO:
+            if (copysz / len(text)) > TEXT_TRASH_RATIO:
                 state = "copy"
             print(copysz)
         if state == "good":
@@ -219,7 +219,10 @@ def index_pages(db,hostname,extracted_pages):
             del doc["url"]
             contentcol.update_one({"url":link},{"$set":doc},upsert=True)
             for chs in doc["paragraph_checksums"]:
+                try:
                     checkcol.insert_one({"_id":chs})
+                except pymongo.errors.DuplicateKeyError as err:
+                    pass
         linkcol.update_one({"url":original_link},{"$set":{"status":state}})
 
 
@@ -238,8 +241,6 @@ def extract_links(link_batch,responses,hostname,rules,default_status="frontlink"
                 badrobot += 1
                 continue
             status = str(default_status)
-            if courlan.is_navigation_page(link):
-                status = "navigation"
             #print(link,status)
             links[link] = status
     outlinks = []
@@ -264,19 +265,27 @@ def index_links(db,extracted_links):
 
 def get_links(db,hostname,status,batch_size):
     linkcol = db["links"]
-    res  = linkcol.find({"status":status,"host":hostname},{"url":1},limit=batch_size)
-    links = []
+    #res  = linkcol.find({"status":status,"host":hostname},{"url":1},limit=batch_size)
+    res = linkcol.aggregate([
+        { "$match": { "status": status,"host":hostname } },
+        { "$sample": { "size": batch_size } }
+    ])
+    links = set()
     for i,doc in enumerate(res):
         #print(">>>>>" + status)
         #print(doc);
-        print(">>>>links")
-        print(doc)
-        links.append(doc["url"])
+        links.add(doc["url"])
         if i >= batch_size:
             break
-    return links
+    return list(links)
 
 
+def fetch_sitemap_links(start_link):
+    out = []
+    navigation_links = trafilatura.sitemaps.sitemap_search(start_link,target_lang=LANGUAGE)
+    for link in navigation_links:
+        out.append((link,"frontlink"))
+    return out
 
 def process_links(db,hostname,status,links=[],rules=None,batch_size=BATCHSIZE):
     #print(links)
@@ -372,23 +381,24 @@ def visit(start_link):
     myclient = pymongo.MongoClient(CONNECTION)
     db=myclient[DBNAME]
     start_link,hostname = courlan.check_url(start_link)
-    rules = fetch_robot(hostname)
-    print(rules)
     batch_size = BATCHSIZE
-    navigation_links = get_links(db,hostname,"navigation",batch_size)
-    if start_link is not None:
-        navigation_links.append(start_link)
-    print(f"Navigation links {len(navigation_links)}")
-    process_links(db,hostname,"frontlink",navigation_links,rules)
+
+    print("Getting  frontlinks")
     links = get_links(db,hostname,"frontlink",batch_size)
-    bl = len(links) - batch_size
     print(f"Got {len(links)} frontlinks")
-    if bl > 0:
+    if len(links) < batch_size:
+        print("Fetching sitemap links")
+        sitemap_links = fetch_sitemap_links(start_link)
+        index_links(db,sitemap_links)
+    links.append(start_link)
+
+    print("Processing frontlinks")
+    rules = fetch_robot(hostname)
+    process_links(db,hostname,"frontlink",links,rules)
     print("Getting  backlinks")
-        front_links = get_links(db,hostname,"backlink",bl)
-        links += front_links
+    back_links = get_links(db,hostname,"backlink",batch_size)
     print("Processing backlinks")
-    process_links(db,hostname,"backlink",links,rules=rules)
+    process_links(db,hostname,"backlink",back_links,rules=rules)
     link_summary(db,hostname)
 
 if __name__ == "__main__":