How ibisanmi1/pepphyschem-hl works
PepPhysChem-HL is a research platform for peptide physicochemical profiling and deep-learning half-life estimation, exposed through both a command-line interface and a Gradio web interface. Users submit a sequence or CSV/FASTA batch; the runtime prepares sequences and features, predicts half-life, optionally computes physicochemical profiles, and returns results and batch analyses. The graph includes the separate model-inference utility and major documented analysis outputs. Some internal wiring is based on the SOURCE INDEX and README; unsampled implementation details are not assumed.
Main entry points: run_PepPhysChem_HL.py CLI pipeline · inference.py Feature inference · data_processing.py Input preparation · physicochemical_analyzer.py Physchem profiling
Components 16
User interfaces
- Web interface
app.py - CLI pipeline
run_PepPhysChem_HL.py - Feature inference
scripts/inference.py
Prediction workflow
- Input preparation
src/data_processing.py - Embedding datasets
src/embedding_datasets.py - Sequence filters
src/modelling_filters.py - Batch reports
run_PepPhysChem_HL.py
Analysis engine
- Physchem profiling
src/physicochemical_analyzer.py - Neural models
src/models.py - Embedding models
src/embedding_models.py - Inference utilities
src/utils.py
Model resources
- Peptide datasets
data/half_life_dataset.csv - Model checkpoints
checkpoints - Result files
Other components
- Researcher
- Hugging Face Hub
Connections 20
- Researcher → Web interface: submits sequences
README.md - Researcher → CLI pipeline: submits sequences
README.md - Researcher → Feature inference: requests predictions
scripts/inference.py - Web interface → Model checkpoints: loads weights
- Web interface → Hugging Face Hub: downloads checkpoints
app.py - CLI pipeline → Embedding datasets: prepares embeddings
run_PepPhysChem_HL.py - CLI pipeline → Embedding models: uses models
run_PepPhysChem_HL.py - CLI pipeline → Physchem profiling: profiles peptides
run_PepPhysChem_HL.py - CLI pipeline → Model checkpoints: loads checkpoint
README.md - CLI pipeline → Batch reports: generates analyses
README.md - CLI pipeline → Result files: writes results
README.md - Input preparation → Physchem profiling: extracts features
src/data_processing.py - Input preparation → Sequence filters: filters sequences
src/data_processing.py - Embedding datasets → Physchem profiling: extracts descriptors
src/embedding_datasets.py - Embedding datasets → Sequence filters: filters sequences
src/embedding_datasets.py - Feature inference → Input preparation: prepares features
scripts/inference.py - Feature inference → Neural models: runs predictions
scripts/inference.py - Feature inference → Physchem profiling: profiles inputs
scripts/inference.py - Peptide datasets → Input preparation: supplies records
- Model checkpoints → Neural models: parameterizes