added sem. segm. example for compartment prediction (examples/semseg_axon.py),...

added sem. segm. example for compartment prediction (examples/semseg_axon.py), fixed order mis-match between ssd.ssv_ids and arrays loaded from ssd.load_cached_data(..), updated docs

README.md

@@ -5,20 +5,21 @@ Refactored version of SyConn for automated synaptic connectivity inference based
  see below or checkout the branch [dorkenwald2017nm](https://github.com/StructuralNeurobiologyLab/SyConn/tree/dorkenwald2017nm).
 
 Current features:
-- introduction of supervoxel and agglomerated supervoxel classes
-- added support for (sub-) cellular compartment (spines, axon/dendrite/soma) and cell type classification with skeleton- [\[1\]](https://www.nature.com/articles/nmeth.4206) and multiview-based [\[2\]](https://www.nature.com/articles/s41467-019-10836-3) approaches
-- cell organelle prediction, extraction and mesh generation
+- object-oriented handler for supervoxels (e.g. cell fragments, predicted cellular organelles like mitochondria, vesicle clouds etc.)
+- prediction of subcellular structures, supervoxel extraction and mesh generation
+- (sub-) cellular compartment (spines, bouton and axon/dendrite/soma) and cell type classification with multiview- [\[2\]](https://www.nature.com/articles/s41467-019-10836-3) and with skeleton-based approaches [\[1\]](https://www.nature.com/articles/nmeth.4206)
 - glia identification and separation [\[2\]](https://www.nature.com/articles/s41467-019-10836-3)
-- generation of connectivity matrix
+- connectivity matrix export
+- KNOSSOS plugin for interactive exploration (work in progress)
 
-If you use parts of this code base in your academic projects, please cite the corresponding publication.
+If you use parts of this code base in your academic projects, please cite the corresponding publication(s).
 
 
 Documentation
 -------------
 The documentation for the refactored version is still work-in-progress and can be found [here](docs/doc.md). Alternatively see the latest [readthedocs build](https://syconn.readthedocs.io/en/latest/).
 
-For SyConn v1, please have a look at the old [documentation](https://structuralneurobiologylab.github.io/SyConn/documentation/). We also present more general information about SyConn on our [Website](https://structuralneurobiologylab.github.io/SyConn/).
+For SyConn v1, please refer to the old [documentation](https://structuralneurobiologylab.github.io/SyConn/documentation/). We also present more general information about SyConn on our [Website](https://structuralneurobiologylab.github.io/SyConn/).
 
 
 The Team
@@ -35,8 +36,8 @@ The original code snippet (under the Apache License 2.0) used for our project ca
 [here](https://github.com/deepmind/dm_control/blob/30069ac11b60ee71acbd9159547d0bc334d63281/dm_control/_render/pyopengl/egl_ext.py).
 
