Creating a New Chat Model in Askimo

This guide explains how to implement a new chat model provider in Askimo. By following these steps, you can integrate any chat model API with the Askimo CLI.

Architecture Overview

Askimo uses a modular architecture for chat models with the following key components:

1. ChatService: Interface that defines the contract for all chat models
2. ChatModelFactory: Interface for creating chat model instances
3. ProviderSettings: Interface for model-specific configuration
4. ModelProvider: Enum that identifies different model providers
5. ModelRegistry: Central registry that manages all model factories

Each model provider (like OpenAI or Ollama) has its own implementation of these interfaces.

Implementation Steps

1. Add LangChain4j Dependency

First, add the appropriate LangChain4j dependency for your provider to the build.gradle.kts file:

dependencies {
    // Existing dependencies
    implementation("dev.langchain4j:langchain4j:1.2.0")
    implementation("dev.langchain4j:langchain4j-open-ai:1.2.0")
    implementation("dev.langchain4j:langchain4j-ollama:1.2.0")
    
    // Add your provider's LangChain4j implementation
    implementation("dev.langchain4j:langchain4j-your-provider:1.2.0")
}

You need to find the appropriate LangChain4j implementation for your provider. Check the LangChain4j GitHub repository or Maven Central for available implementations. If there isn’t an existing implementation for your provider, you may need to create your own or adapt one of the existing implementations.

2. Create a New Provider Enum Value

First, add your provider to the ModelProvider enum in io.askimo.core.providers.ModelProvider:

@Serializable
enum class ModelProvider {
    @SerialName("OPEN_AI") OPEN_AI,
    @SerialName("OLLAMA") OLLAMA,
    @SerialName("YOUR_PROVIDER") YOUR_PROVIDER,  // Add your provider here
    @SerialName("UNKNOWN") UNKNOWN,
}

3. Create Provider Settings

Create a settings class that implements ProviderSettings interface. This class will store configuration specific to your provider:

// File: io.askimo.core.providers.yourprovider.YourProviderSettings.kt

@Serializable
data class YourProviderSettings(
    var apiKey: String = "", 
    override var presets: Presets = Presets(Style.BALANCED, Verbosity.NORMAL),
) : ProviderSettings {
    override fun describe(): List<String> =
        listOf(
            "apiKey:      ${apiKey.take(5)}***",
            "presets:     $presets",
        )
}

4. Implement the Model Factory

Create a factory class that implements ChatModelFactory. This class will be responsible for creating instances of your model:

// File: io.askimo.core.providers.yourprovider.YourProviderModelFactory.kt

class YourProviderModelFactory : ChatModelFactory {
    override val provider: ModelProvider = ModelProvider.YOUR_PROVIDER
    
    override fun availableModels(settings: ProviderSettings): List<String> =
        try {
            // Implement logic to fetch available models from your provider
            // This could be an API call or a hardcoded list
            
            // Example:
            // val client = YourProviderClient(settings.apiKey)
            // client.listModels().map { it.id }
            
            listOf("model1", "model2", "model3")  // Replace with actual implementation
        } catch (e: Exception) {
            println("⚠️ Failed to fetch models from YourProvider: ${e.message}")
            emptyList()
        }
    
    override fun defaultModel(): String = "default-model-name"  // Set your default model
    
    override fun defaultSettings(): ProviderSettings =
        YourProviderSettings(
        )
    
    override fun create(
        model: String,
        settings: ProviderSettings,
        memory: ChatMemory,
    ): ChatService {
        require(settings is YourProviderSettings) {
            "Invalid settings type for YourProvider: ${settings::class.simpleName}"
        }
        
        // create the chat model
        return AiServices
            .builder(ChatService::class.java)
            .streamingChatModel(chatModel)
            .chatMemory(memory)
            .build()
    }
}

5. Register Your Factory

Register your factory in the ProviderRegistry. The best place to do this is by modifying the init block in Provideregistry.kt:

init {
    // Register known factories
    register(OpenAiModelFactory())
    register(OllamaModelFactory())
    register(YourProviderModelFactory())  // Add your factory here
}

Alternatively, you can register your factory programmatically at runtime:

ProviderRegistry.register(YourProviderModelFactory())

Example: Implementation Reference

For reference, here are the key components of existing implementations:

OpenAI Implementation

• Settings: OpenAiSettings - Contains API key and presets
• Factory: OpenAiModelFactory - Creates OpenAI models and fetches available models

Ollama Implementation

• Settings: OllamaSettings - Contains base URL and presets
• Factory: OllamaModelFactory - Creates Ollama models and fetches available models

Testing Your Implementation

After implementing your provider, you can test it by:

1. Building and running the Askimo CLI
2. Setting your provider as the active provider:

   askmio> :setprovider YOUR_PROVIDER
   

3. Setting any required parameters:

   askimo> :setparam api_key your-api-key
   

4. Listing available models:

   askimo> :models
   

5. Chatting with a specific model:

   askimo> :setparam model your-model-name
   askimo> What is the capital of Viet Nam?
   

Conclusion

By following these steps, you can integrate any chat model provider with Askimo. The modular architecture makes it easy to add new providers while maintaining a consistent interface for users.

Remember to handle errors gracefully and provide clear feedback to users when something goes wrong with your provider’s API.