Conversation Platform 4 Marketing (CP4M)

CP4M is a conversational marketing platform which enables advertisers to integrate their customer-facing chatbots with FB Messenger/WhatsApp, in order to meet customers where they are and drive native conversations on the advertiser's owned infra.

Details

CP4M acts as a bridge between existing Messenger and Whatsapp APIs and existing LLM APIs (OpenAI, LLaMa, etc.)

Quickstart

1 Configuration file

Almost all of CP4M's configurations are set via a configuration file. This file can be in either JSON or TOML format, we use TOML here for brevity.

1.1 create a file to hold the configuration

$ mkdir /tmp/cp4m
$ touch /tmp/cp4m/cp4m.toml

1.2 copy one of the texts below into your newly created TOML file

Example: Whatsapp & OpenAI

port = 8080

[[plugins]]
name = "openai_test"
type = "openai"
model = "gpt-3.5-turbo"
api_key = "<your api key here>"

[[stores]]
name = "memory_test"
type = "memory"
storage_duration_hours = 1
storage_capacity_mbs = 1

[[handlers]]
name = "whatsapp_test"
type = "whatsapp"
verify_token = "<your verification token here>"
app_secret = "<your verification app secret here>"
access_token = "<you access token here>"

[[pre_processors]]
type = "aws_s3"
name = "aws_s3"
aws_access_key_id="<your-aws-console-access-key>"
aws_secret_access_key="<your-aws-console-secret-key>"
region = "<your-aws-s3-bucket-region>"
bucket = "<your-aws-bucket-name>"

[[services]]
webhook_path = "/<webhook-pathname>"
plugin = "openai_test"
store = "memory_test"
handler = "whatsapp_test"
pre_processors =["aws_s3"]

Example: Messenger & Llama 2 (via Hugging Face)

port = 8080

[[plugins]]
name = "hf_test"
type = "hugging_face"
endpoint = "https://example.com"
token_limit = 1000
api_key = "<your api_token here>"

[[stores]]
name = "memory_test"
type = "memory"
storage_duration_hours = 1
storage_capacity_mbs = 1

[[handlers]]
type = "messenger"
name = "messenger_test"
verify_token = "<your verification token here>"
app_secret = "<your verification app secret here>"
page_access_token = "<your page access token here>"

[[services]]
webhook_path = "/<webhook-pathname>"
plugin = "hf_test"
store = "memory_test"
handler = "messenger_test"

2 Build and Run CP4M

We provide a Dockerfile to build and run CP4M however you're also welcome to follow the steps in the Dockerfile to build locally.

2.1 Clone repo

$ git clone [email protected]:facebookincubator/CP4M.git

2.2 build and run CP4M via docker

The below commands changes your directory to the CP4M directory, builds a docker image named cp4m and then runs that images and exposing the port 8080 for webhooks.

$ cd CP4M
$ docker build -t cp4m .
$ docker run -v /tmp/cp4m:/tmp/cp4m -e CP4M_CONFIGURATION_FILE=/tmp/cp4m/cp4m.toml -p 8080:8080 cp4m

License

CP4M is MIT licensed.