Data management and analysis of metal–organic framework synthesis using data models

Table of Contents

• Overview
• Features
• Installation
• Usage
• Contributing

---

Overview

Preprint available at: https://chemrxiv.org/doi/full/10.26434/chemrxiv.10001842/v1 This project formats, validates, serializes, and analyzes the given MOF synthesis data using data models. It was developed to demonstrate the usefulness of data models for realizing FAIR data and software management in chemical synthesis projects. The data models and codes can be reused by those who are interested in developing or using such a workflow.

---

Features

• Two example datasets of MOF synthesis
• Formatting synthesis and characterization data into a well-defined and interoperable structure with the data models
• Rigorous data validation using JSON schema
• Data serialization into known formats, XDL and MPIF
• Phase mole fraction analysis of PXRD data using reference patterns
• Phase yield calculation based on mole fractions and yields using molar masses
• Decision tree modeling of the main products and convex hull analysis of phase yields with synthesis parameters
• Modular scripts based on the data model APIs

---

Installation

# Install uv
pip install uv
# Clone the repository
git clone https://github.com/FAIRChemistry/mof-synthesis-data-modeling.git
cd mof-synthesis-data-modeling
# Install dependencies in a virtual environment
uv pip install -e .

Install Graphviz for decision tree visualization

macOS: brew install graphviz
Debian/Ubuntu: sudo apt install graphviz
Windows: https://graphviz.org/download/

---

Usage

Formatting, validation, and serialization into XDL

uv run scripts/format_and_serialize_all.py

Serialization into MPIF

cd scripts/mofsy2mpif
uv run npm install
uv run npm start
cd ../..

Phase mole fraction analysis of PXRD patterns

uv run marimo edit

1. Workspace > scripts > pxrd_analysis.mo.py
2. Run all slate cells (right bottom)

Decision tree modeling

uv run scripts/generate_decision_trees.py

Convex hull plotting

uv run marimo edit

1. Workspace > scripts > convex-hull.mo.py
2. Run all slate cells (right bottom)

---

Contributing

• Use uv run ... as a substitution for python ....
• New packages can be added using uv add ... instead of pip install ....
• Best results using the VSCode debugger are achieved by running uv sync --all-packages. Then just run Python Debugger: Clear Cache and Reload Window in the command palette.