Adapters
Adapters are basically just pipelines to transfer messages from the chat itself to the bot's brain. All they do is basically handling their own permissions (and self-commands like join) and differencing direct commands from natural language requests. Token must authenticate themselves through a valid token generated on the dashboard (in request header x-access-token or in body token).
Adapters can pass data to the brain using the data: {} object in body. Data will be passed to skill handlers. It is good practise to send an unique channel identifier and the name of the user that entered the command.
Connecting to the Brain
You may implement a connector that use the HTTP API of the brain or use websockets (socket-io). Socket events emitted by the adapter are:
| event | HTTP equivalent | params | callback | description |
|---|---|---|---|---|
nlp |
/nlp |
{ phrase, data: {} } | (err, body) | Send a phrase to be analyzed by the brain. |
command |
/command |
{ command, data: {} } | (err, body) | Send a command to be executed by the brain. |
converse |
/converse |
{ thread_id, phrase, data: {} } | (err, body) | Send a phrase to be analyzed to the brain. |
hook-accept |
/hooks |
hook_id | (error) | Accept the creation of a hook. |
Socket events received by the adapter are:
| event | callback | description |
|---|---|---|
hook |
(hook_id, body) | Hook triggered by the brain, body will contain the message to display |