performance improvements on get_peers script

xmrnodes/cli.py

@@ -88,78 +88,75 @@ def check_node(_node):
             _hc.delete_instance()
         node.delete_instance()
 
+def upsert_peer(peer):
+    exists = Peer.select().where(Peer.url == peer).first()
+    if exists:
+        exists.datetime = datetime.utcnow()
+        exists.save()
+    else:
+        with geoip2.database.Reader("./data/GeoLite2-City.mmdb") as geodb:
+            try:
+                u = urlparse(peer)
+                _url = f"{u.scheme}://{u.netloc}".lower()
+                geodata = geodb.city(u.hostname)
+                p = Peer(
+                    url=_url,
+                    country=geodata.country.name,
+                    city=geodata.city.name,
+                    postal=geodata.postal.code,
+                    lat=geodata.location.latitude,
+                    lon=geodata.location.longitude,
+                )
+                p.save()
+            except Exception as e:
+                pass
+
+def _get_peers():
+    """
+    This command keeps will go through the oldest nodes and scan them for more peers.
+    Unresponsive peers get deleted. Responsive peers get their datestamp refreshed to move
+    to the top of the list. It will only crawl a subset of peers and is intended to be
+    run in intervals. The script will automatically prune out peers over time.
+    """
+    # crawl existing peers
+    peers = Peer.select().order_by(Peer.datetime.asc()).limit(20)
+    for peer in peers:
+        try:
+            new_peers = retrieve_peers(peer.hostname, peer.port)
+            if new_peers:
+                new = []
+                for new_peer in new_peers:
+                    exists = Peer.select().where(Peer.url == new_peer).first()
+                    if not exists:
+                        new.append(new_peer)
+                print(f"+++ Found {len(new)} more peers from {peer.url}")
+                upsert_peer(peer.url)
+                for new_peer in new_peers:
+                    upsert_peer(new_peer)
+            else:
+                raise Exception('dead node')
+        except Exception as e:
+            print(f"--- Dead peer {peer.url}")
+            peer.delete_instance()
+
+    # if no peers are available in the database then get a list of peers to scan from upstream node
+    if not peers:
+        print(f"[.] Retrieving peers from {config.NODE_HOST}:{config.NODE_PORT}")
+        peers_to_scan = retrieve_peers(config.NODE_HOST, config.NODE_PORT)
+        print(f"[+] Found {len(peers_to_scan)} initial peers to begin scraping.")
+        for peer in peers_to_scan:
+            upsert_peer(peer)
+    
+    # rw_cache("map_peers", list(Peer.select().execute()))
 
 @bp.cli.command("get_peers")
 def get_peers():
-    """
-    This command requests peers from the configured upstream node and fans out 
-    to recursively scrape all other peers on the network. This will take
-    several hours to run.
-    """
-    # keep track of all peers
-    all_peers = []
-    print("[+] Preparing to crawl Monero p2p network")
-    print(f"[.] Retrieving initial peers from {config.NODE_HOST}:{config.NODE_PORT}")
-
-    # start initial list of peers to scan
-    peers_to_scan = retrieve_peers(config.NODE_HOST, config.NODE_PORT)
-    print(f"[+] Found {len(peers_to_scan)} initial peers to begin scraping.")
-    sleep(3)
-
-    # helper function to add a new peer to the db or update an existing one
-    def save_peer(peer):
-        with geoip2.database.Reader("./data/GeoLite2-City.mmdb") as reader:
-            _url = urlparse(peer)
-            url = f"{_url.scheme}://{_url.netloc}".lower()
-            # add new peer if not in db
-            if not Peer.select().where(Peer.url == peer).exists():
-                response = reader.city(_url.hostname)
-                p = Peer(
-                    url=peer,
-                    country=response.country.name,
-                    city=response.city.name,
-                    postal=response.postal.code,
-                    lat=response.location.latitude,
-                    lon=response.location.longitude,
-                )
-                p.save()
-                print(f"{peer} - saving new peer")
-            # or update if it does
-            else:
-                p = Peer.select().where(Peer.url == peer).first()
-                p.datetime = datetime.now()
-                p.save()
-            return _url
-    
-    # iterate over the whole list until all peers have been scanned
-    # add new peers to the list
-    # skip the peer if we've seen it already
     try:
-        while peers_to_scan:
-            _peer = peers_to_scan[0]
-            peers_to_scan.pop(0)
-            if _peer in all_peers:
-                print(f'already found {_peer}')
-                continue
-            all_peers.append(_peer)
-            try:
-                peer = save_peer(_peer)
-                peers_to_scan += retrieve_peers(peer.hostname, peer.port)
-            except:
-                pass
+        _get_peers()
     except KeyboardInterrupt:
-        print('Stopped.')
-
-    print(
-        f"[+] Found {len(all_peers)} peers from {config.NODE_HOST}:{config.NODE_PORT}"
-    )
-    print("[+] Deleting old Monero p2p peers")
-    for p in Peer.select():
-        if p.hours_elapsed() > config.PEER_LIFETIME:
-            print(f"[.] Deleting {p.url}")
-            p.delete_instance()
-    rw_cache("map_peers", list(Peer.select().execute()))
-
+        print("Stopped")
+    except Exception as e:
+        print(f"Error: {e}")
 
 @bp.cli.command("validate")
 def validate():