using Codex.Dal.Entities;

namespace Codex.Dal.Services;

/// <summary>
/// Service for interacting with Ollama LLM endpoints
/// </summary>
public interface IOllamaService
{
    /// <summary>
    /// Generates a response from an Ollama model given conversation context
    /// </summary>
    /// <param name="endpoint">Ollama endpoint URL (e.g., "http://localhost:11434")</param>
    /// <param name="model">Model name (e.g., "phi", "codellama:7b")</param>
    /// <param name="systemPrompt">System prompt defining agent behavior</param>
    /// <param name="contextMessages">Previous conversation messages for context</param>
    /// <param name="userMessage">Current user message to respond to</param>
    /// <param name="temperature">Temperature parameter (0.0 to 2.0)</param>
    /// <param name="maxTokens">Maximum tokens to generate</param>
    /// <param name="cancellationToken">Cancellation token</param>
    /// <returns>Response from the model with token counts</returns>
    Task<OllamaResponse> GenerateAsync(
        string endpoint,
        string model,
        string systemPrompt,
        List<ConversationMessage> contextMessages,
        string userMessage,
        double temperature,
        int maxTokens,
        CancellationToken cancellationToken = default
    );
}

/// <summary>
/// Response from Ollama generation request
/// </summary>
public record OllamaResponse
{
    /// <summary>
    /// Generated response content
    /// </summary>
    public string Content { get; init; } = string.Empty;

    /// <summary>
    /// Number of tokens in the input prompt
    /// </summary>
    public int? InputTokens { get; init; }

    /// <summary>
    /// Number of tokens in the generated output
    /// </summary>
    public int? OutputTokens { get; init; }
}