add phasics zip tif group file format

qpformat/file_formats/__init__.py

@@ -3,7 +3,7 @@
 
 from .dataset import DataSet
 # from .group_hdf5_qpimage import GroupHdf5Qpimage
-# from .group_zip_tif_phasics import GroupZipTifPhasics
+from .group_zip_tif_phasics import GroupZipTifPhasics
 from .single_hdf5_qimage import SingleHdf5Qpimage
 from .single_tif_phasics import SingleTifPhasics
 
@@ -88,7 +88,7 @@ class UnknownFileFormatError(BaseException):
 
 # the order is important for
 formats = [GroupFolder,
-           # GroupZipTifPhasics,
+           GroupZipTifPhasics,
            # GroupHdf5Qpimage,
            SingleHdf5Qpimage,
            SingleTifPhasics,

qpformat/file_formats/group_zip_tif_phasics.py

@@ -0,0 +1,81 @@
+import io
+import functools
+import zipfile
+
+from .dataset import DataSet
+from .single_tif_phasics import SingleTifPhasics
+
+
+class GroupZipTifPhasics(DataSet):
+    def __init__(self, *args, **kwargs):
+        super(GroupZipTifPhasics, self).__init__(*args, **kwargs)
+        self._files = None
+        self._dataset = None
+
+    def __len__(self):
+        return len(self.files)
+
+    def _get_dataset(self, idx):
+        if self._dataset is None:
+            self._dataset = [None] * len(self)
+        if self._dataset[idx] is None:
+            # Use ``zipfile.ZipFile.open`` to return an open file
+            zf = zipfile.ZipFile(self.path)
+            pt = zf.open(self.files[idx])
+            fd = io.BytesIO(pt.read())
+            self._dataset[idx] = SingleTifPhasics(path=fd,
+                                                  meta_data=self.meta_data)
+        assert len(self._dataset[idx]) == 1, "unknown phasics tif file"
+        return self._dataset[idx]
+
+    @staticmethod
+    @functools.lru_cache(maxsize=32)
+    def _index_files(path):
+        """Search zip file for SID PHA files"""
+        with zipfile.ZipFile(path) as zf:
+            names = sorted(zf.namelist())
+            names = [nn for nn in names if nn.endswith(".tif")]
+            names = [nn for nn in names if nn.startswith("SID PHA")]
+            phasefiles = []
+            for name in names:
+                with zf.open(name) as pt:
+                    fd = io.BytesIO(pt.read())
+                    if SingleTifPhasics.verify(fd):
+                        phasefiles.append(name)
+            return phasefiles
+
+    @property
+    def files(self):
+        if self._files is None:
+            self._files = GroupZipTifPhasics._index_files(self.path)
+        return self._files
+
+    def get_qpimage_raw(self, idx=0):
+        """Return QPImage without background correction"""
+        ds = self._get_dataset(idx)
+        return ds.get_qpimage_raw(idx=0)
+
+    def get_time(self, idx=0):
+        ds = self._get_dataset(idx)
+        return ds.get_time(idx=0)
+
+    @staticmethod
+    def verify(path):
+        """Verify phasics zip tif file format"""
+        valid = False
+        try:
+            zf = zipfile.ZipFile(path)
+        except:
+            pass
+        else:
+            names = sorted(zf.namelist())
+            names = [nn for nn in names if nn.endswith(".tif")]
+            names = [nn for nn in names if nn.startswith("SID PHA")]
+            for name in names:
+                with zf.open(name) as pt:
+                    fd = io.BytesIO(pt.read())
+                    if SingleTifPhasics.verify(fd):
+                        valid = True
+                        break
+            zf.close()
+        return valid

qpformat/file_formats/single_tif_phasics.py

@@ -25,7 +25,7 @@ def __init__(self, path, meta_data={}):
         Parameters
         ----------
         path: str
-            path to the experimental data file.
+            path to the experimental data file or an open file object
         meta_data: dict
             dictionary containing meta data.
             see :py:class:`qpimage.VALID_META_KEYS`.
@@ -55,12 +55,20 @@ def __init__(self, path, meta_data={}):
     def __len__(self):
         return 1
 
