anuradhakadam035-ops/swaroop-project-

Purpose and Topology

VIDYASETU is an AI-assisted career-guidance platform for students. It combines registration/profile data, MCQ assessments, skills, and education context to produce career matches, skill-gap analysis, learning roadmaps, courses, jobs, resumes, interview support, and reports. The repository is a static multi-page frontend plus a Python backend API.

Application Entry Points

• Backend API bootstrap and route composition: backend/main.py

• Backend configuration and database access boundary: backend/config.py, backend/database.py

• Browser entry pages include frontend/index.html, frontend/dashboard.html, and role-specific admin pages such as frontend/admin.html.

• Frontend JavaScript calls backend APIs through frontend/css/js/api.js; feature pages have corresponding feature scripts.

API and Domain Boundary

FastAPI-style routers organize the API by user workflow rather than a single monolith: authentication, profile/dashboard, assessment, career/recommendations, and AI features. Representative API modules are backend/routers/auth.py, backend/routers/assessment.py, backend/routers/recommendation.py, backend/routers/roadmap.py, and backend/routers/resume.py.

Schemas define request/response contracts separately from persistent entities, while models represent the stored domain: backend/schemas/assessment.py, backend/schemas/recommendation.py, backend/models/assessment.py, backend/models/recommendation.py.

Identity, Roles, and Persistence

Users and administrators are distinct authenticated actors. User authentication/security is isolated in backend/security/auth.py; administrator authentication is isolated in backend/security/admin_auth.py. Core persisted entities cover users, assessments/questions/answers/results, recommendations, resumes, jobs, chats, interviews, and admin-managed courses: backend/models/user.py, backend/models/assessment_result.py, backend/models/student_answer.py, backend/models/admin_course.py.

This provides the key boundary between HTTP/UI input, validated schemas, authorization, and database-backed records.

Assessment-to-Recommendation Pipeline

The primary control flow is: user registers/authenticates → completes assessment questions → answers and results are stored → recommendation logic evaluates profile/skills/assessment signals → generated career matches and gaps are returned to dashboard/report views.

Relevant stages are implemented in backend/routers/assessment.py, backend/services/recommendation_engine.py, backend/services/skill_service.py, and backend/services/roadmap_generator.py. Rule-based career knowledge and skill extraction/matching utilities support the pipeline: backend/utils/career_rules.py, backend/utils/skill_extractor.py, backend/utils/skill_matcher.py.

AI and Career-Development Services

Gemini is the explicit AI-provider integration boundary: backend/core/gemini.py. AI-facing router/service layers support general AI output, chatbot conversations, and interview assistance: backend/routers/ai.py, backend/services/ai_service.py, backend/services/chatbot_ai.py, backend/services/interview_ai.py.

Derived guidance is exposed as recommendations, roadmaps, courses, and jobs through backend/routers/recommendation.py, backend/routers/course.py, and backend/routers/jobs.py.

Resume, ATS, and Document Processing

Resume workflows ingest and analyze candidate documents, calculate ATS-oriented feedback, and generate reportable output. The API/service/utilities boundary is visible in backend/routers/resume.py, backend/services/resume_service.py, backend/services/ats_service.py, and backend/utils/pdf_reader.py. PDF generation is separated in backend/services/pdf_service.py.

Administration and Analytics

Administrative routes manage admins, students, courses, jobs, questions, and dashboard operations: backend/routers/admin.py, backend/routers/admin_questions.py, backend/routers/admin_course.py, backend/routers/admin_students.py. Analytics has its own API/service layer—backend/routers/analytics.py, backend/services/analtyics_service.py—allowing aggregate reporting to remain distinct from student-facing recommendation generation.