How gokkul-ks/placebot works
PlaceBot is a placement-preparation chatbot delivered as a FastAPI service with a single-page chat interface. A user submits a message; the response engine handles conversational shortcuts, searches the curated Q&A dataset semantically, optionally expands short keyword queries, then routes by confidence to a dataset answer, Ollama, or a fallback. Configuration, request/response models, and the UI are supporting product layers. The sampled sources establish the chatbot’s internal dependencies; the README documents the API and major capabilities. main.py and the UI source were not sampled, so their implementation-level wiring is not asserted.
Main entry points: chatbot.py Response router · main.py FastAPI service · index.html Chat UI · ollama_client.py Ollama client
Components 11
Chat Experience
- Chat UI
static/index.html - FastAPI service
main.py - API models
schemas.py
Response Engine
- Response router
chatbot.py - Text rules
text_utils.py
Knowledge Search
- Semantic matching
chatbot.py - Placement Q&A
data.csv
Configuration & Integrations
- Settings
config.py - Ollama client
ollama_client.py
Other components
- Chat user
- Ollama service
Connections 10
- Chat user → Chat UI: asks
README.md - Chat UI → FastAPI service: submits message
- FastAPI service → Response router: gets reply
README.md - FastAPI service → API models: validates exchange
- Response router → Text rules: normalizes and shortcuts
chatbot.py - Response router → Semantic matching: searches questions
chatbot.py - Response router → Placement Q&A: loads Q&A
README.md - Response router → Settings: reads thresholds
chatbot.py - Response router → Ollama client: requests generation
chatbot.py - Ollama client → Ollama service: calls API