How krishgarg344/intelligent-customer-case-processing-system works
The product is a FastAPI customer-ticket processing service. A client submits ticket text; the endpoint predicts category and priority, maps the category to a department, generates a Gemini draft reply, and returns those results. The README also describes dataset preparation and model-training stages, represented by the available dataset and notebooks. The saved-model artifacts mentioned by the API are not present in the supplied tree, so model-training-to-runtime artifact flow is not shown.
Main entry points: main.py Ticket Endpoint · labeled_tickets_with_priority.csv Labeled Tickets · ticket_classification.ipynb Category Training · priority_prediction.ipynb Priority Training
Components 10
Ticket API
- Ticket Endpoint
main.py
Ticket Models
- Category Inference
main.py - Priority Inference
main.py - Category Training
models/ticket_classification.ipynb - Priority Training
models/priority_prediction.ipynb - Labeled Tickets
dataset/labeled_tickets_with_priority.csv
Routing and Replies
- Department Mapping
department_assignment.py - Draft Generation
main.py
Other components
- CRM Client
- Gemini API
Connections 9
- CRM Client → Ticket Endpoint: submits ticket
main.py - Ticket Endpoint → Category Inference: predicts category
main.py - Ticket Endpoint → Priority Inference: predicts priority
main.py - Ticket Endpoint → Department Mapping: assigns department
main.py - Ticket Endpoint → Draft Generation: generates draft
main.py - Draft Generation → Gemini API: requests content
main.py - Ticket Endpoint → CRM Client: returns results
main.py - Labeled Tickets → Category Training: training data
README.md - Labeled Tickets → Priority Training: labeled tickets
README.md