chore: merge dev into release

chore: merge dev into release

moseq2_pca/__init__.py

@@ -1 +1 @@
-__version__ = 'v1.1.3'
+__version__ = 'v1.2.0'

moseq2_pca/cli.py

@@ -100,8 +100,17 @@ def common_dask_parameters(function):
 @click.option('--output-file', default='pca', type=str, help='Name of h5 file for storing pca results')
 @click.option('--local-processes', default=False, type=bool, help='Used with a local cluster. If True: use processes, If False: use threads')
 @click.option('--overwrite-pca-train', default=False, type=bool, help='Used to bypass the pca overwrite question. If True: skip question, run automatically')
+@click.option('--camera-type', default='k2', type=str, help='specify the camera type (k2 or azure), default is k2')
 def train_pca(input_dir, output_dir, output_file, **cli_args):
     # function writes output pca path to config_data
+    if cli_args.get('camera_type') == 'azure':
+        # check if parameters are set to k2 default, change to azure default
+        print('Updating parameters for Azure Kinect camera...')
+        if cli_args['gaussfilter_space'] == (1.5, 1):
+            cli_args['gaussfilter_space'] = (2.25, 1.5)
+        if cli_args['tailfilter_size'] == (9, 9):
+            cli_args['tailfilter_size'] = (13, 13)
+
     config_data = train_pca_wrapper(input_dir, cli_args, output_dir, output_file)
     # write config_data to config_file if there is one
     if cli_args.get('config_file'):

moseq2_pca/helpers/data.py

@@ -22,43 +22,44 @@ def get_pca_paths(config_data, output_dir):
     pca_file_scores (str): path to pca_scores file
     """
 
-    # Get path to pre-computed PCA file
-    pca_file_components = join(output_dir, 'pca.h5')
-    if 'pca_file_components' not in config_data:
-        config_data['pca_file_components'] = pca_file_components
-    elif config_data['pca_file_components'] is not None:
-        pca_file_components = config_data['pca_file_components']
+    # Check if there is PCA file from config_data
+    if config_data.get('pca_file', None) is not None:
+        pca_file = config_data['pca_file']
+    else:
+        # Assume PCA file is in output_dir
+        pca_file = join(output_dir, 'pca.h5')
+        config_data['pca_file'] = pca_file
 
-    if not exists(pca_file_components):
-        raise IOError(f'Could not find PCA components file {pca_file_components}')
+    if not exists(pca_file):
+        raise IOError(f'Could not find PCA components file {pca_file}')
 
     # Get path to PCA Scores
     pca_file_scores = config_data.get('pca_file_scores', join(output_dir, 'pca_scores.h5'))
     config_data['pca_file_scores'] = pca_file_scores
 
-    return config_data, pca_file_components, pca_file_scores
+    return config_data, pca_file, pca_file_scores
 
-def load_pcs_for_cp(pca_file_components, config_data):
+def load_pcs_for_cp(pca_file, config_data):
     """
     Load computed Principal Components for Model-free Changepoint Analysis.
 
     Args:
-    pca_file_components (str): path to pca h5 file to read PCs
+    pca_file (str): path to pca h5 file to read PCs
     config_data (dict): config parameters
 
     Returns:
-    pca_components (str): path to pca components
+    pca_file (str): path to pca components
     changepoint_params (dict): dict of relevant changepoint parameters
     missing_data (bool): Indicates whether to use mask_params for missing data pca
     mask_params (dict): Mask parameters to use when computing CPs
     """
 
-    print(f'Loading PCs from {pca_file_components}')
-    with h5py.File(pca_file_components, 'r') as f:
+    print(f'Loading PCs from {pca_file}')
+    with h5py.File(pca_file, 'r') as f:
         pca_components = f[config_data['pca_path']][()]
 
     # get the yaml for pca, check parameters, if we used fft, be sure to turn on here...
-    pca_yaml = splitext(pca_file_components)[0] + '.yaml'
+    pca_yaml = splitext(pca_file)[0] + '.yaml'
 
     if exists(pca_yaml):
         with open(pca_yaml, 'r') as f:

moseq2_pca/helpers/wrappers.py

@@ -235,7 +235,7 @@ def apply_pca_wrapper(input_dir, config_data, output_dir, output_file):
     config_data, pca_file, pca_file_scores = get_pca_paths(config_data, output_dir)
 
     print('Loading PCs from', pca_file)
-    with h5py.File(config_data['pca_file_components'], 'r') as f:
+    with h5py.File(config_data['pca_file'], 'r') as f:
         pca_components = f[config_data['pca_path']][()]
 
     # Get the yaml for pca, check parameters, if we used fft, be sure to turn on here...
@@ -313,10 +313,10 @@ def compute_changepoints_wrapper(input_dir, config_data, output_dir, output_file
     save_file = join(output_dir, output_file)
 
     # Get paths to PCA, PCA Scores file
-    config_data, pca_file_components, pca_file_scores = get_pca_paths(config_data, output_dir)
+    config_data, pca_file, pca_file_scores = get_pca_paths(config_data, output_dir)
 
     # Load Principal components, set up changepoint parameter dict, and optionally load reconstructed PCs.
-    pca_components, changepoint_params, missing_data, mask_params = load_pcs_for_cp(pca_file_components, config_data)
+    pca_components, changepoint_params, missing_data, mask_params = load_pcs_for_cp(pca_file, config_data)
 
     # Initialize Dask client
     client, cluster, workers = \

tests/integration_tests/test_cli.py

@@ -113,7 +113,7 @@ def test_compute_changepoints(self):
         with open(config, 'r') as f:
             config_data = yaml.safe_load(f)
 
-        config_data['pca_file_components'] = None
+        config_data['pca_file'] = None
         config_data['pca_file_scores'] = None
 
         with open(config, 'w') as f:

tests/integration_tests/test_gui.py

@@ -97,7 +97,7 @@ def test_apply_pca_command(self):
 
         with open(config_file, 'r') as f:
             config_data = yaml.safe_load(f)
-            config_data['pca_file_components'] = join(outpath, 'pca.h5')
+            config_data['pca_file'] = join(outpath, 'pca.h5')
 
             config_data['use_fft'] = True
             config_data['missing_data'] = False
@@ -132,7 +132,7 @@ def test_compute_changepoints_command(self):
 
         with open(config_file, 'r') as f:
             config_data = yaml.safe_load(f)
-            config_data['pca_file_components'] = join(outpath, 'pca.h5')
+            config_data['pca_file'] = join(outpath, 'pca.h5')
 
             config_data['use_fft'] = True
             config_data['missing_data'] = False