RNS-Things/reticulum-meshchatX
1
0
Fork 0
Code Issues 1 Pull Requests Actions Packages Projects Releases 8 Wiki Activity

Implement spam and blocking features in database and API

Browse Source 
- Incremented database version to 6 and added 'is_spam' column to LxmfMessage model.
- Introduced BlockedDestination and SpamKeyword models for managing blocked destinations and spam keywords.
- Added API endpoints for managing blocked destinations and spam keywords, including create, read, and delete operations.
- Enhanced message handling to mark messages as spam based on keywords and block messages from blocked sources.
This commit is contained in:
ivan
2025-11-30 20:41:48 -06:00
parent 5fcc86d65a
commit a3cb84fa06
2 changed files with 231 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

203
meshchat.py
View File

@@ -85,6 +85,8 @@ class ReticulumMeshChat:
database.LxmfMessage,
database.LxmfConversationReadState,
database.LxmfUserIcon,
database.BlockedDestination,
database.SpamKeyword,
])
# init config
@@ -160,7 +162,7 @@ class ReticulumMeshChat:
self.download_id_counter = 0
# register audio call identity
self.audio_call_manager = AudioCallManager(identity=self.identity)
self.audio_call_manager = AudioCallManager(identity=self.identity, is_destination_blocked_callback=self.is_destination_blocked)
self.audio_call_manager.register_incoming_call_callback(self.on_incoming_audio_call)
# start background thread for auto announce loop
@@ -2359,6 +2361,122 @@ class ReticulumMeshChat:
"message": "ok",
})
# get blocked destinations
@routes.get("/api/v1/blocked-destinations")
async def index(request):
blocked = database.BlockedDestination.select()
blocked_list = []
for b in blocked:
created_at = b.created_at
if isinstance(created_at, str):
created_at_str = created_at
else:
created_at_str = created_at.isoformat() if hasattr(created_at, 'isoformat') else str(created_at)
blocked_list.append({
"destination_hash": b.destination_hash,
"created_at": created_at_str
})
return web.json_response({
"blocked_destinations": blocked_list,
})
# add blocked destination
@routes.post("/api/v1/blocked-destinations")
async def index(request):
data = await request.json()
destination_hash = data.get("destination_hash", "")
if not destination_hash or len(destination_hash) != 32:
return web.json_response({"error": "Invalid destination hash"}, status=400)
try:
database.BlockedDestination.create(destination_hash=destination_hash)
# drop any existing paths to this destination
try:
RNS.Transport.drop_path(bytes.fromhex(destination_hash))
except:
pass
return web.json_response({"message": "ok"})
except:
return web.json_response({"error": "Destination already blocked"}, status=400)
# remove blocked destination
@routes.delete("/api/v1/blocked-destinations/{destination_hash}")
async def index(request):
destination_hash = request.match_info.get("destination_hash", "")
if not destination_hash or len(destination_hash) != 32:
return web.json_response({"error": "Invalid destination hash"}, status=400)
try:
blocked = database.BlockedDestination.get_or_none(database.BlockedDestination.destination_hash == destination_hash)
if blocked:
blocked.delete_instance()
return web.json_response({"message": "ok"})
else:
return web.json_response({"error": "Destination not blocked"}, status=404)
except Exception as e:
return web.json_response({"error": str(e)}, status=500)
# get spam keywords
@routes.get("/api/v1/spam-keywords")
async def index(request):
keywords = database.SpamKeyword.select()
keyword_list = [{"id": k.id, "keyword": k.keyword, "created_at": k.created_at.isoformat()} for k in keywords]
return web.json_response({
"spam_keywords": keyword_list,
})
# add spam keyword
@routes.post("/api/v1/spam-keywords")
async def index(request):
data = await request.json()
keyword = data.get("keyword", "").strip()
if not keyword:
return web.json_response({"error": "Keyword is required"}, status=400)
try:
database.SpamKeyword.create(keyword=keyword)
return web.json_response({"message": "ok"})
except:
return web.json_response({"error": "Keyword already exists"}, status=400)
# remove spam keyword
@routes.delete("/api/v1/spam-keywords/{keyword_id}")
async def index(request):
keyword_id = request.match_info.get("keyword_id", "")
try:
keyword_id = int(keyword_id)
except:
return web.json_response({"error": "Invalid keyword ID"}, status=400)
try:
keyword = database.SpamKeyword.get_or_none(database.SpamKeyword.id == keyword_id)
if keyword:
keyword.delete_instance()
return web.json_response({"message": "ok"})
else:
return web.json_response({"error": "Keyword not found"}, status=404)
except Exception as e:
return web.json_response({"error": str(e)}, status=500)
# mark message as spam or not spam
@routes.post("/api/v1/lxmf-messages/{hash}/spam")
async def index(request):
message_hash = request.match_info.get("hash", "")
data = await request.json()
is_spam = data.get("is_spam", False)
try:
message = database.LxmfMessage.get_or_none(database.LxmfMessage.hash == message_hash)
if message:
message.is_spam = is_spam
message.updated_at = datetime.now(timezone.utc)
message.save()
return web.json_response({"message": "ok"})
