Merge branch 'block_relays' into 'master'

Block relays

See merge request pleroma/relay!26

relay.yaml.example

@@ -20,4 +20,11 @@ ap:
   whitelist_enabled: false
   whitelist:
   - 'good-instance.example.com'
-  - 'another.good-instance.example.com'
+  - 'another.good-instance.example.com'
+  # uncomment the lines below to prevent certain activitypub software from posting
+  # to the relay (all known relays by default). this uses the software name in nodeinfo
+  #blocked_software:
+  #- 'activityrelay'
+  #- 'aoderelay'
+  #- 'social.seattle.wa.us-relay'
+  #- 'unciarelay'

relay/__init__.py

@@ -9,9 +9,13 @@ import yaml
 
 def load_config():
     with open('relay.yaml') as f:
-        yaml_file = yaml.load(f)
-        whitelist = yaml_file['ap'].get('whitelist', [])
-        blocked = yaml_file['ap'].get('blocked_instances', [])
+        options = {}
+
+        ## Prevent a warning message for pyyaml 5.1+
+        if getattr(yaml, 'FullLoader', None):
+            options['Loader'] = yaml.FullLoader
+
+        yaml_file = yaml.load(f, **options)
 
         config = {
             'db': yaml_file.get('db', 'relay.jsonld'),
@@ -19,9 +23,10 @@ def load_config():
             'port': int(yaml_file.get('port', 8080)),
             'note': yaml_file.get('note', 'Make a note about your instance here.'),
             'ap': {
-                'blocked_instances': [] if blocked is None else blocked,
+                'blocked_software': [v.lower() for v in yaml_file['ap'].get('blocked_software', [])],
+                'blocked_instances': yaml_file['ap'].get('blocked_instances', []),
                 'host': yaml_file['ap'].get('host', 'localhost'),
-                'whitelist': [] if whitelist is None else whitelist,
+                'whitelist': yaml_file['ap'].get('whitelist', []),
                 'whitelist_enabled': yaml_file['ap'].get('whitelist_enabled', False)
             }
         }
@@ -30,7 +35,6 @@ def load_config():
 
 CONFIG = load_config()
 
-
 from .http_signatures import http_signatures_middleware
 
 

relay/actor.py

@@ -103,6 +103,29 @@ async def push_message_to_actor(actor, message, our_key_id):
             logging.info('Caught %r while pushing to %r.', e, inbox)
 
 
+async def fetch_nodeinfo(domain):
+    headers = {'Accept': 'application/json'}
+    nodeinfo_url = None
+
+    wk_nodeinfo = await fetch_actor(f'https://{domain}/.well-known/nodeinfo', headers=headers)
+
+    if not wk_nodeinfo:
+        return
+
+    for link in wk_nodeinfo.get('links', ''):
+        if link['rel'] == 'http://nodeinfo.diaspora.software/ns/schema/2.0':
+            nodeinfo_url = link['href']
+            break
+
+    if not nodeinfo_url:
+        return
+
+    nodeinfo_data = await fetch_actor(nodeinfo_url, headers=headers)
+    software = nodeinfo_data.get('software')
+
+    return software.get('name') if software else None
+
+
 async def follow_remote_actor(actor_uri):
     actor = await fetch_actor(actor_uri)
     
@@ -235,6 +258,7 @@ async def handle_follow(actor, data, request):
     following = DATABASE.get('relay-list', [])
     inbox = get_actor_inbox(actor)
 
+
     if urlsplit(inbox).hostname in AP_CONFIG['blocked_instances']:
         return
 
@@ -294,6 +318,12 @@ async def inbox(request):
     data = await request.json()
     instance = urlsplit(data['actor']).hostname
 
+    if AP_CONFIG['blocked_software']:
+        software = await fetch_nodeinfo(instance)
+
+        if software and software.lower() in AP_CONFIG['blocked_software']:
+            raise aiohttp.web.HTTPUnauthorized(body='relays have been blocked', content_type='text/plain')
+
     if 'actor' not in data or not request['validated']:
         raise aiohttp.web.HTTPUnauthorized(body='access denied', content_type='text/plain')
 

relay/remote_actor.py

@@ -12,13 +12,18 @@ CACHE_TTL = CONFIG.get('cache-ttl', 3600)
 ACTORS = TTLCache(CACHE_SIZE, CACHE_TTL)
 
 
-async def fetch_actor(uri, force=False):
+async def fetch_actor(uri, headers={}, force=False):
     if uri in ACTORS and not force:
         return ACTORS[uri]
 
+    new_headers = {'Accept': 'application/activity+json'}
+
+    for k,v in headers.items():
+        new_headers[k.capitalize()] = v
+
     try:
         async with aiohttp.ClientSession(trace_configs=[http_debug()]) as session:
-            async with session.get(uri, headers={'Accept': 'application/activity+json'}) as resp:
+            async with session.get(uri, headers=new_headers) as resp:
                 if resp.status != 200:
                     return None
                 ACTORS[uri] = (await resp.json(encoding='utf-8', content_type=None))

requirements.txt

@@ -6,7 +6,7 @@ idna==2.7
 idna-ssl==1.1.0
 multidict==4.3.1
 pycryptodome==3.9.4
-PyYAML==3.13
+PyYAML>=5.1
 simplejson==3.16.0
 yarl==1.2.6
 cachetools