scripts/summarize_xrd_thresholds.py

#!/usr/bin/env python
import csv
from pathlib import Path

import numpy as np

out = []
def emit(metric, *values):
    line = ' '.join([metric, *map(str, values)])
    print(line)
    out.append((metric, values))

stats = Path('/workspace/mp-20/xrd_stats/mp20_xrd_material_stats.csv')
peaks = Path('/workspace/mp-20/xrd_stats/mp20_xrd_peaks.csv')
mins = []
lasts = []
retry = []
with stats.open() as f:
    for row in csv.DictReader(f):
        if row['status'] == 'ok':
            mins.append(float(row['min_peak_gap']))
            lasts.append(float(row['last_peak_2theta']))
            if row['error']:
                retry.append(row)
mins = np.array(mins)
lasts = np.array(lasts)
emit('materials', len(lasts))
emit('boundary_retry', len(retry), retry[0]['material_id'] if retry else '')
for t in [0.001, 0.005, 0.01, 0.02, 0.05, 0.1, 0.2]:
    emit(f'per_material_min_gap_le_{t:g}', int((mins <= t).sum()), float((mins <= t).mean()))
for t in [120, 130, 140, 150, 160, 170, 175, 178, 179, 180]:
    emit(f'last_peak_le_{t:g}', int((lasts <= t).sum()), float((lasts <= t).mean()))
thresholds = [0.001, 0.005, 0.01, 0.02, 0.05, 0.1, 0.2]
counts = dict.fromkeys(thresholds, 0)
total = 0
prev_key = None
prev_theta = None
smallest = []
with peaks.open() as f:
    reader = csv.DictReader(f)
    for row in reader:
        key = (row['split'], row['index'], row['material_id'])
        theta = float(row['two_theta'])
        if key == prev_key:
            gap = theta - prev_theta
            total += 1
            for t in thresholds:
                if gap <= t:
                    counts[t] += 1
            if len(smallest) < 10 or gap < smallest[-1][0]:
                smallest.append((gap, key, prev_theta, theta))
                smallest.sort(key=lambda x: x[0])
                smallest = smallest[:10]
        prev_key = key
        prev_theta = theta
emit('all_adjacent_total', total)
for t in thresholds:
    emit(f'all_adjacent_gap_le_{t:g}', counts[t], counts[t] / total)
emit('smallest_gaps', '')
for rank, item in enumerate(smallest, start=1):
    gap, key, left, right = item
    emit(f'smallest_gap_{rank}', gap, '|'.join(key), left, right)
with open('/workspace/mp-20/xrd_stats/mp20_xrd_threshold_summary.csv', 'w', encoding='utf-8') as f:
    f.write('metric,value1,value2,value3,value4\n')
    for metric, values in out:
        vals = list(values)[:4]
        vals += [''] * (4 - len(vals))
        f.write(metric + ',' + ','.join(map(str, vals)) + '\n')