joshua12-5/biopinn

How joshua12-5/biopinn works

BIOPINN models drug transport and biological response in a three-zone tumor, trains a parametric PINN surrogate, and uses it for evaluation, treatment optimization, and interactive results exploration. The main workflow is synthetic FDM data generation → PINN training → surrogate-based biological prediction and analysis. The graph also preserves the documented ablation and manuscript-results branches. Some analysis and dashboard details are documented or indexed but not fully excerpted, so their connections are limited to visible references and README statements.

Main entry points: server.py Dashboard Service · evaluate.py Model Evaluation · optimize.py Dose Optimization · train.py Training Engine

Components 16

Tumor Physics

Configuration src/config.py
Microenvironment src/microenvironment.py
FDM Solver src/fdm_solver.py

PINN Learning

Dataset Pipeline src/data_pipeline.py
BIOPINN Model src/model.py
Physics Losses src/losses.py
Training Engine src/train.py

Analysis

Biological Response src/biology.py
Model Evaluation src/evaluate.py
Dose Optimization src/optimize.py
Physics Ablation src/ablation.py
Scientific Plots src/visualize.py
Manuscript Assets src/results.py

Results Interface

Dashboard Service app/server.py
Dashboard UI app/static/js/dashboard.js

Other components

Researcher

Connections 31

Researcher → Dataset Pipeline: generates data README.md
Dataset Pipeline → FDM Solver: solves simulations src/data_pipeline.py
FDM Solver → Microenvironment: gets fields src/fdm_solver.py
Dataset Pipeline → Training Engine: provides tensors README.md
Training Engine → BIOPINN Model: trains src/train.py
Training Engine → Physics Losses: computes losses src/train.py
Physics Losses → Microenvironment: uses fields src/losses.py
Researcher → Training Engine: trains model README.md
Biological Response → BIOPINN Model: predicts with src/biology.py
Model Evaluation → Biological Response: predicts fields src/evaluate.py
Model Evaluation → FDM Solver: resolves reference src/evaluate.py
Model Evaluation → Physics Losses: measures residuals src/evaluate.py
Dose Optimization → Biological Response: evaluates response src/optimize.py
Dose Optimization → FDM Solver: compares solver src/optimize.py
Dose Optimization → Microenvironment: uses fields src/optimize.py
Physics Ablation → Training Engine: trains baseline src/ablation.py
Physics Ablation → Model Evaluation: compares models src/ablation.py
Scientific Plots → Microenvironment: uses radial grid src/visualize.py
Manuscript Assets → Dose Optimization: builds assets from src/results.py
Manuscript Assets → Model Evaluation: builds assets from src/results.py
Manuscript Assets → Physics Ablation: builds assets from src/results.py
Manuscript Assets → Biological Response: builds assets from src/results.py
Researcher → Dashboard Service: explores results README.md
Dashboard Service → Biological Response: requests response app/server.py
Dashboard Service → Model Evaluation: requests evaluation app/server.py
Dashboard Service → Dose Optimization: requests optimization app/server.py
Dashboard Service → Physics Ablation: requests ablation app/server.py
Dashboard UI → Dashboard Service: calls service
Configuration → FDM Solver: configures
Configuration → Training Engine: configures src/train.py
Configuration → Dose Optimization: configures

More Python diagrams: vectornguyen76/face-recognition · x35f/alpha2 · aws-samples/aws-cdk-project-structure-python · codedog-ai/codedog · datacurve-ai/pier

Drawn by GitDiagram on October 1, 2026 from the repository's file tree, README and source files. Read as Markdown ·