PDF comparison

src/diffpy/similarity/compare_pdfs.py

@@ -0,0 +1,106 @@
+from pathlib import Path
+
+import numpy as np
+
+from diffpy.pdfmorph.pdfmorphpy import pdfmorph
+from diffpy.utils.parsers import loadData
+
+grnames = [
+    "Si_O_only_10_common_grs",
+    "Ti_O_only_10_common_grs",
+    "Zr_O_only_10_common_grs",
+]
+
+gr_list = []
+
+for grname in grnames:
+    grdir = Path(f"PDFs/{grname}")
+
+    gr_list_component = list(grdir.iterdir())
+    gr_list_component.sort()
+
+    gr_list.extend(gr_list_component)
+
+grname = "All_O_only_10_common_grs"
+
+datmatrw = []
+datmatrw_morph = []
+
+datmatpcc = []
+datmatpcc_morph = []
+
+labels = []
+
+for grfile1 in gr_list:
+    with open(grfile1, "r") as grread:
+        for line in grread:
+            if "composition =" in line:
+                labels.append(line.split("=")[1].strip())
+
+    datrw_row = []
+    datpcc_row = []
+
+    datrw_row_morph = []
+    datpcc_row_morph = []
+
+    for grfile2 in gr_list:
+        rawdata = pdfmorph(grfile1, grfile2, noplot="")[0]
+        morphdata = pdfmorph(grfile1, grfile2, noplot="", stretch=0, scale=1)[
+            0
+        ]
+
+        datrw_row.append(rawdata["Rw"])
+        datpcc_row.append(rawdata["Pearson"])
+
+        datrw_row_morph.append(morphdata["Rw"])
+        datpcc_row_morph.append(morphdata["Pearson"])
+
+    datmatrw.append(datrw_row)
+    datmatrw_morph.append(datrw_row_morph)
+
+    datmatpcc.append(datpcc_row)
+    datmatpcc_morph.append(datpcc_row_morph)
+
+datmatrw = np.array(datmatrw)
+datmatrw = np.sqrt(datmatrw * datmatrw.T)
+
+datmatrw_morph = np.array(datmatrw_morph)
+datmatrw_morph = np.sqrt(datmatrw_morph * datmatrw_morph.T)
+
+datmatpcc = np.array(datmatpcc)
+datmatpcc = (1 - datmatpcc) / 2
+datmatpcc_morph = np.array(datmatpcc_morph)
+datmatpcc_morph = (1 - datmatpcc_morph) / 2
+
+import matplotlib.pyplot as plt
+import pandas as pd
+import seaborn as sns
+
+for idx, df in enumerate(
+    [datmatrw, datmatrw_morph, datmatpcc, datmatpcc_morph]
+):
+    dm = pd.DataFrame(df, index=labels, columns=labels)
+
+    mask = np.triu(
+        np.ones_like(np.array(dm), dtype=bool)
+    )  # make a lower triangular mask
+    # Plot the heatmap with the mask
+    plt.figure(figsize=(15, 12))
+    sns.heatmap(dm, annot=False, vmin=0, vmax=1)
+    plt.grid(False)
+    plt.tight_layout()
+
+    iname = None
+    if idx == 0:
+        iname = "Rw"
+    elif idx == 1:
+        iname = "Rw-morphed"
+    elif idx == 2:
+        iname = "Pcc"
+    elif idx == 3:
+        iname = "Pcc-morphed"
+    dm.to_csv(Path("figs/CSVs") / f"{grname}-{iname}.csv")
+    plt.savefig(Path("figs") / f"{grname}-{iname}.png", dpi=300)
+    plt.savefig(Path("figs") / f"{grname}-{iname}.svg", dpi=300)
+
+    plt.show()

