Merge 4601903 into 55144ce

Author: be-smith

pydatalab/example_data/XRD/JO_KL_16_ZF_3_60deg_0.02step_12degpermin_001.rasx

@@ -12,14 +12,14 @@
 from pydatalab.logger import LOGGER
 from pydatalab.mongo import flask_mongo
 
-from .utils import parse_xrdml
+from .utils import parse_rasx_zip, parse_xrdml
 
 
 class XRDBlock(DataBlock):
     blocktype = "xrd"
     name = "Powder XRD"
     description = "Visualize XRD patterns and perform simple baseline corrections."
-    accepted_file_extensions = (".xrdml", ".xy", ".dat", ".xye")
+    accepted_file_extensions = (".xrdml", ".xy", ".dat", ".xye", ".rasx")
 
     defaults = {"wavelength": 1.54060}
 
@@ -38,6 +38,8 @@ def load_pattern(
 
         if ext == ".xrdml":
             df = parse_xrdml(location)
+        elif ext == ".rasx":
+            df = parse_rasx_zip(location)
 
         else:
             columns = ["twotheta", "intensity", "error"]

pydatalab/src/pydatalab/apps/xrd/blocks.py

@@ -1,6 +1,8 @@
 import os
 import re
 import warnings
+import zipfile
+from pathlib import Path
 from typing import List, Tuple
 
 import numpy as np
@@ -146,3 +148,39 @@ def toXY(intensities: List[float], start: float, end: float) -> str:
     angles = np.linspace(start, end, num=len(intensities))
     xylines = ["{:.5f} {:.3f}\r\n".format(a, i) for a, i in zip(angles, intensities)]
     return "".join(xylines)
+
+
+def parse_rasx_zip(filename: str) -> pd.DataFrame:
+    """Parses an RASX zip file and returns a pandas DataFrame with columns
+    twotheta and intensity.
+
+    Parameters:
+        filename: The file to parse.
+
+    """
+    # Unzip the file
+    zip_path = Path(filename)
+    destination = zip_path.parent / "extracted"
+    with zipfile.ZipFile(zip_path, "r") as zip_ref:
+        zip_ref.extractall(destination)
+
+    # Find the .txt data file inside the unzipped .rasx archive
+    # Seems to normally unzip to a folder called "Data0" with one .txt file inside
+    data_file = None
+    for file in destination.glob("Data*/*.txt"):
+        data_file = file
+        break
+
+    if data_file is None:
+        raise FileNotFoundError("No .txt file found in the .rasx archive.")
+
+    xrd_data = pd.read_csv(data_file, sep="\t", header=None)
+    xrd_data.columns = ["twotheta", "intensity", "imnotsure"]
+    # Extract the data
+    # twotheta, intensities = getTwoThetaIntensities(s)
+    return pd.DataFrame(
+        {
+            "twotheta": xrd_data["twotheta"],
+            "intensity": xrd_data["intensity"],
+        }
+    )

pydatalab/src/pydatalab/apps/xrd/utils.py

@@ -13,19 +13,21 @@ def data_files():
 
 def test_load(data_files):
     for f in data_files:
-        df, y_options = XRDBlock.load_pattern(f)
-        assert all(y in df.columns for y in y_options)
+        if f.suffix in XRDBlock.accepted_file_extensions:
+            df, y_options = XRDBlock.load_pattern(f)
+            assert all(y in df.columns for y in y_options)
 
 
 def test_plot(data_files):
     f = next(data_files)
-    df, y_options = XRDBlock.load_pattern(f)
-    p = selectable_axes_plot(
-        [df],
-        x_options=["2θ (°)", "Q (Å⁻¹)", "d (Å)"],
-        y_options=y_options,
-        plot_line=True,
-        plot_points=True,
-        point_size=3,
-    )
-    assert p
+    if f.suffix in XRDBlock.accepted_file_extensions:
+        df, y_options = XRDBlock.load_pattern(f)
+        p = selectable_axes_plot(
+            [df],
+            x_options=["2θ (°)", "Q (Å⁻¹)", "d (Å)"],
+            y_options=y_options,
+            plot_line=True,
+            plot_points=True,
+            point_size=3,
+        )
+        assert p