Can set input dataset parameters

conf/testing.py

@@ -42,6 +42,16 @@ tifffile = {
     'z': 7,
 }
 
+filename = 'mono_zstack_mask.tif'
+monozstackmask = {
+    'filename': filename,
+    'path': root / filename,
+    'w': 256,
+    'h': 256,
+    'c': 1,
+    'z': 85
+}
+
 ilastik = {
     'pixel_classifier': 'demo_px.ilp',
     'object_classifier': 'demo_obj.ilp',

extensions/chaeo/examples/transfer_labels_to_ilastik_object_classifier.py

-import csv
 from pathlib import Path
 import h5py
-import ilastik.applets.objectClassification
+import json
 import numpy as np
 import pandas as pd
-import tifffile
+import uuid
 
-from extensions.ilastik.models import IlastikObjectClassifierModel
 from model_server.accessors import generate_file_accessor
 
 def get_dataset_info(h5):
@@ -14,11 +12,17 @@ def get_dataset_info(h5):
     info = {}
     for gk in ['Raw Data', 'Segmentation Image']:
         info[gk] = {}
-        for dk in ['location', 'filePath', 'shape']:
+        for dk in ['location', 'filePath', 'shape', 'nickname']:
             try:
                 info[gk][dk] = h5[f'{lane}/{gk}/{dk}'][()]
             except Exception as e:
                 print(e)
+        try:
+            info[gk]['id'] = uuid.UUID(h5[f'{lane}/{gk}/datasetId'][()].decode())
+        except ValueError as e:
+            info[gk]['id'] = '<invalid UUID>'
+        info[gk]['axistags'] = json.loads(h5[f'{lane}/{gk}/axistags'][()].decode())
+        info[gk]['axes'] = [ax['key'] for ax in info[gk]['axistags']['axes']]
     return info
 
 def transfer_labels_to_ilastik_ilp(ilp, df_stack_meta):
@@ -58,42 +62,55 @@ def transfer_labels_to_ilastik_ilp(ilp, df_stack_meta):
 
 def generate_ilastik_object_classifier(template_ilp, where_training: str):
 
-    # validate input data
+    # validate z-stack input data
     where = Path(where_training)
     zstacks = {
-        'raw': {
+        'Raw Data': {
             'path': where / 'zstack_train_raw.tif',
         },
-        'seg': {
+        'Segmentation Image': {
             'path': where / 'zstack_train_mask.tif',
         }
     }
 
     for k, v in zstacks.items():
-        # assert v['path'].exists(), 'Could not find input z-stack: ' + v['path']
-        # ff = tifffile.imread(v['path'])
-        # v['nz'] = ff.shape[0]
-        # v['hw'] = ff.shape[1:2]
-        # v['dtype'] = ff.dtype
         v['acc'] = generate_file_accessor(v['path'])
-    assert zstacks['raw']['acc'].is_binary()
 
-    assert zstacks['raw']
+    assert zstacks['Segmentation Image']['acc'].is_mask()
+
+    assert len(set([v['acc'].hw for k, v in zstacks.items()])) == 1  # same height and width
+    assert len(set([v['acc'].nz for k, v in zstacks.items()])) == 1  # same z-depth
+
+    # now load CSV
+    csv_path = where / 'train_stack.csv'
+    assert csv_path.exists()
+    df_meta = pd.read_csv(csv_path)
+    assert np.all(
+        df_meta['zi'].sort_values().to_numpy() == np.arange(0, zstacks['Raw Data']['acc'].nz)
+    )
 
     with h5py.File(template_ilp, 'r+') as h5:
-        pass
+        info = get_dataset_info(h5)
+
+        def set_ds(grp, ds, val):
+            ds = h5[f'Input Data/infos/lane0000/{grp}/{ds}']
+            ds[()] = val
+            return ds[()]
+
+        for hg in ['Raw Data', 'Segmentation Image']:
+            assert info[hg]['location'] == b'FileSystem'
+            assert info[hg]['axes'] == ['t', 'y', 'x']
+            set_ds(hg, 'filePath', zstacks[hg]['path'].__str__())
+            set_ds(hg, 'nickname', zstacks[hg]['path'].stem)
+            shape_zyx = [zstacks[hg]['acc'].shape_dict[ax] for ax in ['Z', 'Y', 'X']]
+            set_ds(hg, 'shape', np.array(shape_zyx))
+        new_info = get_dataset_info(h5)
 
