Pubsub aqua (#19)

* add branch to ci to test release

* add some comments

* release!

* delete release branch

* DHT api in terms of PubSub

Co-authored-by: DieMyst <dmitry.shakhtarin@fluence.ai>
This commit is contained in:
Dmitry Kurinskiy 2021-06-08 18:26:38 +03:00 committed by GitHub
parent a24a530be1
commit 19d4a28f61
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 3700 additions and 13 deletions

3611
npm/package-lock.json generated

File diff suppressed because it is too large Load Diff

View File

@ -33,7 +33,7 @@
},
"homepage": "https://github.com/fluencelabs/aqua-dht#readme",
"devDependencies": {
"@fluencelabs/aqua-cli": "^0.1.4-136",
"@fluencelabs/aqua-cli": "^0.1.5-140",
"typescript": "^3.9.5"
}
}

100
npm/pubsub.aqua Normal file
View File

@ -0,0 +1,100 @@
import "dht.aqua"
import "@fluencelabs/aqua-lib/builtin.aqua"
-- Get peers closest to the topic's hash in Kademlia network
-- These peers are expected to store list of subscribers of this topic
func getNeighbours(node_id: PeerId, topic: string) -> []PeerId:
on node_id:
k <- Op.string_to_b58(topic)
nodes <- Kademlia.neighborhood(k, false)
<- nodes
-- If this peer have set node_id as a subscriber for topic,
-- this call will prevent subscriber from re-subscribing
-- so that eventually it will disappear from the subscribers list
func removeSubscriber(node_id: PeerId, topic: string):
on node_id:
t <- Peer.timestamp_sec()
AquaDHT.clear_host_value(topic, t)
-- Create a topic: register it on the closest peers
-- node_id is a peer with Kademlia access to start with
func initTopic(node_id: PeerId, topic: string):
nodes <- getNeighbours(node_id, topic)
for n <- nodes par:
on n:
try:
t <- Peer.timestamp_sec()
AquaDHT.register_key(topic, t, false, 0)
-- Create a topic and subscribe to it
-- %init_peer_id% (current client) will become a subscriber
func initTopicAndSubscribe(node_id: PeerId, topic: string, value: string, relay_id: ?PeerId, service_id: ?string):
nodes <- getNeighbours(node_id, topic)
for n <- nodes par:
on n:
try:
t <- Peer.timestamp_sec()
AquaDHT.register_key(topic, t, false, 0)
AquaDHT.put_value(topic, value, t, relay_id, service_id, 0)
-- Create a topic and make the given node a subscriber to it
func initTopicAndSubscribeNode(subscriber_node_id: PeerId, topic: string, value: string, service_id: ?string):
nil: ?string
on subscriber_node_id:
t <- Peer.timestamp_sec()
AquaDHT.register_key(topic, t, false, 0)
r <- AquaDHT.put_host_value(topic, value, t, nil, service_id, 0)
nodes <- getNeighbours(subscriber_node_id, topic)
for n <- nodes par:
on n:
try:
tt <- Peer.timestamp_sec()
AquaDHT.register_key(topic, tt, false, 0)
AquaDHT.propagate_host_value(r, tt, 0)
-- Subscribe to a topic
-- Note: topic must be already initiated
func subscribe(node_id: PeerId, topic: string, value: string, relay_id: ?PeerId, service_id: ?string):
nodes <- getNeighbours(node_id, topic)
for n <- nodes par:
on n:
try:
t <- Peer.timestamp_sec()
AquaDHT.put_value(topic, value, t, relay_id, service_id, 0)
-- Subscribe a node to the given topic
-- Note: topic must be already initiated
func subscribeNode(subscriber_node_id: PeerId, topic: string, value: string, service_id: ?string):
on subscriber_node_id:
nil: ?string
t <- Peer.timestamp_sec()
r <- AquaDHT.put_host_value(topic, value, t, nil, service_id, 0)
nodes <- getNeighbours(subscriber_node_id, topic)
for n <- nodes par:
on n:
try:
tt <- Peer.timestamp_sec()
AquaDHT.register_key(topic, tt, false, 0)
AquaDHT.propagate_host_value(r, tt, 0)
-- Find the list of subscribers for the given topic
func findSubscribers(node_id: PeerId, topic: string) -> []Record:
nodes <- getNeighbours(node_id, topic)
res: *GetValuesResult
for n <- nodes par:
on n:
try:
t <- Peer.timestamp_sec()
res <- AquaDHT.get_values(topic, t)
on node_id:
v <- AquaDHT.merge_two(res!.result, res!1.result)
<- v.result
-- Execute the given code on subscribers
-- Note that you can provide another Aqua function as an argument to this one
func executeOnSubscribers(node_id: PeerId, topic: string, call: Record -> ()):
subs <- findSubscribers(node_id, topic)
for r <- subs par:
on r.peer_id via r.relay_id:
call(r)