freewillai

Free Will AI

Requirements to test in testnet

1- We have to deploy the contracts on testnet and get their addresses.

export FREEWILLAI_TOKEN_ADDRESS=<token-address>
export FREEWILLAI_TASK_RUNNER_ADDRESS=<task_runner-address>

2- Each node needs a private key so we have to create an account for each node on this testnet

python -m freewillai.node -s 200 -p <node-private-key>

3- The user/client also needs a private key

export PRIVATE_KEY=<user-private-key>

# or put it in .env file
# .env
PRIVATE_KEY=<user-private-key>

4- Set the endpoint provider to environment

export FREEWILLAI_PROVIDER=http://<host>:<port>...

Commands to deploy demo (Digital Ocean server command line format):

Spin up anvil and ipfs for workers and client interaction

sudo ./docker-manager setup

Spin up workers. There is 9 workers availables

one worker (attached)

sudo ./docker-manager up worker 1
:'                             ^^^ 
                          worker_id (1-9)'

group of workers (detached)

sudo ./docker-manager up workers 5
:'                              ^^^
                   workers amount to get up (1-9) 
                   Empty will run all of workers (9)'

Spin up repl on demo.freewillai.org

sudo ./docker-manager up repl

# To make it tolerant of out-of-memory issues
sudo OOM_TOLERANCE=1 ./docker-manager up repl

To stop containers replace up to down.

Example:

sudo ./docker-manager down repl
sudo ./docker-manager down workers 5
sudo ./docker-manager down worker 1

Also we can get up all (anvil, ipfs, 9 workers and repl)

sudo ./docker-manager up all

Get down all docker containers even non-freewilai containers
(maybe change it to get down just freewillai containers)

sudo ./docker-manager down all 

View worker logs

sudo ./docker-manager logs worker 2
:'                               ^^^ 
                            worker_id (1-9)'

Troubleshooting

Suddenly repl stopped and demo.freewillai.org page has 404 error
It often is due to out of memory issue. To address it you just need to restart the repl

sudo ./docker-manager down repl
sudo ./docker-manager up repl

Now we can run this command to restart repl every time it dies

sudo OOM_TOLERANCE=1 ./docker-manager up repl

(C) Copyright FreeWillAI