ohlstrom/Crystallography-Open-Database-XRD-Patterns

1 2---3license: cc0-1.04task_categories:5  - other6tags:7  - chemistry8  - crystallography9  - xrd10  - materials-science11pretty_name: COD XRD Patterns12size_categories:13  - 100K<n<1M14---15 16# Crystallography Open Database — Synthetic XRD Patterns17 18436,196 synthetic powder X-ray diffraction (XRD) patterns generated from every CIF file in the [Crystallography Open Database](https://www.crystallography.net/cod/) (COD).19 20## What's in the dataset21 22Each pattern is a simulated powder diffractogram computed from a published crystal structure. The dataset covers the full chemical and structural diversity of COD: metals, minerals, organics, MOFs, and everything in between. No measured (experimental) patterns are included — these are purely computational.23 24## Files and subsets25 26| File | Patterns | Purpose |27|---|---|---|28| `COD_xrd_patterns_1000.pt` | 1,000 | Quick prototyping |29| `COD_xrd_patterns_10000.pt` | 10,000 | Development / hyperparameter tuning |30| `COD_xrd_patterns_50000.pt` | 50,000 | Development / hyperparameter tuning |31| `COD_xrd_patterns_100000.pt` | 100,000 | Development / hyperparameter tuning |32| `COD_xrd_patterns.pt` | 436,196 | Full dataset |33 34The smaller subsets are random samples drawn from the full set.35 36## Data format37 38Each `.pt` file is a Python dictionary saved with `torch.save()`. Load it with:39 40```python41import torch42 43data = torch.load("COD_xrd_patterns.pt")44 45patterns = data["patterns"]    # torch.Tensor, shape (N, 4500)46filenames = data["filenames"]  # list[str], length N47```48 49- **`patterns`**: A float tensor where each row is a 4,500-point XRD intensity vector.50- **`filenames`**: The corresponding COD CIF filenames (e.g. `"4326570.cif"`).51 52## Generation parameters53 54| Parameter | Value |55|---|---|56| Software | pymatgen `XRDCalculator` |57| Radiation | Cu K-alpha (lambda = 1.54184 angstrom) |58| 2-theta range | 0 to 90 degrees |59| Step size | 0.02 degrees |60| Points per pattern | 4,500 |61| Normalisation | Min-max scaled to [0, 1] |62 63Patterns were generated by reading each CIF file, computing the theoretical diffraction peak positions and intensities, and interpolating onto the uniform 2-theta grid. Regions with no diffraction peaks are zero-padded.64 65## Finding the original CIF files66 67Every filename maps directly to a COD entry. Strip the `.cif` extension to get the COD ID, then download the structure file:68 69```70Filename:  4326570.cif71COD ID:    432657072URL:       https://www.crystallography.net/cod/4326570.cif73```74 75## License76 77CC0 1.0 Universal — same as the Crystallography Open Database itself.78 79## Citation80 81If you use this dataset, please cite:82 83**COD:**84Grazulis, S. et al. "Crystallography Open Database — an open-access collection of crystal structures." *Journal of Applied Crystallography*, 42(4), 726–729, 2009.85 86**pymatgen:**87Ong, S. P. et al. "Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis." *Computational Materials Science*, 68, 314–319, 2013.88