Building the Tutor Service Layer in C#

Building the Tutor Service Layer

In the previous lesson, we explored the SessionManager class, which plays a crucial role in managing tutoring session data within our application. Now, we will take the next step in our journey by building the Tutor Service Layer. This layer is essential for integrating the DeepSeek language model with tutoring sessions, allowing us to process student queries and generate tailored explanations. By the end of this lesson, you will understand how to set up the TutorService class, create tutoring sessions, and process academic questions using DeepSeek models via HTTP requests.

The service layer acts as a bridge between the model layer, where data is managed, and the AI model, which generates educational responses. It is responsible for orchestrating the flow of data and ensuring that student interactions are handled effectively. Let's dive into the details of setting up this important component.

Setting Up the TutorService Class

The TutorService class is the heart of our service layer. It is responsible for managing tutoring sessions and interacting with the DeepSeek model to generate educational responses. To begin, we need to set up the class and its components.

First, we include the necessary namespaces, such as System, System.IO, System.Net.Http, and any project-specific namespaces for session management. We also use Guid.NewGuid() to generate unique session IDs. Here’s how the class is initialized in C#:

using System;
using System.IO;
using System.Net.Http;
using PersonalTutor.Models;

namespace PersonalTutor.Services
{
    public class TutorService
    {
        private SessionManager sessionManager;
        private HttpClient httpClient;
        private string systemPrompt;

        public TutorService()
        {
            sessionManager = new SessionManager();
            httpClient = new HttpClient();
            systemPrompt = LoadSystemPrompt("data/system_prompt.txt");
        }

        // ... other methods will be added here
    }
}

In this setup, we instantiate SessionManager to manage tutoring data, initialize an HttpClient for making HTTP requests to the DeepSeek API, and load the systemPrompt using the LoadSystemPrompt method, which we'll discuss next.

Processing Student Queries

The ProcessQuery method is where the educational magic happens. It processes student questions, interacts with the DeepSeek model to generate tutoring explanations, and updates the session history. Below, we outline the steps involved in this process, followed by the corresponding code implementation:

  1. Retrieve the session using GetSession, and throw an exception if the session is not found.
  2. Add the student's query to the session history.
  3. Send the conversation, including the system prompt and all previous exchanges, to the DeepSeek model via an HTTP request to generate a response.
  4. Add the tutor's explanation to the session history and return it to the student.
  5. Handle any errors with the HTTP client gracefully.
public async Task<string> ProcessQuery(string studentId, string sessionId, string query)
{
    var session = sessionManager.GetSession(studentId, sessionId);
    if (session == null)
    {
        throw new ArgumentException("Session not found");
    }

    // Add student query
    sessionManager.AddMessage(studentId, sessionId, "user", query);

    try
    {
        // Retrieve the conversation using the SessionManager
        var conversation = sessionManager.GetConversation(studentId, sessionId)
            .Select(msg => new { role = msg.Role, content = msg.Content })
            .ToArray();

        // Prepare API key and endpoint
        var apiKey = Environment.GetEnvironmentVariable("OPENAI_API_KEY");
        var baseUri = Environment.GetEnvironmentVariable("OPENAI_BASE_URL");
        if (string.IsNullOrEmpty(apiKey))
        {
            throw new InvalidOperationException("OPENAI_API_KEY environment variable is not set.");
        }
        if (string.IsNullOrEmpty(baseUri))
        {
            throw new InvalidOperationException("OPENAI_BASE_URL environment variable is not set.");
        }
        var endpoint = baseUri.TrimEnd('/') + "/v1/chat/completions";

        // Prepare payload
        var payload = new
        {
            model = "deepseek-ai/DeepSeek-V3",
            messages = conversation,
            temperature = 0.7,
            max_tokens = 500
        };
        var jsonPayload = JsonSerializer.Serialize(payload);

        // Prepare HTTP request
        using var request = new HttpRequestMessage(HttpMethod.Post, endpoint);
        request.Headers.Authorization = new AuthenticationHeaderValue("Bearer", apiKey);
        request.Content = new StringContent(jsonPayload, Encoding.UTF8, "application/json");

        // Send request
        var response = await httpClient.SendAsync(request);
        var body = await response.Content.ReadAsStringAsync();

        if (!response.IsSuccessStatusCode)
        {
            throw new InvalidOperationException($"Error from DeepSeek API ({response.StatusCode}): {body}");
        }

        using var doc = JsonDocument.Parse(body);
        var aiMessage = doc.RootElement
                           .GetProperty("choices")[0]
                           .GetProperty("message")
                           .GetProperty("content")
                           .GetString()
                           ?.Trim() ?? "";

        // Add AI response to session history
        sessionManager.AddMessage(studentId, sessionId, "assistant", aiMessage);

        return aiMessage;
    }
    catch (Exception e)
    {
        throw new InvalidOperationException($"Error getting AI response: {e.Message}");
    }
}

In the context of a personal tutor, we configure our DeepSeek model with specific parameters to optimize its educational performance. The temperature is set to 0.7, which balances accuracy and creativity in the tutor's explanations, ensuring they are both informative and engaging. The max_tokens is set to 500, allowing the model to provide detailed educational content without overwhelming the student, thus maintaining an effective learning experience.

Sign up

Join the 1M+ learners on CodeSignal

Be a part of our community of 1M+ users who develop and demonstrate their skills on CodeSignal