+    @staticmethod
+    def _get_tif(path):
+        if not isinstance(path, str):
+            # Seek open file zero to avoid error in tifffile:
+            # "ValueError: invalid TIFF file"
+            path.seek(0)
+        return tifffile.TiffFile(path)
+
     def get_qpimage_raw(self, idx=0):
         """Return QPImage without background correction"""
         if idx != 0:
             raise ValueError("Single file format, only one entry (`idx!=0`)!")
         # Load experimental data
-        with tifffile.TiffFile(self.path) as tf:
+        with SingleTifPhasics._get_tif(self.path) as tf:
             # page 0 contains intensity
             # page 1 contains phase in nm
             # page 2 contains phase in wavelengths
@@ -114,7 +122,7 @@ def get_time(self, idx=0):
 
     @staticmethod
     def _get_meta_data(path, section, name):
-        with tifffile.TiffFile(path) as tf:
+        with SingleTifPhasics._get_tif(path) as tf:
             meta = str(tf.pages[0].tags["61238"].value)
 
         meta = meta.strip("'b")
@@ -142,18 +150,17 @@ def verify(path):
         Returns `True` if the file format matches.
         """
         valid = False
-        if path.endswith(".tif"):
-            try:
-                tf = tifffile.TiffFile(path)
-            except:
-                pass
-            else:
-                if (len(tf) == 3 and
-                    "61243" in tf.pages[0].tags and
-                    "61242" in tf.pages[0].tags and
-                    "61238" in tf.pages[0].tags and
-                    "max_sample_value" in tf.pages[0].tags and
-                    (tf.pages[0].tags["61242"].value !=
-                     tf.pages[1].tags["61242"].value)):
-                    valid = True
+        try:
+            tf = SingleTifPhasics._get_tif(path)
+        except:
+            pass
+        else:
+            if (len(tf) == 3 and
+                "61243" in tf.pages[0].tags and
+                "61242" in tf.pages[0].tags and
+                "61238" in tf.pages[0].tags and
+                "max_sample_value" in tf.pages[0].tags and
+                (tf.pages[0].tags["61242"].value !=
+                 tf.pages[1].tags["61242"].value)):
+                valid = True
         return valid

tests/test_group_zip_tif_phasics.py

@@ -0,0 +1,44 @@
+"""
+The test file "group_phasics.zip" was created from an original
+Phasics zip file. The "ConfigStack.txt" was adapted to reflect
+only three measured images. All data files that are not relevant
+for qpformat are included as empty files to reflect the original
+zip file structure. The "SID PHA*.tif" files were created with
+the script given in the doc string of "test_single_tif_phasics.py".
+"""
+from os.path import abspath, dirname, join
+import sys
+
+import numpy as np
+
+# Add parent directory to beginning of path variable
+sys.path.insert(0, dirname(dirname(abspath(__file__))))
+import qpformat  # noqa: E402
+
+
+def test_load_data():
+    path = join(dirname(abspath(__file__)), "data/group_phasics.zip")
+    ds = qpformat.load_data(path)
+    assert ds.path == path
+    assert "GroupZipTifPhasics" in ds.__repr__()
+
+
+def test_data_content():
+    path = join(dirname(abspath(__file__)), "data/group_phasics.zip")
+    ds = qpformat.load_data(path)
+    assert len(ds) == 3
+    assert ds.get_time(0) == 1461949418.29027
+    assert ds.get_time(1) == 1461949418.62727
+    assert ds.get_time(2) == 1461949419.11427
+    qpi = ds.get_qpimage(0)
+    assert qpi.meta["wavelength"] == 550e-9
+    assert np.allclose(qpi.amp.max(), 183.96992660669972)
+    assert np.allclose(qpi.pha.max() - qpi.pha.min(), 0.189023369599572)
+
+
+if __name__ == "__main__":
+    # Run all tests
+    loc = locals()
+    for key in list(loc.keys()):
+        if key.startswith("test_") and hasattr(loc[key], "__call__"):
+            loc[key]()