openai classification action OpenAI Classification Action

Provided by: "Apache Software Foundation"

Support Level for this Kamelet is: "Preview"

Classifies a sentence using the OpenAI API.

It requires an OpenAI account (beta accounts available at https://beta.openai.com/).

Training data for classification can be provided in advance by uploading a file with purpose="classifications", see: https://beta.openai.com/docs/api-reference/files/upload.

The Kamelet accepts data in:

Configuration Options

The following table summarizes the configuration options available for the openai-classification-action Kamelet:

Property Name Description Type Default Example

authorizationToken *

Authorization Token

The authorization token to use to contact the openAI API

string

engine

Engine

The OpenAI engine to use

string

"davinci"

file

Examples File

The ID of an uploaded file containing examples for classification

string

"file-9HbeJXxtG7pg1YTXRx1v1fmn"

format

Format

The response format ('text/plain' or 'application/json' are supported)

string

"text/plain"
Fields marked with (*) are mandatory.

Usage

This section summarizes how the openai-classification-action can be used in various contexts.

Knative Action

The openai-classification-action Kamelet can be used as intermediate step in a Knative binding.

openai-classification-action-binding.yaml
apiVersion: camel.apache.org/v1alpha1
kind: KameletBinding
metadata:
  name: openai-classification-action-binding
spec:
  source:
    ref:
      kind: Kamelet
      apiVersion: camel.apache.org/v1alpha1
      name: timer-source
    properties:
      message: "Hello"
  steps:
  - ref:
      kind: Kamelet
      apiVersion: camel.apache.org/v1alpha1
      name: openai-classification-action
    properties:
      authorizationToken: "The Authorization Token"
  sink:
    ref:
      kind: Channel
      apiVersion: messaging.knative.dev/v1
      name: mychannel

Make sure you have Camel K installed into the Kubernetes cluster you’re connected to.

Save the openai-classification-action-binding.yaml file into your hard drive, then configure it according to your needs.

You can run the action using the following command:

kubectl apply -f openai-classification-action-binding.yaml

Dependencies

The Kamelet needs the following dependencies:

  • github:apache.camel-kamelets:camel-kamelets-utils:main-SNAPSHOT

  • camel:jackson

  • camel:jsonpath

  • camel:core

  • camel:kamelet

  • camel:http

Binding to Knative using the Kamel CLI:

The procedure described above can be simplified into a single execution of the kamel bind command:

kamel bind timer-source?message=Hello --step openai-classification-action -p "step-0.authorizationToken=The Authorization Token" channel:mychannel

This will create the KameletBinding under the hood and apply it to the current namespace in the cluster.

Kafka Action

The openai-classification-action Kamelet can be used as intermediate step in a Kafka binding.

openai-classification-action-binding.yaml
apiVersion: camel.apache.org/v1alpha1
kind: KameletBinding
metadata:
  name: openai-classification-action-binding
spec:
  source:
    ref:
      kind: Kamelet
      apiVersion: camel.apache.org/v1alpha1
      name: timer-source
    properties:
      message: "Hello"
  steps:
  - ref:
      kind: Kamelet
      apiVersion: camel.apache.org/v1alpha1
      name: openai-classification-action
    properties:
      authorizationToken: "The Authorization Token"
  sink:
    ref:
      kind: KafkaTopic
      apiVersion: kafka.strimzi.io/v1beta1
      name: my-topic

Ensure that you’ve installed Strimzi and created a topic named my-topic in the current namespace. Make also sure you have Camel K installed into the Kubernetes cluster you’re connected to.

Save the openai-classification-action-binding.yaml file into your hard drive, then configure it according to your needs.

You can run the action using the following command:

kubectl apply -f openai-classification-action-binding.yaml

Binding to Kafka using the Kamel CLI:

The procedure described above can be simplified into a single execution of the kamel bind command:

kamel bind timer-source?message=Hello --step openai-classification-action -p "step-0.authorizationToken=The Authorization Token" kafka.strimzi.io/v1beta1:KafkaTopic:my-topic

This will create the KameletBinding under the hood and apply it to the current namespace in the cluster.

Kamelet source file

Have a look at the following link:

https://github.com/apache/camel-kamelets/blob/main/openai-classification-action.kamelet.yaml

Back to top