Enable RTDC copier to accept list of features (#247)

* enh: enable selected features to be copied

* test: copy a list of features with rtdc_copy

* Update CHANGELOG

* reg: iterate over feature list

* Include negative control test

---------

Co-authored-by: ralajan <raghava.alajangi@mpl.mpg.de>
Co-authored-by: Paul Müller <paulmueller@users.noreply.github.com>

CHANGELOG

@@ -1,3 +1,5 @@
+0.57.2
+ - enh: allow to specify list of features for data copier
 0.57.1
  - fix: RTDCWriter.rectify_metadata fails when image feature is empty
  - fix: handle empty write requests in export.hdf5 and RTDCWriter (#242)

dclab/rtdc_dataset/copier.py

@@ -1,7 +1,7 @@
 """Helper methods for copying .rtdc data"""
 from __future__ import annotations
 
-from typing import Literal
+from typing import List, Literal
 
 import h5py
 import h5py.h5o
@@ -15,11 +15,30 @@
 
 def rtdc_copy(src_h5file: h5py.Group,
               dst_h5file: h5py.Group,
-              features: Literal['all', 'scalar', 'none'] = "all",
+              features: List[str] | Literal['all', 'scalar', 'none'] = "all",
               include_logs: bool = True,
               include_tables: bool = True,
               meta_prefix: str = ""):
-    """Create a compressed copy of an RT-DC file"""
+    """Create a compressed copy of an RT-DC file
+
+    Parameters
+    ----------
+    src_h5file: h5py.Group
+        Input HDF5 file
+    dst_h5file: h5py.Group
+        Output HDF5 file
+    features: list of strings or one of ['all', 'scalar', 'none']
+        If this is a list then it specifies the features that are copied from
+        `src_h5file` to `dst_h5file`. Alternatively, you may specify 'all'
+        (copy all features), 'scalar' (copy only scalar features), or 'none'
+        (don't copy any features).
+    include_logs: bool
+        Copy the logs from `src_h5file` to `dst_h5file`.
+    include_tables: bool
+        Copy the tables from `src_h5file` to `dst_h5file`.
+    meta_prefix: str
+        Add this prefix to the name of the logs and tables in `dst_h5file`.
+    """
     # metadata
     for akey in src_h5file.attrs:
         dst_h5file.attrs[akey] = src_h5file.attrs[akey]
@@ -55,11 +74,24 @@ def rtdc_copy(src_h5file: h5py.Group,
                 **hdf5plugin.Zstd(clevel=5))
 
     # events
-    if features != "none":
-        scalar_only = features == "scalar"
+    if isinstance(features, list):
+        feature_iter = features
+    elif features == "all":
+        feature_iter = list(src_h5file["events"])
+    elif features == "scalar":
+        feature_iter = [feat for feat in src_h5file["events"]
+                        if feature_exists(feat, scalar_only=True)]
+    elif features == "none":
+        feature_iter = []
+    else:
+        raise ValueError(f"`features` must be either a list of feature names "
+                         f"or one of 'all', 'scalar' or 'none', got "
+                         f"'{features}'")
+
+    if feature_iter:
         dst_h5file.require_group("events")
-        for feat in src_h5file["events"]:
-            if feature_exists(feat, scalar_only=scalar_only):
+        for feat in feature_iter:
+            if feature_exists(feat):
                 # Skip all defective features. These are features that
                 # are known to be invalid (e.g. ancillary features that
                 # were computed falsely) and must be recomputed by dclab.

tests/test_rtdc_copier.py

@@ -113,7 +113,7 @@ def test_copy_tables():
     tab_data = np.zeros((10, len(columns)))
     tab_data[:, 0] = np.arange(10)
     tab_data[:, 1] = 1000
-    tab_data[:, 2] = np.linspace(np.pi, 2*np.pi, 10)
+    tab_data[:, 2] = np.linspace(np.pi, 2 * np.pi, 10)
     rec_arr = np.rec.array(tab_data, dtype=ds_dt)
     # sanity check
     assert np.all(rec_arr["bread"][:].flatten() == np.arange(10))
@@ -140,7 +140,7 @@ def test_copy_tables():
         assert np.all(tab_data["bread"][:].flatten() == np.arange(10))
         assert np.all(tab_data["beer"][:].flatten() == 1000)
         assert np.all(tab_data["chocolate"][:].flatten() == np.linspace(
-            np.pi, 2*np.pi, 10))
+            np.pi, 2 * np.pi, 10))
 
 
 def test_copy_tables_hdf5_issue_3214():
@@ -209,3 +209,24 @@ def test_copy_scalar_features_only():
     with h5py.File(path_copy) as hc:
         assert "image" not in hc["events"]
         assert "deform" in hc["events"]
+
+
+def test_copy_specified_feature_list():
+    path = retrieve_data("fmt-hdf5_image-bg_2020.zip")
+    path_copy = path.with_name("test_copy.rtdc")
+
+    # copy
+    with h5py.File(path) as h5, h5py.File(path_copy, "w") as hc:
+        # make sure image data is there
+        assert "image" in h5["events"]
+        assert "area_um" in h5["events"]
+        assert "deform" in h5["events"]
+        rtdc_copy(src_h5file=h5,
+                  dst_h5file=hc,
+                  features=["image", "deform"])
+
+    # Make sure this worked
+    with h5py.File(path_copy) as hc:
+        assert "image" in hc["events"]
+        assert "area_um" not in hc["events"]
+        assert "deform" in hc["events"]