A service for generating structured scoring dimensions from problem statements, reference answers, total score, and subject metadata. It is designed as a reusable building block inside AI grading workflows.
- Accepts problem content, reference answers, total score, and subject information
- Calls a workflow service to generate scoring rubrics
- Returns structured score-dimension results for downstream grading systems
- Filters oversized base64 image payloads before processing
AI grading systems become much more useful when they can produce structured scoring rubrics instead of only raw scores. This repository represents that middle-layer capability.
app/main.py: FastAPI application entryapp/api/routers/score_dimension.py: score-dimension endpointsapp/services/dify_service.py: workflow integration logicapp/core/config.py: runtime configuration
pip install -r requirements.txt
uvicorn app.main:app --host 0.0.0.0 --port 8080- Scoring rubric generation
- FastAPI service structure
- Workflow integration
- Input sanitation for base64-heavy content
This public repository is a cleaned release version. Local environment files, logs, caches, and virtual environments were excluded.