How kartikeya1911/adslm-adaptive_datascience_language_model works
ADSLM is an AutoML and data-science copilot for uploaded tabular data. Users submit a dataset through the Streamlit UI; the FastAPI backend profiles it, detects a task, preprocesses data, benchmarks models, and returns explanations, governance checks, scale estimates, and reports. The README documents the hybrid Streamlit-to-FastAPI deployment and report downloads. The route-to-service relationships are supported by the route imports and documented pipeline; training-to-evaluation is supported by the training source index. Some service internals and frontend calls were not excerpted, so those links rely on README documentation.
Main entry points: routes.py API Routes · app.py Streamlit UI · training.py Model Training · reports Report Files
Components 19
User Interface
- Streamlit UI
frontend/app.py
API Layer
- FastAPI App
main.py - API Routes
app/api/routes.py - Runtime Config
app/core/config.py
Data & ML Pipeline
- Dataset Profiling
app/services/dataset_analyzer.py - Task Detection
app/services/task_detection.py - Leakage-Safe Prep
app/services/preprocessing.py - Model Recommendation
app/services/model_recommendation.py - Model Training
app/services/training.py - Model Evaluation
app/services/evaluation.py
Insights & Governance
- Explainability
app/services/explainability.py - Narrative Insights
app/services/insight_generator.py - Expertise Adaptation
app/services/expertise_adaptation.py - Governance Audit
app/services/regulatory_compliance.py - Scale Profiling
app/services/bigdata_engine.py - Report Generation
app/services/report_generator.py - Saved Models
saved_models - Report Files
reports
Other components
- Dataset User
Connections 20
- Dataset User → Streamlit UI: uploads data
README.md - Streamlit UI → FastAPI App: sends requests
README.md - FastAPI App → API Routes: registers routes
main.py - API Routes → Dataset Profiling: profiles data
app/api/routes.py - API Routes → Task Detection: detects task
README.md - API Routes → Leakage-Safe Prep: preprocesses data
README.md - API Routes → Model Recommendation: selects candidates
README.md - API Routes → Model Training: trains models
README.md - Model Training → Model Evaluation: evaluates models
app/services/training.py - API Routes → Explainability: explains model
app/api/routes.py - API Routes → Narrative Insights: generates insights
README.md - API Routes → Expertise Adaptation: adapts insights
README.md - API Routes → Governance Audit: audits compliance
app/api/routes.py - API Routes → Scale Profiling: profiles scale
app/api/routes.py - API Routes → Report Generation: builds reports
app/api/routes.py - Model Training → Saved Models: saves model
app/services/training.py - Report Generation → Report Files: writes reports
app/services/report_generator.py - Streamlit UI → Report Files: retrieves reports
README.md - Runtime Config → Saved Models: configures storage
app/core/config.py - Runtime Config → Report Files: configures storage
app/core/config.py