works

mongo/mongocrawler.py

@@ -172,7 +172,6 @@ def fetch_page(link:str)->(str,str):
     html = None
     if response is not None :
         good = True
-        print(response)
         if response.status != 200:
             good = False
             LOGGER.error('not a 200 response: %s for URL %s', response.status, url)
@@ -212,8 +211,16 @@ def extract_page(final_link,html):
     if html is not None:
         doc = trafilatura.bare_extraction(html,url=final_link,with_metadata=True,include_formatting=False,target_language=LANGUAGE,favor_precision=True)
         if doc is not None:
-            if not "text" in doc or len(doc["text"]) < MIN_TEXT_SIZE:
+            lines = doc["text"].split("\n")
+            # filter out tables
+            good_lines = []
+            for line in lines:
+                if line.startswith("|") or line.startswith("1 2 3 4") or line.startswith("12345"):
+                    continue
+                good_lines.append(line)
+            doc["text"] = "\n".join(good_lines)
             # text too small
+            if not "text" in doc or len(doc["text"]) < MIN_TEXT_SIZE:
                 doc = None
     return doc
 
@@ -642,7 +649,8 @@ def parseurl(link):
     print(rules.crawl_delay("*"))
     html = trafilatura.fetch_url(link,decode=True)
     get_bs_links(link,html)
-    doc = trafilatura.bare_extraction(html)
+    doc = extract_page(link,html)
+    if doc is not None:
         import pprint
         pprint.pprint(doc)
         internal_links, external_links = get_bs_links(link,html)
@@ -719,6 +727,21 @@ def visit(hostname,filter_content=True):
 def crawl_summary():
     myclient = pymongo.MongoClient(CONNECTION)
     db=myclient[DBNAME]
+    contentcol = db["content"]
+    res = contentcol.aggregate([
+        {"$group":{"_id":None,"total_text_size":{"$sum":"$text_size"}}}
+    ])
+    print(">>>>> Total text size in content")
+    for item in res:
+        print(item)
+    linkscol = db["links"]
+    # find counts of link statuses
+    res = linkscol.aggregate([
+        {"$group":{"_id":"$status","count":{"$sum":1}}}
+    ])
+    print(">>>>> Link status counts")
+    for item in res:
+        print(item["_id"],item["count"])
     batchcol = db["batches"]
     yesterday =  datetime.datetime.today() - datetime.timedelta(days=1)
     print(yesterday)
@@ -732,6 +755,7 @@ def crawl_summary():
                    }
          },
         {"$sort":{"original_text_size":-1}},
+        {"$limit":100},
     ])
     print(">>>> Batches")
     headers = ["_id","document_count","good_document_count","batch_size","original_text_size"]
@@ -739,29 +763,41 @@ def crawl_summary():
     for item in res:
         values = [str(item[x]) for x in headers]
         print("\t".join(values))
-    contentcol = db["content"]
-    res = contentcol.aggregate([
-        {"$group":{"_id":None,total_text_size:{"$sum":"$text_size"}}}
-    ])
-    print(">>>>> Total text size in content")
-    for item in res:
-        print(res)
 
+def extr(hdoc):
+    url = hdoc["url"]
+    html = binascii.a2b_qp(hdoc["quoted_html"])
+    doc =  extract_page(url,html)
+    return doc
 
 def import_html():
     myclient= pymongo.MongoClient(CONNECTION)
     db=myclient[DBNAME]
-    for l in sys.stdin:
+    contentcol = db["content"]
+    buffer = []
+    counter = 0
+    for i,l in enumerate(sys.stdin):
         hdoc = json.loads(l)
         url = hdoc["url"]
-        # beautifusoup is to unify encoding
-        html = BeautifulSoup(binascii.a2b_qp(hdoc["quoted_html"])).prettify()
-        doc = extract_page(url,html)
-        if doc is not None:
-            print("------=====-")
-            print(doc)
-            status = index_page(db,url,url,html,doc)
-            print(status)
+        r = contentcol.find_one({"url":url},projection=["_id"])
+        if r is not None:
+            print(">>>>" +  str(i) + " copy: " + url)
+            continue
+        buffer.append(hdoc)
+        if len(buffer) < 128:
+            continue
+        from multiprocessing import Pool
+        with Pool(8) as p:
+            outs = p.map(extr,buffer)
+            for hdoc,doc in zip(buffer,outs):
+                if doc is None:
+                    print("bad html" + hdoc["url"])
+                    continue
+                status = index_page(db,hdoc["url"],hdoc["url"],hdoc["quoted_html"],doc)
+                counter += 1
+                print( ">>> " + str(counter) + " " + hdoc["url"] + " " + status)
+            del buffer[:]
+
 
 def sample_domains():
     myclient = pymongo.MongoClient(CONNECTION)