How yaroslavsobolev/robowski-maps works
Project type and purpose
• This is a scientific software suite (Python-based monorepo) that automates robotic experiments (pipetting, washing, microspectrometry), ingests raw spectral and run data, performs calibration and spectral unmixing, fits kinetic models, and visualizes chemical reaction hyperspaces for a Nature paper.
High-level components
A. Hardware Control Layer
• zeus-pipetter: scripts (main.py, scheduler.py, pipetter.py, etc.) and Arduino/GRBL firmware for pipetting robot automation and NanoDrop spectrometer control
• washing_station: AxiDraw/FlowEZ pump interface to wash vial plates
• CRAIC automation: PyAutoGUI + OpenCV barcode capture to drive CRAIC microspectrometer GUI (under robowski/uv_vis_absorption_spectroscopy/craic_automation)
B. Data Ingestion & Calibration
• raw data files: spectrophotometer CSV/DSW/MSP files, run_info CSV, etc. stored under run-specific directories or BPRF/test folders
• calibration routines: calibrator.py, bkg_pca.py, absorbance_errorbar_model.py (under robowski/uv_vis_absorption_spectroscopy)
• environment variable ROBOCHEM_DATA_PATH points to large raw dataset archive
C. Spectral Unmixing & Data Processing
• spectral_unmixing_tutorial.ipynb & robowski/uv_vis_absorption_spectroscopy/examples: high-level “calibration_*.py” then “process_*.py” scripts produce results/product_concentration.csv
• summary_of_reaction_yield_data: consolidated CSV yield tables
D. Kinetic Modeling
• robowski/kinetics_models/E1_kinetics, simpleSN1_kinetics, Ugi_kinetics modules implement model fitting, bootstrapping, MCMC (emcee), and output numpy/CSV artifacts
E. Visualization
• robowski/visualize_results: static and animated viewers (mayavi), scripts to generate 3D/4D maps (plot_3d_map_for_*.py, view_4d_data_for_*.py)
• scripts under misc_scripts/figures_for_articles for publication-quality figures
F. Utilities, Theory & Tests
• geometric_algorithms, smoothness_theory, outlier_handling, misc_scripts modules for auxiliary computations, simulation, outlier marking
• tests folder contains end-to-end and unit tests for data parsing, processing, and expected outputs
Relationships & data flow
• Hardware Control → Raw Data Files
↳ Data Ingestion reads raw CSV/DSW/MSP
↳ Calibration produces calibration models (.npy, PCA components)
↳ Spectral Unmixing uses calibration + raw spectra → product_concentration.csv
↳ Summary tables assembled in summary_of_reaction_yield_data
↳ Kinetic Modeling consumes yields → model output arrays/plots
↳ Visualization consumes yields & model predictions → 3D/4D maps
↳ Notebooks link everything interactively
Architectural patterns & principles
• Layered, modular monorepo
• Hardware interface modules isolated from core library
• Core library (“robowski” package) organized by functionality (uv_vis_absorption_spectroscopy, kinetics_models, visualize_results, etc.)
• Extensibility: new reactions or hardware can be added as new subfolders under examples or new modules
• Reproducibility: environment.yml / requirements.txt for environment isolation; tests ensure regression safety