python-grpcflow

playground for python grpc to serve Dialogflow fulfillment

setup

python3 -m venv ./env
source env/bin/activate
pip install -r requirements.txt

server

python -m grpc_tools.protoc \
    -I. \
    --python_out=. \
    --grpc_python_out=. \
    grpcflow.proto

for windows

python -m grpc_tools.protoc     -I.     --python_out=.     --grpc_python_out=.     grpcflow.proto

start the server:

export PORT=50061
python server.py

for windows

set PORT=50061
python server.py

Now the server should be listening on port 50061. We'll use the tool grpcurl to manually interact with it. On Linux and Mac you can install it with curl -s https://grpc.io/get_grpcurl | bash.

grpcurl \
    -d '{"responseId": "aabbccdd-4096", 
    "session": "project/fujen-smart-iot/agent/sessions/deadbeaf", 
    "queryResult": { 
        "queryText": "your name?", 
        "parameters": {}, 
        "allRequiredParamsPresent": true, 
        "fulfillmentText": "my name is dialogflow", 
        "fulfillmentMessages": {"text": {"text":["my name is dialogflow"]}}, 
        "outputContexts": [{"name":"projects/fujen-smart-iot/agent/sessions/aabbccdd/contexts/__system_counters__"}], 
        "parameters":{"no-input":0,"no-match":0}, 
        "intent": {"name":"projects/fujen-smart-iot/agent/intents/aabbccdd",  "displayName":"get-agent-name"}, 
        "intentDetectionConfidence":1, 
        "languageCode":"zh-tw" 
        } 
    }' \
    --plaintext \
    -proto grpcflow.proto \
    localhost:50061 \
    DialogflowWebhook.fulfillmentWebhook

test not handled intents

grpcurl \
    -d '{"responseId": "aabbccdd-4096", 
    "session": "project/fujen-smart-iot/agent/sessions/deadbeaf", 
    "queryResult": { 
        "queryText": "your name?", 
        "parameters": {}, 
        "allRequiredParamsPresent": true, 
        "fulfillmentText": "my name is dialogflow", 
        "fulfillmentMessages": {"text": {"text":["default name is dialogflow"]}}, 
        "outputContexts": [], 
        "parameters":{}, 
        "intent": {"name":"intents/aabbccdd",  "displayName":"get-my-name"}, 
        "intentDetectionConfidence":1, 
        "languageCode":"zh-tw" 
        } 
    }' \
    --plaintext \
    -proto grpcflow.proto \
    localhost:50061 \
    DialogflowWebhook.fulfillmentWebhook

client

sample node.js client is published as @tomyjwu/grpcflow-client and source code is located in https://github.com/tomyjwu/grpcflow

container

Now we can build our image. In order to deploy to Cloud Run, we'll be pushing to the gcr.io container registry, so we'll tag it accordingly.

export GCP_PROJECT=<Your GCP Project Name>
export GCP_PROJECT=fujen-smart-iot
docker build -t gcr.io/$GCP_PROJECT/grpcflow-server:latest .

Now, before we deploy to Cloud Run, let's make sure that we've containerized our application properly. We'll test it by spinning up a local container.

docker run -d -p 50061:50061 -e PORT=50061 gcr.io/$GCP_PROJECT/grpcflow-server:latest

If all goes well, grpcurl will give us the same result as before:

grpcurl \
    -d '{"responseId": "aabbccdd-4096", 
    "session": "project/fujen-smart-iot/agent/sessions/deadbeaf", 
    "queryResult": { 
        "fulfillmentMessages": {"text": {"text":["default name is dialogflow"]}}, 
        "intent": {"name":"intents/aabbccdd",  "displayName":"get-agent-name"}, 
        "languageCode":"zh-tw" 
        } 
    }' \
    --plaintext \
    -proto grpcflow.proto \
    localhost:50061 \
    DialogflowWebhook.fulfillmentWebhook

Deploying to Cloud Run

Cloud Run needs to pull our application from a container registry, so the first step is to push the image we just built.

Make sure that you can use gcloud and are able to push to gcr.io.

gcloud auth login
gcloud auth configure-docker

check your active project

gcloud config get-value project

if not correct, please set it by

gcloud config set project $GCP_PROJECT
```bash

Now we can push our image.

```bash
docker push gcr.io/$GCP_PROJECT/grpcflow-server:latest

Finally, we deploy our application to Cloud Run:

gcloud run deploy --image gcr.io/$GCP_PROJECT/grpcflow-server:latest --platform managed

You may be prompted for auth. If so, choose the unauthenticated option.

This command will give you a message like

Service [grpcflow-server] revision [grpcflow-server-00001-baw] has been deployed and is serving 100 percent of traffic at https://grpcflow-server-xyspwhk3xq-uc.a.run.app

We can now access the gRPC service at grpcflow-server-xyspwhk3xq-uc.a.run.app:443. Go ahead and leave the https:// prefix off. Notice that this endpoint is secured with TLS even though the server we wrote is using a plaintext connection. Cloud Run provides a proxy that provides TLS for us. We'll account for that in our grpcurl invocation by leaving off the --plaintext flag.

grpcurl \
    -d '{"responseId": "aabbccdd-4096", 
    "session": "project/fujen-smart-iot/agent/sessions/deadbeaf", 
    "queryResult": { 
        "fulfillmentMessages": {"text": {"text":["default name is dialogflow"]}}, 
        "intent": {"name":"intents/aabbccdd",  "displayName":"get-agent-name"}, 
        "languageCode":"zh-tw" 
        } 
    }' \
    -proto grpcflow.proto \
    grpcflow-server-xyspwhk3xq-uc.a.run.app:443 \
    DialogflowWebhook.fulfillmentWebhook

And now you've got an auto-scaling gRPC Dialogflow webhook service!

tomyjwu / python-grpcflow