dclab 0.64.1__tar.gz → 0.64.3__tar.gz

{dclab-0.64.1 → dclab-0.64.3}/CHANGELOG

@@ -1,3 +1,13 @@
+0.64.3
+ - ci: use GitHub windows-latest runner for building wheel
+0.64.2
+ - fix: string comparison TypeError when verifying basins, in niche cases
+   when basin's run identifier was not specified or could not be computed
+ - enh: raise NotImplementedError when subsetting basin contour data
+ - enh: make measurement identifier computation robust against empty strings
+ - ref: remove zombie code for fetching ML model label from model instance
+ - ref: minor cleanup in feature definitions
+ - docs: minor cleanup
 0.64.1
  - enh: introduce `KernelDensityEstimator.get_at`
  - enh: add 10th, 25th, 75th, and 90th percentile statistics methods

{dclab-0.64.1 → dclab-0.64.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dclab
-Version: 0.64.1
+Version: 0.64.3
 Summary: Library for real-time deformability cytometry (RT-DC)
 Author: Benedikt Hartmann, Eoghan O'Connell, Maik Herbig, Maximilian Schlögel, Nadia Sbaa, Paul Müller, Philipp Rosendahl, Raghava Alajangi
 Maintainer-email: Paul Müller <dev@craban.de>

{dclab-0.64.1 → dclab-0.64.3}/dclab/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '0.64.1'
-__version_tuple__ = version_tuple = (0, 64, 1)
+__version__ = version = '0.64.3'
+__version_tuple__ = version_tuple = (0, 64, 3)

{dclab-0.64.1 → dclab-0.64.3}/dclab/definitions/feat_logic.py

@@ -1,6 +1,11 @@
+import re
+
 from . import feat_const
 
 
+ML_SCORE_REGEX = re.compile(r"^ml_score_[a-z0-9]{3}$")
+
+
 def check_feature_shape(name, data):
     """Check if (non)-scalar feature matches with its data's dimensionality
 
@@ -17,11 +22,15 @@ def check_feature_shape(name, data):
         If the data's shape does not match its scalar description
     """
     if len(data.shape) == 1 and not scalar_feature_exists(name):
-        raise ValueError(f"Feature '{name}' is not a scalar feature, but "
-                         "a 1D array was given for `data`!")
+        raise ValueError(
+            f"Feature '{name}' is not a scalar feature, but "
+            "a 1D array was given for `data`!"
+        )
     elif len(data.shape) != 1 and scalar_feature_exists(name):
-        raise ValueError(f"Feature '{name}' is a scalar feature, but the "
-                         "`data` array is not 1D!")
+        raise ValueError(
+            f"Feature '{name}' is a scalar feature, but the "
+            "`data` array is not 1D!"
+        )
 
 
 def feature_exists(name, scalar_only=False):
@@ -56,15 +65,9 @@ def feature_exists(name, scalar_only=False):
     elif not scalar_only and name in feat_const.feature_names:
         # non-scalar feature
         valid = True
-    else:
-        # check whether we have an `ml_score_???` feature
-        valid_chars = "0123456789abcdefghijklmnopqrstuvwxyz"
-        if (name.startswith("ml_score_")
-            and len(name) == len("ml_score_???")
-            and name[-3] in valid_chars
-            and name[-2] in valid_chars
-                and name[-1] in valid_chars):
-            valid = True
+    elif ML_SCORE_REGEX.match(name):
+        # machine-learning score feature ml_score_???
+        valid = True
     return valid
 
 
@@ -93,8 +96,10 @@ def feature_register(name, label=None, is_scalar=True):
     allowed_chars = "abcdefghijklmnopqrstuvwxyz_1234567890"
     feat = "".join([f for f in name if f in allowed_chars])
     if feat != name:
-        raise ValueError("`feature` must only contain lower-case characters, "
-                         f"digits, and underscores; got '{name}'!")
+        raise ValueError(
+            "`feature` must only contain lower-case characters, "
+            f"digits, and underscores; got '{name}'!"
+        )
     if label is None:
         label = f"User-defined feature {name}"
     if feature_exists(name):
@@ -156,22 +161,16 @@ def get_feature_label(name, rtdc_ds=None, with_unit=True):
     TODO: extract feature label from ancillary information when an rtdc_ds is
     given.
     """
-    # TODO: Is there another way of avoiding this circular import?
-    from ..rtdc_dataset.feat_anc_core.ancillary_feature import AncillaryFeature
-    assert feature_exists(name)
     if name in feat_const.feature_name2label:
         label = feat_const.feature_name2label[name]
+    elif ML_SCORE_REGEX.match(name):
+        # use a generic name for machine-learning features
+        label = f"ML score {name[-3:].upper()}"
     else:
-        # First check whether an ancillary feature with that
-        # name exists.
-        for af in AncillaryFeature.features:
-            if af.feature_name == name:
-                labelid = af.data.outputs.index(name)
-                label = af.data.output_labels[labelid]
-                break
-        else:
-            # If that did not work, use a generic name.
-            label = "ML score {}".format(name[-3:].upper())
+        exists = feature_exists(name)
+        msg = f"Could not find label for '{name}'"
+        msg += " (feature does not exist)" if not exists else ""
+        raise ValueError(msg)
     if not with_unit:
         if label.endswith("]") and label.count("["):
             label = label.rsplit("[", 1)[0].strip()