src/diffpy/similarity/compare_pdfs_single.py

@@ -0,0 +1,94 @@
+from pathlib import Path
+
+import numpy as np
+
+from diffpy.pdfmorph.pdfmorphpy import pdfmorph
+from diffpy.utils.parsers import loadData
+
+grname = "Si_O_only_10_common_grs"
+grdir = Path(f"PDFs/{grname}")
+
+gr_list = list(grdir.iterdir())
+gr_list.sort()
+
+datmatrw = []
+datmatrw_morph = []
+
+datmatpcc = []
+datmatpcc_morph = []
+
+labels = []
+
+for grfile1 in gr_list:
+    with open(grfile1, "r") as grread:
+        for line in grread:
+            if "composition =" in line:
+                labels.append(line.split("=")[1].strip())
+
+    datrw_row = []
+    datpcc_row = []
+
+    datrw_row_morph = []
+    datpcc_row_morph = []
+
+    for grfile2 in gr_list:
+        rawdata = pdfmorph(grfile1, grfile2, noplot="")[0]
+        morphdata = pdfmorph(grfile1, grfile2, noplot="", stretch=0, scale=1)[
+            0
+        ]
+
+        datrw_row.append(rawdata["Rw"])
+        datpcc_row.append(rawdata["Pearson"])
+
+        datrw_row_morph.append(morphdata["Rw"])
+        datpcc_row_morph.append(morphdata["Pearson"])
+
+    datmatrw.append(datrw_row)
+    datmatrw_morph.append(datrw_row_morph)
+
+    datmatpcc.append(datpcc_row)
+    datmatpcc_morph.append(datpcc_row_morph)
+
+datmatrw = np.array(datmatrw)
+datmatrw = np.sqrt(datmatrw * datmatrw.T)
+
+datmatrw_morph = np.array(datmatrw_morph)
+datmatrw_morph = np.sqrt(datmatrw_morph * datmatrw_morph.T)
+
+datmatpcc = np.array(datmatpcc)
+datmatpcc = (1 - datmatpcc) / 2
+datmatpcc_morph = np.array(datmatpcc_morph)
+datmatpcc_morph = (1 - datmatpcc_morph) / 2
+
+import matplotlib.pyplot as plt
+import pandas as pd
+import seaborn as sns
+
+for idx, df in enumerate(
+    [datmatrw, datmatrw_morph, datmatpcc, datmatpcc_morph]
+):
+    dm = pd.DataFrame(df, index=labels, columns=labels)
+
+    mask = np.triu(
+        np.ones_like(np.array(dm), dtype=bool)
+    )  # make a lower triangular mask
+    # Plot the heatmap with the mask
+    plt.figure(figsize=(10, 8))
+    sns.heatmap(dm, annot=True, vmin=0, vmax=1)
+    plt.grid(False)
+    plt.tight_layout()
+
+    iname = None
+    if idx == 0:
+        iname = "Rw"
+    elif idx == 1:
+        iname = "Rw-morphed"
+    elif idx == 2:
+        iname = "Pcc"
+    elif idx == 3:
+        iname = "Pcc-morphed"
+    dm.to_csv(Path("figs/CSVs") / f"{grname}-{iname}.csv")
+    plt.savefig(Path("figs") / f"{grname}-{iname}.png", dpi=300)
+    plt.savefig(Path("figs") / f"{grname}-{iname}.svg", dpi=300)
+
+    plt.show()

src/diffpy/similarity/convert_cif_pdf.py

@@ -0,0 +1,47 @@
+from pathlib import Path
+
+from diffpy.srreal.pdfcalculator import PDFCalculator
+from diffpy.structure import Structure
+
+Uisodefault = 0.005
+
+typename = "Zr_O_only"
+cifdir = Path(f"CIFs/{typename}")
+
+import numpy as np
+from pyobjcryst import loadCrystal
+
+cfg = {
+    "qmax": 25,
+    "rmin": 0,
+    "rmax": 30,
+}
+
+pc0 = PDFCalculator(**cfg)
+
+pdfdir = Path(f"PDFs/{typename}_grs")
+
+for ciffile in cifdir.iterdir():
+    crystal = loadCrystal(ciffile)
+    for sc in crystal.GetScatteringComponentList():
+        sp = sc.mpScattPow
+        sp.Biso = sp.Biso or 8 * np.pi**2 * Uisodefault
+
+    r0, g0 = pc0(crystal)
+
+    crystalcomp = None
+    with open(ciffile, "r") as cifread:
+        for line in cifread:
+            if "_chemical_formula_structural" in line:
+                crystalcomp = line.split()[1]
+
+    with open(f"PDFs/{typename}_grs/{ciffile.stem}.cgr", "w") as grfile:
+        grfile.write("i# [PDF] Computed by diffpy-cmi\n")
+        grfile.write(f"composition = {crystalcomp}\n")
+        grfile.write(f"qmax = {cfg['qmax']}\n")
+        grfile.write(f"rmin = {cfg['rmin']}\n")
+        grfile.write(f"rmax = {cfg['rmax']}\n")
+
+        grfile.write("\n")
+        grfile.write(f"# Labels: r, gr\n")
+        np.savetxt(grfile, np.array([r0, g0]).T)