Emotify

Detect face's sentiment from given image uri or base64 image data

Dev

Follow the steps below to setup the credential and GCP project

• Make sure you have install node, npm and serverless
• Make sure you have create a project on GCP with service account key
• Copy your service account's json file to ~/.gcloud/keyfile.json
• Change provider.project in serverless.yml to match your project-key (you can get it from JSON file)

Run function locally

Have a read at https://cloud.google.com/functions/docs/emulator

# Install functions-emulator 
npm install -g @google-cloud/functions-emulator

# Config the GCP project
  # Replace $YOUR_PROJECT_ID with GCP project id
functions config set projectId $YOUR_PROJECT_ID

# Start emulator
functions start

# Deploy a function locally
functions deploy http --trigger-http

# Execute a function
  # Replace $BASE64 with base64 encoded image data
functions call http --data='{"content": "$BASE64"}'

Deploy function

npm i
serverless deploy

You will get function endpoint in the console

Execute function

# GET endpoint with image URI
  # Replace $FUNCTION_URI with actual function URI endpoint
  # Replace $URI with actual image URI
curl -G $FUNCTION_URI --data-urlencode "uri=$URI"

# POST endpoint with image URI
  # Replace $FUNCTION_URI with actual function URI endpoint
  # Replace $BASE64 with base64 encoded image data
curl -X POST \
  $FUNCTION_URI \
  -H 'Content-Type: application/json' \
  -H 'cache-control: no-cache' \
  -d '{
    "content": "$BASE64"
}'