existing_CSD_data.py

import ccdc
import sys
import os
import numpy as np

entries = ccdc.io.EntryReader('CSD')
family = sys.argv[1]
common_name = sys.argv[2]
Polymorphs = []
volumes_by_polymorph = {}
temperature_by_polymorph = {}
Refcode_by_polymorph = {}
dois_by_polymorph = {}
for i in range(9):
    if i == 0:
        num = ""
    else:
        num = "%02i" % i
    full_name = f"{family}{num}"
    print(full_name)
    try:
        entry = entries.entry(full_name)
    except Exception as e:
        print(f"Could not retrieve crystal for {full_name}: {e}")
        break
    polymorph = entry.polymorph if entry.polymorph else "Unknown"
    volume = entry.crystal.cell_volume
    if volume is not None:
        if polymorph not in volumes_by_polymorph:
            volumes_by_polymorph[polymorph] = []
        volumes_by_polymorph[polymorph].append(volume)
    temperature = entry.temperature if entry.temperature else "Unknown"
    if temperature is not None:
        if polymorph not in temperature_by_polymorph:
            temperature_by_polymorph[polymorph] = []
        temperature_by_polymorph[polymorph].append(temperature)
    refcode = entry.identifier
    if polymorph not in Refcode_by_polymorph:
        Refcode_by_polymorph[polymorph] = []
    Refcode_by_polymorph[polymorph].append(refcode)
    doi = entry.publication.doi if entry.publication and entry.publication.doi else "Unknown"
    if polymorph not in dois_by_polymorph:
        dois_by_polymorph[polymorph] = []
    dois_by_polymorph[polymorph].append(doi)
if os.path.exists(common_name):
    print(f"folder {common_name} exists.")
else:
    print(f"making folder {common_name}")
    os.mkdir(common_name)

for polymorph, volumes in volumes_by_polymorph.items():
    output_file = open(os.path.join(common_name, f"{polymorph}_volume.csv"), 'w')
    output_file.write("%s\n" % (Refcode_by_polymorph[polymorph][0]))  # should be best R factor refcode rather than first really
    output_file.write("Identifier, Property              , Value (Angstrom^3),          Std,   N,    Name,    Reference    ,                 Comment\n")
    output_file.write(f"1, Unit Cell Volume @  {temperature_by_polymorph[polymorph][0]}, {np.mean(volumes):.2f}, {np.std(volumes):.2f}, {len(volumes)},    ,{dois_by_polymorph[polymorph][0]}, generated by automatic script; add Name and double check temperature etc.\n")