{dclab-0.64.1 → dclab-0.64.3}/dclab/rtdc_dataset/core.py

@@ -835,9 +835,9 @@ class RTDCBase(abc.ABC):
         identifier = self.config.get("experiment", {}).get("run identifier",
                                                            None)
         if identifier is None:
-            time = self.config.get("experiment", {}).get("time", None)
-            date = self.config.get("experiment", {}).get("date", None)
-            sid = self.config.get("setup", {}).get("identifier", None)
+            time = self.config.get("experiment", {}).get("time", None) or None
+            date = self.config.get("experiment", {}).get("date", None) or None
+            sid = self.config.get("setup", {}).get("identifier", None) or None
             if None not in [time, date, sid]:
                 # only compute an identifier if all of the above are defined.
                 hasher = hashlib.md5(f"{time}_{date}_{sid}".encode("utf-8"))

{dclab-0.64.1 → dclab-0.64.3}/dclab/rtdc_dataset/feat_basin.py

@@ -458,22 +458,30 @@ class Basin(abc.ABC):
             if not self._measurement_identifier_verified:
                 if self.measurement_identifier is None:
                     # No measurement identifier was presented by the
-                    # referencing dataset. Don't perform any checks.
+                    # referencing dataset. We are in the dark.
+                    # Don't perform any checks.
                     self._measurement_identifier_verified = True
                 else:
-                    if self.mapping == "same":
-                        # When we have identical mapping, then the measurement
-                        # identifier has to match exactly.
-                        verifier = str.__eq__
+                    # This is the measurement identifier of the basin.
+                    basin_identifier = self.get_measurement_identifier()
+                    if basin_identifier is None:
+                        # Again, we are in the dark, because the basin dataset
+                        # does not have an identifier. This is an undesirable
+                        # situation, but there is nothing we can do about it.
+                        self._measurement_identifier_verified = True
                     else:
-                        # When we have non-identical mapping (e.g. exported
-                        # data), then the measurement identifier has to
-                        # partially match.
-                        verifier = str.startswith
-                    self._measurement_identifier_verified = verifier(
-                        self.measurement_identifier,
-                        self.get_measurement_identifier()
-                    )
+                        if self.mapping == "same":
+                            # When we have identical mapping, then the
+                            # measurement identifier has to match exactly.
+                            verifier = str.__eq__
+                        else:
+                            # When we have non-identical mapping (e.g. exported
+                            # data), then the measurement identifier has to
+                            # partially match.
+                            verifier = str.startswith
+                        self._measurement_identifier_verified = verifier(
+                            self.measurement_identifier, basin_identifier)
+
             check_rid = self._measurement_identifier_verified
         else:
             check_rid = True
@@ -538,8 +546,12 @@ class BasinProxy:
 
     def __getitem__(self, feat):
         if feat not in self._features:
-            feat_obj = BasinProxyFeature(feat_obj=self.ds[feat],
-                                         basinmap=self.basinmap)
+            if feat == "contour":
+                raise NotImplementedError("Feature 'contour' cannot be "
+                                          "handled by BasinProxy.")
+            else:
+                feat_obj = BasinProxyFeature(feat_obj=self.ds[feat],
+                                             basinmap=self.basinmap)
             self._features[feat] = feat_obj
         return self._features[feat]
 

{dclab-0.64.1 → dclab-0.64.3}/dclab.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dclab
-Version: 0.64.1
+Version: 0.64.3
 Summary: Library for real-time deformability cytometry (RT-DC)
 Author: Benedikt Hartmann, Eoghan O'Connell, Maik Herbig, Maximilian Schlögel, Nadia Sbaa, Paul Müller, Philipp Rosendahl, Raghava Alajangi
 Maintainer-email: Paul Müller <dev@craban.de>

