Trainkore SDK

Features

Create chat prompts.
Retrieve chat prompts by ID.
Fetch logs by project ID.
Fetch logs by log ID.

Installation

You can install the Trainkore SDK via npm:

npm install trainkore

Usage

Initialize Trainkore SDK

To start using the SDK, initialize it with your API key. The API key can be found on the Trainkore platform.

import Trainkore from 'trainkore';
 
const apiKey = 'your-trainkore-api-key';  // Replace with your API key(Settings page)
 
const trainkore = new Trainkore(apiKey);

Create a Chat Prompt

Once the SDK is initialized, you can create a chat prompt by calling the create method.

 
(async () => {
  try {
    const chatPrompt = await trainkore.chatPrompt.create({
      projId: 'your-project-id',
      versionId: 'your-version-id',
      userId: 'your-user-id',
      userInput: {
        userInput: {
          '1': 'write an email on taking leave from office',
          '2': 'Write an essay on {{topic}}',
          '3': 'Tell me more about {{movie_name}}'
        },
        selectedOptions: {
          '1': { value: 'user', label: 'user' },
          '2': { value: 'user', label: 'user' },
          '3': { value: 'user', label: 'user' }
        },
        inputValue: {
          topic: 'dhoni',
          movie_name: 'My Name is Khan'
        }
      },
      "userInputData": {
        "userInputData": {
          "1" :"Who is lord Ram?",
          "2" :"Who is Aamir khan?",
          "3" :"Write python code to get maximum number in an array"
 
        },
        "selectedOptionsData": {
            "1": {
                "value": "user",
                "label": "user"
            },
            "2": {
                "value": "user",
                "label": "user"
            },
            "3": {
                "value": "human",
                "label": "user"
            }
        }
    },
      parametersValue: {
        stopSliderValue: 50,
        topP: 0.1,
        temperature: 0.3,
        max_tokens: 150,
        frequency_penalty: 0.1,
        presence_penalty: 0.1,
        model: { value: 'gpt-4o', label: 'gpt-4o' }
      }
    });
    console.log('Chat Prompt Created:', chatPrompt);
  } catch (error) {
    console.error('Error creating chat prompt:', error.message);
  }
})();

Retrieve a Chat Prompt by ID

You can retrieve a specific chat prompt by its ID using the getChatPromptById method.

 
(async () => {
  try {
    const chatId = 'your-chat-id'; // Replace with your chat prompt ID(You'll find in app)
    const chatPrompt = await trainkore.chatPrompt.getChatPromptById(chatId);
    console.log('Retrieved Chat Prompt:', chatPrompt);
  } catch (error) {
    console.error('Error retrieving chat prompt:', error.message);
  }
})();

Fetch Logs by Project ID

To fetch logs associated with a project by its projId, use the getLogsByProjectId method.

 
(async () => {
  try {
    const projId = 'your-project-id';  // Replace with your project ID("You'll find in app")
    const logs = await trainkore.log.getLogsByProjectId(projId);
    console.log('Project Logs:', logs);
  } catch (error) {
    console.error('Error fetching logs:', error.message);
  }
})();

Fetch Log by Log ID

To retrieve a specific log by logId, use the getLogByLogId method.

 
(async () => {
  try {
    const logId = 'your-log-id';  // Replace with your log ID(You'll find in app)
    const log = await trainkore.log.getLogByLogId(logId);
    console.log('Log Details:', log);
  } catch (error) {
    console.error('Error fetching log:', error.message);
  }
})();

Contributing

Contributions are welcome! Please submit a pull request or file an issue if you encounter any problems.