else:
return web.json_response({"error": "Message not found"}, status=404)
except Exception as e:
return web.json_response({"error": str(e)}, status=500)
# called when web app has started
async def on_startup(app):
@@ -3065,6 +3183,7 @@ class ReticulumMeshChat:
"rssi": db_lxmf_message.rssi,
"snr": db_lxmf_message.snr,
"quality": db_lxmf_message.quality,
"is_spam": db_lxmf_message.is_spam,
"created_at": db_lxmf_message.created_at,
"updated_at": db_lxmf_message.updated_at,
}
@@ -3089,10 +3208,51 @@ class ReticulumMeshChat:
query = query.on_conflict(conflict_target=[database.LxmfUserIcon.destination_hash], update=data)
query.execute()
# check if a destination is blocked
def is_destination_blocked(self, destination_hash: str) -> bool:
try:
blocked = database.BlockedDestination.get_or_none(
database.BlockedDestination.destination_hash == destination_hash
)
return blocked is not None
except:
return False
# check if message content matches spam keywords
def check_spam_keywords(self, title: str, content: str) -> bool:
try:
spam_keywords = database.SpamKeyword.select()
search_text = (title + " " + content).lower()
for keyword in spam_keywords:
if keyword.keyword.lower() in search_text:
return True
return False
except:
return False
# check if message has attachments and should be rejected
def has_attachments(self, lxmf_fields: dict) -> bool:
try:
if LXMF.FIELD_FILE_ATTACHMENTS in lxmf_fields:
return len(lxmf_fields[LXMF.FIELD_FILE_ATTACHMENTS]) > 0
if LXMF.FIELD_IMAGE in lxmf_fields:
return True
if LXMF.FIELD_AUDIO in lxmf_fields:
return True
return False
except:
return False
# handle an lxmf delivery from reticulum
# NOTE: cant be async, as Reticulum doesn't await it
def on_lxmf_delivery(self, lxmf_message: LXMF.LXMessage):
try:
source_hash = lxmf_message.source_hash.hex()
# check if source is blocked - reject immediately
if self.is_destination_blocked(source_hash):
print(f"Rejecting LXMF message from blocked source: {source_hash}")
return
# check if this lxmf message contains a telemetry request command from sideband
is_sideband_telemetry_request = False
@@ -3107,8 +3267,28 @@ class ReticulumMeshChat:
print("Ignoring received LXMF message as it is a telemetry request command")
return
# upsert lxmf message to database
self.db_upsert_lxmf_message(lxmf_message)
# check for spam keywords
is_spam = False
message_title = lxmf_message.title if hasattr(lxmf_message, 'title') else ""
message_content = lxmf_message.content if hasattr(lxmf_message, 'content') else ""
# check spam keywords
if self.check_spam_keywords(message_title, message_content):
is_spam = True
print(f"Marking LXMF message as spam due to keyword match: {source_hash}")
# reject attachments from blocked sources (already checked above, but double-check)
if self.has_attachments(lxmf_fields):
if self.is_destination_blocked(source_hash):
print(f"Rejecting LXMF message with attachments from blocked source: {source_hash}")
return
# reject attachments from spam sources
if is_spam:
print(f"Rejecting LXMF message with attachments from spam source: {source_hash}")
return
# upsert lxmf message to database with spam flag
self.db_upsert_lxmf_message(lxmf_message, is_spam=is_spam)
# update lxmf user icon if icon appearance field is available
try:
@@ -3178,7 +3358,7 @@ class ReticulumMeshChat:
self.message_router.handle_outbound(lxmf_message)
# upserts the provided lxmf message to the database
def db_upsert_lxmf_message(self, lxmf_message: LXMF.LXMessage):
def db_upsert_lxmf_message(self, lxmf_message: LXMF.LXMessage, is_spam: bool = False):
# convert lxmf message to dict
lxmf_message_dict = self.convert_lxmf_message_to_dict(lxmf_message)
@@ -3201,6 +3381,7 @@ class ReticulumMeshChat:
"rssi": lxmf_message_dict["rssi"],
"snr": lxmf_message_dict["snr"],
"quality": lxmf_message_dict["quality"],
"is_spam": is_spam,
"updated_at": datetime.now(timezone.utc),
}
@@ -3444,6 +3625,13 @@ class ReticulumMeshChat:
# NOTE: cant be async, as Reticulum doesn't await it
def on_audio_call_announce_received(self, aspect, destination_hash, announced_identity, app_data, announce_packet_hash):
# check if source is blocked - drop announce and path if blocked
identity_hash = announced_identity.hash.hex()
if self.is_destination_blocked(identity_hash):
print(f"Dropping audio call announce from blocked source: {identity_hash}")
RNS.Transport.drop_path(destination_hash)
return
# log received announce
print("Received an announce from " + RNS.prettyhexrep(destination_hash) + " for [call.audio]")
@@ -3468,6 +3656,13 @@ class ReticulumMeshChat:
# NOTE: cant be async, as Reticulum doesn't await it
def on_lxmf_announce_received(self, aspect, destination_hash, announced_identity, app_data, announce_packet_hash):
# check if source is blocked - drop announce and path if blocked
identity_hash = announced_identity.hash.hex()
if self.is_destination_blocked(identity_hash):
print(f"Dropping announce from blocked source: {identity_hash}")
RNS.Transport.drop_path(destination_hash)
return
# log received announce
print("Received an announce from " + RNS.prettyhexrep(destination_hash) + " for [lxmf.delivery]")