dcnum 0.25.9__tar.gz → 0.25.10__tar.gz

{dcnum-0.25.9 → dcnum-0.25.10}/CHANGELOG

@@ -1,3 +1,6 @@
+0.25.10
+ - fix: search for validation strings in logs did not use re.MULTILINE
+ - enh: allow to validate torch model applicability based on meta values
 0.25.9
  - fix: md5 sum caused PermissionError on Windows for background computation
 0.25.8

{dcnum-0.25.9 → dcnum-0.25.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dcnum
-Version: 0.25.9
+Version: 0.25.10
 Summary: numerics toolbox for imaging deformability cytometry
 Author: Maximilian Schlögel, Paul Müller, Raghava Alajangi
 Maintainer-email: Paul Müller <dev@craban.de>

{dcnum-0.25.9 → dcnum-0.25.10}/src/dcnum/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '0.25.9'
-__version_tuple__ = version_tuple = (0, 25, 9)
+__version__ = version = '0.25.10'
+__version_tuple__ = version_tuple = (0, 25, 10)

{dcnum-0.25.9 → dcnum-0.25.10}/src/dcnum/segm/segm_torch/segm_torch_base.py

@@ -3,6 +3,8 @@ import pathlib
 import re
 from typing import Dict
 
+import numpy as np
+
 from ...meta import paths
 
 from ..segmenter import Segmenter, SegmenterNotApplicableError
@@ -116,10 +118,15 @@ class TorchSegmenterBase(Segmenter):
         """
         key = item["key"]
         if key in data_dict:
-            regexp = re.compile(item["regexp"])
-            matched = bool(regexp.match(data_dict[key]))
-            negate = item.get("regexp-negate", False)
-            valid = matched if not negate else not matched
+            valid = True
+            if "regexp" in item:
+                re_match = bool(re.search(item["regexp"], data_dict[key],
+                                          re.MULTILINE))
+                negate = item.get("regexp-negate", False)
+                valid = valid and (re_match if not negate else not re_match)
+            if "value" in item:
+                valid = valid and np.allclose(item["value"], data_dict[key],
+                                              atol=0, rtol=0.01)
             if not valid:
                 reasons_list.append(item.get("reason", "unknown reason"))
         elif not item.get("allow-missing-key", False):

{dcnum-0.25.9 → dcnum-0.25.10}/src/dcnum.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dcnum
-Version: 0.25.9
+Version: 0.25.10
 Summary: numerics toolbox for imaging deformability cytometry
 Author: Maximilian Schlögel, Paul Müller, Raghava Alajangi
 Maintainer-email: Paul Müller <dev@craban.de>

{dcnum-0.25.9 → dcnum-0.25.10}/src/dcnum.egg-info/SOURCES.txt

@@ -125,4 +125,5 @@ tests/data/fmt-hdf5_cytoshot_full-features_legacy_allev_2023.zip
 tests/data/fmt-hdf5_shapein_empty.zip
 tests/data/fmt-hdf5_shapein_raw-with-variable-length-logs.zip
 tests/data/segm-torch-model_unet-dcnum-test_g1_910c2.zip
+tests/data/segm-torch-model_unet-dcnum-test_g2_17ec6.zip
 tests/data/segm-torch-test-data_unet-dcnum-test_g1_910c2.zip

{dcnum-0.25.9 → dcnum-0.25.10}/tests/test_segm_torch.py

@@ -54,6 +54,43 @@ def test_segm_torch_validate_model_file_logs():
             )
 
 
+def test_segm_torch_validate_model_file_logs_negate():
+    """Test whether model validation fails for invalid logs"""
+    model_file = retrieve_model(
+        "segm-torch-model_unet-dcnum-test_g2_17ec6.zip")
+    sm = segm.segm_torch.SegmentTorchMPO
+
+    # Creating a specific log file will mak the model invalid
+    path = retrieve_data(
+        "fmt-hdf5_cytoshot_full-features_2024.zip")
+
+    with read.HDF5Data(path) as hd:
+        # sanity check
+        assert "dclab-compress" in hd.logs
+        with pytest.raises(
+                segm_torch_base.SegmenterNotApplicableError,
+                match="must not be compressed 2024-05-07"):
+            sm.validate_applicability(
+                segmenter_kwargs={"model_file": model_file},
+                meta=hd.meta,
+                logs=hd.logs
+            )
+
+    # Remove the offending log
+    with h5py.File(path, "a") as h5:
+        del h5["logs/dclab-compress"]
+
+    # Try again, this should work now.
+    with read.HDF5Data(path) as hd:
+        # sanity check
+        assert "dclab-compress" not in hd.logs
+        sm.validate_applicability(
+            segmenter_kwargs={"model_file": model_file},
+            meta=hd.meta,
+            logs=hd.logs
+        )
+
+
 def test_segm_torch_validate_model_file_meta():
     """Test whether model validation fails for invalid metadata"""
     model_file = retrieve_model(
@@ -96,6 +133,45 @@ def test_segm_torch_validate_model_file_meta():
             )
 
 
+def test_segm_torch_validate_model_file_meta_value():
+    """Test whether model validation fails for invalid metadata"""
+    model_file = retrieve_model(
+        "segm-torch-model_unet-dcnum-test_g2_17ec6.zip")
+    sm = segm.segm_torch.SegmentTorchMPO
+
+    # Create a test dataset with metadata that will make the model invalid
+    path = retrieve_data(
+        "fmt-hdf5_cytoshot_full-features_2023.zip")
+
+    with h5py.File(path, "a") as h5:
+        h5.attrs["setup:channel width"] = 30.
+
+    with read.HDF5Data(path) as hd:
+        # sanity check
+        assert hd.meta["setup:channel width"] == 30
+        with pytest.raises(
+                segm_torch_base.SegmenterNotApplicableError,
+                match="channel width must be 20 micrometers"):
+            sm.validate_applicability(
+                segmenter_kwargs={"model_file": model_file},
+                meta=hd.meta,
+                logs=hd.logs
+            )
+
+    # Repeat the same thing, this time fixing the attribute
+    with h5py.File(path, "a") as h5:
+        h5.attrs["setup:channel width"] = 20.
+
+    with read.HDF5Data(path) as hd:
+        # sanity check
+        assert hd.meta["setup:channel width"] == 20
+        sm.validate_applicability(
+            segmenter_kwargs={"model_file": model_file},
+            meta=hd.meta,
+            logs=hd.logs
+        )
+
+
 def test_segm_torch_mpo():
     """Basic PyTorch segmenter"""
     path = retrieve_data(