 
 if __name__ == '__main__':
     ilp = 'c:/Users/rhodes/model-server/ilastik/test_autolabel_obj - Copy.ilp'
-    # ilp = 'c:/Users/rhodes/model-server/ilastik/test_template_obj.ilp'
-
-    df = pd.read_csv(
-        'c:/Users/rhodes/projects/proj0011-plankton-seg/exp0009/output/labeled_patches-20231014-0002/train_stack.csv'
-    )
-    # transfer_labels_to_ilastik_ilp(ilp, df)
 
     generate_ilastik_object_classifier(
-        ilp,
-        'c:/Users/rhodes/projects/proj0011-plankton-seg/exp0009/output/labeled_patches-20231014-0002'
+        'c:/Users/rhodes/model-server/ilastik/test_template_obj.ilp',
+        'c:/Users/rhodes/projects/proj0011-plankton-seg/exp0009/output/labeled_patches-20231014-0004'
     )
\ No newline at end of file

extensions/chaeo/workflows.py

@@ -250,6 +250,7 @@ def transfer_ecotaxa_labels_to_patch_stacks(
 
             stack_meta.append({'zi': fi, 'patch_filename': fn, 'annotation_class': ac, 'annotation_class_id': aci})
             acc_bm = generate_file_accessor(Path(where_masks) / fn)
+            assert acc_bm.is_mask()
             assert acc_bm.hw == patch_size, f'Unexpected patch size {patch_size}'
             assert acc_bm.chroma == 1
             assert acc_bm.nz == 1

model_server/accessors.py

@@ -34,7 +34,11 @@ class GenericImageDataAccessor(ABC):
         return True if self.shape_dict['Z'] > 1 else False
 
     def is_mask(self):
-        return self._data.dtype == 'bool'
+        if self._data.dtype == 'bool':
+            return True
+        elif self._data.dtype == 'uint8':
+            return np.all(np.unique(self._data) == [0, 255])
+        return False
 
     def get_one_channel_data (self, channel: int):
         c = int(channel)
@@ -101,12 +105,18 @@ class TifSingleSeriesFileAccessor(GenericImageFileAccessor):
             raise DataShapeError(f'Expect only one series in {fpath}')
 
         se = tf.series[0]
-        sd = {ch: se.shape[se.axes.index(ch)] for ch in se.axes}
 
-        idx = {k: sd[k] for k in ['Y', 'X', 'C', 'Z']}
+        order = ['Y', 'X', 'C', 'Z']
+        axs = [a for a in se.axes if a in order]
+        da = se.asarray()
+
+        if 'C' not in axs:
+            axs.append('C')
+            da = np.expand_dims(da, len(da.shape))
+
         yxcz = np.moveaxis(
-            se.asarray(),
-            [se.axes.index(ch) for ch in idx],
+            da,
+            [axs.index(k) for k in order],
             [0, 1, 2, 3]
         )
 

tests/test_accessors.py

@@ -2,7 +2,7 @@ import unittest
 
 import numpy as np
 
-from conf.testing import czifile, output_path, monopngfile, rgbpngfile, tifffile
+from conf.testing import czifile, output_path, monopngfile, rgbpngfile, tifffile, monozstackmask
 from model_server.accessors import CziImageFileAccessor, DataShapeError, generate_file_accessor, InMemoryDataAccessor, PngFileAccessor, write_accessor_data_to_file, TifSingleSeriesFileAccessor
 
 class TestCziImageFileAccess(unittest.TestCase):
@@ -106,4 +106,8 @@ class TestCziImageFileAccess(unittest.TestCase):
         self.assertEqual(acc.nz, 1)
 
     def test_read_mono_png(self):
-        return self.test_read_png(pngfile=monopngfile)
\ No newline at end of file
+        return self.test_read_png(pngfile=monopngfile)
+
+    def test_read_zstack_mono_mask(self):
+        acc = generate_file_accessor(monozstackmask['path'])
+        self.assertTrue(acc.is_mask())
\ No newline at end of file