 
-References
-----------
+Publications
+------------
 \[1\] [Automated synaptic connectivity inference for volume electron microscopy](https://www.nature.com/articles/nmeth.4206)
 ```
  @ARTICLE{SyConn2017,

examples/semseg_axon.py

+from syconn.reps.super_segmentation import *
+from syconn.reps.super_segmentation_helper import semseg_of_sso_nocache
+from syconn.proc.ssd_assembly import init_sso_from_kzip
+from syconn.handler.prediction import get_semseg_axon_model
+import os
+import argparse
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='SyConn example run')
+    parser.add_argument('--working_dir', type=str,
+                        default=os.path.expanduser("~/SyConn/example_cube1/"),
+                        help='Working directory of SyConn')
+    parser.add_argument('--kzip', type=str, default='',
+                        help='path to kzip file which contains a cell reconstruction (see '
+                             'SuperSegmentationObject().export2kzip())')
+    args = parser.parse_args()
+
+    # path to working directory of example cube - required to load pretrained models
+    path_to_workingdir = os.path.expanduser(args.working_dir)
+
+    # path to cell reconstruction k.zip
+    cell_kzip_fn = os.path.abspath(os.path.expanduser(args.kzip))
+    if not os.path.isfile(cell_kzip_fn):
+        raise FileNotFoundError
+    # set working directory to obtain models
+    global_params.wd = path_to_workingdir
+
+    # get model for compartment detection
+    m = get_semseg_axon_model()
+    view_props = global_params.view_properties_semsegax
+    view_props["verbose"] = True
+
+    # load SSO instance from k.zip file
+    sso = init_sso_from_kzip(cell_kzip_fn, sso_id=1)
+
+    # run prediction and store result in new kzip
+    cell_kzip_fn_axon = cell_kzip_fn[:-6] + '_axon.k.zip'
+    semseg_of_sso_nocache(sso, dest_path=cell_kzip_fn_axon, model=m,
+                          **view_props)

examples/semseg_spine.py

@@ -21,8 +21,9 @@ if __name__ == '__main__':
     path_to_workingdir = os.path.expanduser(args.working_dir)
 
     # path to cell reconstruction k.zip
-    cell_kzip_fn = os.path.expanduser(args.kzip)
-
+    cell_kzip_fn = os.path.abspath(os.path.expanduser(args.kzip))
+    if not os.path.isfile(cell_kzip_fn):
+        raise FileNotFoundError
     # set working directory to obtain models
     global_params.wd = path_to_workingdir
 
@@ -30,10 +31,11 @@ if __name__ == '__main__':
     m = get_semseg_spiness_model()
 
     # load SSO instance from k.zip file
-    sso = init_sso_from_kzip(cell_kzip_fn)
+    sso = init_sso_from_kzip(cell_kzip_fn, sso_id=1)
 
     # run prediction and store result in new kzip
     cell_kzip_fn_spines = cell_kzip_fn[:-6] + '_spines.k.zip'
-    semseg_of_sso_nocache(sso, dest_path=cell_kzip_fn_spines, semseg_key="spinesstest", n_avg=5,
-                          ws=(256, 128), model=m,  nb_views=2, comp_window=8e3, verbose=True)
+    semseg_of_sso_nocache(sso, dest_path=cell_kzip_fn_spines, verbose=True,
+                          semseg_key="spinesstest", n_avg=5, ws=(256, 128),
+                          model=m,  nb_views=2, comp_window=8e3)
 

scripts/kplugin/syconn_knossos_viewer.py

@@ -927,6 +927,9 @@ def int2str_label_converter(label, gt_type):
         l_dc_inv = dict(STN=0, DA=1, MSN=2, LMAN=3, HVC=4, GP=5, FS=6, TAN=7)
         l_dc_inv["?"] = 8
         l_dc = {v: k for k, v in l_dc_inv.items()}
+        # Do not distinguish between FS and INT/?
+        l_dc[8] = "INT"
+        l_dc[6] = "INT"
         try:
             return l_dc[label]
         except KeyError:

syconn/handler/multiviews.py

@@ -191,6 +191,9 @@ def int2str_converter(label, gt_type):
         l_dc_inv = dict(STN=0, DA=1, MSN=2, LMAN=3, HVC=4, GP=5, FS=6, TAN=7)
         l_dc_inv["?"] = 8
         l_dc = {v: k for k, v in l_dc_inv.items()}
+        # Do not distinguish between FS and INT/?
+        l_dc[8] = "INT"
+        l_dc[6] = "INT"
         try:
             return l_dc[label]
         except KeyError:

syconn/reps/super_segmentation_dataset.py

@@ -215,13 +215,16 @@ class SuperSegmentationDataset(object):
     @property
     def ssv_ids(self):
         if self._ssv_ids is None:
-            if len(self.mapping_dict) > 0:
+            # do not change the order of the if statements as it is crucial
+            # for the resulting ordering of self.ssv_ids (only ids.npy matches
+            # with all the other cached numpy arrays).
+            if os.path.exists(self.path + "/ids.npy"):
+                self._ssv_ids = np.load(self.path + "/ids.npy")
+            elif len(self.mapping_dict) > 0:
                 self._ssv_ids = np.array(list(self.mapping_dict.keys()))
             elif self.mapping_dict_exists:
                 self.load_mapping_dict()
                 self._ssv_ids = np.array(list(self.mapping_dict.keys()))
-            elif os.path.exists(self.path + "/ids.npy"):
-                self._ssv_ids = np.load(self.path + "/ids.npy")
             else:
                 paths = glob.glob(self.path + "/so_storage/*/*/*/")
                 self._ssv_ids = np.array([int(os.path.basename(p.strip("/")))
@@ -250,6 +253,7 @@ class SuperSegmentationDataset(object):
         return self._id_changer
 
     def load_cached_data(self, name):
+        # TODO: remove this 's' concept
         if os.path.exists(self.path + name + "s.npy"):
             return np.load(self.path + name + "s.npy", allow_pickle=True)