AI Predictive & Generative Hub

Smart is supercharged with a premium suite of AI predictive widgets, automated syllabi drafters, textbook MCQ visual workspaces, 3D flippable study flashcards, and student leave advisors out-of-the-box, powered by **Google Gemini & OpenAI**.

To move from mock demo sandbox data to live production databases (such as Node.js, Spring Boot, Laravel, or .NET APIs), follow this standard three-step Angular integration workflow.

1. Fetch Real Database Records via HttpClient

Define a backend service method to query your REST endpoint or database server for the latest student marks or attendance logs.

// src/app/core/service/student-data.service.ts
import { Injectable, inject } from '@angular/core';
import { HttpClient } from '@angular/common/http';
import { Observable } from 'rxjs';

@Injectable({ providedIn: 'root' })
export class StudentDataService {
  private http = inject(HttpClient);
  private apiUrl = 'https://your-school-api.com/api/v1';

  // Fetch live student analytics data from SQL database
  getStudentRecords(studentId: string): Observable<any> {
    return this.http.get<any>(`${this.apiUrl}/students/${studentId}/records`);
  }
}

2. Interpolate Dynamic Payloads & Send to AI Service

Inject your custom data service and the template's standalone AiService into your dashboard component. Fetch the real database records first, serialize the payload, and request Gemini / OpenAI parsing.

// src/app/teacher/examination/marks-entry/marks-entry.component.ts
import { Component, OnInit, inject } from '@angular/core';
import { StudentDataService } from '@core/service/student-data.service';
import { AiService } from '@core/service/ai.service';

@Component({ selector: 'app-custom-grading' })
export class CustomGradingComponent implements OnInit {
  private dataService = inject(StudentDataService);
  private aiService = inject(AiService);

  studentName = 'John Doe';
  isLoading = false;
  aiSuggestedFeedback = '';

  generateSmartRemarks() {
    this.isLoading = true;

    // 1. Fetch live scores from the SQL/NoSQL backend
    this.dataService.getStudentRecords('std_1082').subscribe({
      next: (databasePayload) => {
        
        // 2. Build contextual prompt with actual database fields
        const prompt = `You are a teacher. Analyze the following actual database record:
        Student: \${databasePayload.name}
        Subject: Mathematics
        Score: \${databasePayload.marksObtained}/\${databasePayload.maxMarks}
        Attendance Rate: \${databasePayload.attendanceRate}%
        
        Generate exactly 1 constructive feedback remark.`;

        // 3. Dispatch payload to the configured AI Model
        this.aiService.postPrompt(prompt).subscribe({
          next: (aiResult) => {
            this.aiSuggestedFeedback = aiResult.trim();
            this.isLoading = false;
          },
          error: () => this.isLoading = false
        });
      }
    });
  }
}

3. Parse and Map Structured AI JSON Back to the UI

Force models to respond in structured JSON. Once received, safely parse the payload and update the component properties reactively to bind with material UI widgets.

generateJsonMetrics() {
  const prompt = `Output exactly one JSON object: { "riskPercentage": 75, "reasons": ["Low Attendance"] }`;
  
  this.aiService.postPrompt(prompt).subscribe({
    next: (responseText) => {
      try {
        // Enforce parsing safely
        const parsedData = JSON.parse(responseText.trim());
        
        // Update responsive chart/gauge values instantly
        this.riskGaugeValue = parsedData.riskPercentage;
        this.riskInsightsList = parsedData.reasons;
      } catch (e) {
        console.error("AI response did not match expected JSON format", e);
      }
    }
  });
}