{dclab-0.64.1 → dclab-0.64.3}/dclab.egg-info/SOURCES.txt

@@ -246,7 +246,8 @@ tests/test_cli_join.py
 tests/test_cli_repack.py
 tests/test_cli_split.py
 tests/test_config_value_mapping.py
-tests/test_dfn.py
+tests/test_dfn_feat.py
+tests/test_dfn_meta.py
 tests/test_downsampling.py
 tests/test_feat_bright.py
 tests/test_feat_bright_bc.py

{dclab-0.64.1 → dclab-0.64.3}/docs/sec_av_basins/index.rst

@@ -432,7 +432,7 @@ are a few things to notice:
   This relative path makes it possible to copy-paste these two files *together* to
   other locations. You will always be able to open the output file and see the
   basin features defined in the input file. Internally, dclab also checks
-  the :func:`measurement identifier <.RTDCBase.get_measurement_identifier>`
+  the :meth:`run identifier <dclab.rtdc_dataset.RTDCBase.get_measurement_identifier>`
   of the output file against that of the input file to avoid loading basin
   features from the wrong file.
 

{dclab-0.64.1 → dclab-0.64.3}/docs/sec_code_reference.rst

@@ -304,8 +304,8 @@ isoelastics
 
 .. _sec_ref_kde:
 
-kernel density estimators (KDEs)
--------------------------------
+Kernel Density Estimators (KDEs)
+--------------------------------
 .. autoclass:: dclab.kde.KernelDensityEstimator
    :members:
 

dclab-0.64.3/tests/test_dfn_feat.py

@@ -0,0 +1,27 @@
+import pytest
+
+import dclab
+from dclab.definitions import get_feature_label
+
+
+@pytest.mark.parametrize("name,label",
+                         [
+                             ["ml_score_rbc", "ML score RBC"],
+                             ["deform", "Deformation"],
+                             ["area_um", "Area [µm²]"],
+                         ])
+def test_get_feature_label_basic(name, label):
+    assert get_feature_label(name) == label
+
+
+def test_get_feature_label_temporary():
+    name = "unknown"
+    label = "User-defined feature unknown"
+    dclab.register_temporary_feature(name)
+    assert get_feature_label(name) == label
+
+
+def test_get_feature_label_invalid():
+    name = "unknown2"
+    with pytest.raises(ValueError, match="unknown2"):
+        get_feature_label(name)

{dclab-0.64.1 → dclab-0.64.3}/tests/test_rtdc_feat_basin.py

@@ -418,6 +418,48 @@ def test_basin_key_reproducible():
         ]
 
 
+def test_basin_run_identifier_unknown():
+    """
+    When the run identifier of a basin is unknown, the basin
+    should still be accepted.
+    """
+    h5path = retrieve_data("fmt-hdf5_fl_wide-channel_2023.zip")
+    with h5py.File(h5path, "a") as h5:
+        # remove run identifier and metadata from which run identifier
+        # can be computed.
+        for attr in [
+            "experiment:run identifier",
+            "experiment:time",
+            "experiment:date",
+            "setup:identifier",
+        ]:
+            if attr in h5.attrs:
+                del h5.attrs[attr]
+            assert attr not in h5.attrs
+
+    export_same = h5path.with_name("same.rtdc")
+    export_subs = h5path.with_name("subs.rtdc")
+
+    # Check whether that worked and export a subset with basins.
+    with dclab.new_dataset(h5path) as ds:
+        assert ds.get_measurement_identifier() is None
+        # export without filters
+        ds.export.hdf5(export_same, features=["deform"], basins=True)
+        # export with filters
+        ds.filter.manual[::2] = False
+        ds.apply_filter()
+        ds.export.hdf5(export_subs, features=["deform"], basins=True)
+
+    # Try to open the exported datasets (this raised an error before)
+    with dclab.new_dataset(export_same) as ds:
+        assert "image" in ds
+        assert ds["image"][0][0, 0] == 124
+
+    with dclab.new_dataset(export_subs) as ds:
+        assert "image" in ds
+        assert ds["image"][0][0, 0] == 126
+
+
 def test_basin_sorting_basic():
     bnlist = [
         {"type": "remote", "format": "dcor", "ident": 0},

{dclab-0.64.1 → dclab-0.64.3}/tests/test_rtdc_fmt_http_basin.py

@@ -258,7 +258,7 @@ def test_trace_availability_invalid(tmp_path):
         bdat = {
             "type": "remote",
             "format": "http",
-            "urls": ["https://dcor.mpl.mpg.de/api/3/action/site_read"],
+            "urls": ["https://dcor.mpl.mpg.de/api/3/action/status_show"],
             "features": ["trace"],
         }
         blines = json.dumps(bdat, indent=2).split("\n")