Merge branch 'generator_plugin'

Author: awilliamson

keras_retinanet/bin/train.py

@@ -17,6 +17,7 @@
 """
 
 import argparse
+from yapsy.PluginManager import PluginManagerSingleton
 import os
 import sys
 
@@ -36,9 +37,6 @@
 from .. import layers
 from ..callbacks import RedirectModel
 from ..callbacks.eval import Evaluate
-from ..preprocessing.pascal_voc import PascalVocGenerator
-from ..preprocessing.csv_generator import CSVGenerator
-from ..preprocessing.open_images import OpenImagesGenerator
 from ..models.resnet import resnet_retinanet, custom_objects, download_imagenet
 from ..utils.transform import random_transform_generator
 from ..utils.keras_version import check_keras_version
@@ -151,74 +149,47 @@ def create_generators(args):
     # create random transform generator for augmenting training data
     transform_generator = random_transform_generator(flip_x_chance=0.5)
 
-    if args.dataset_type == 'coco':
-        # import here to prevent unnecessary dependency on cocoapi
-        from ..preprocessing.coco import CocoGenerator
+    generators = None
 
-        train_generator = CocoGenerator(
-            args.coco_path,
-            'train2017',
-            transform_generator=transform_generator,
-            batch_size=args.batch_size
-        )
-
-        validation_generator = CocoGenerator(
-            args.coco_path,
-            'val2017',
-            batch_size=args.batch_size
-        )
-    elif args.dataset_type == 'pascal':
-        train_generator = PascalVocGenerator(
-            args.pascal_path,
-            'trainval',
-            transform_generator=transform_generator,
-            batch_size=args.batch_size
-        )
-
-        validation_generator = PascalVocGenerator(
-            args.pascal_path,
-            'test',
-            batch_size=args.batch_size
-        )
-    elif args.dataset_type == 'csv':
-        train_generator = CSVGenerator(
-            args.annotations,
-            args.classes,
-            transform_generator=transform_generator,
-            batch_size=args.batch_size
-        )
+    for plugin in PluginManagerSingleton.get().getAllPlugins():
+        try:
+            getattr(plugin.plugin_object, 'dataset_type')
+        except AttributeError:
+            print("Plugin ({}) does not contain a definition for dataset_type and cannot be utilised.".format(plugin.name))
+        else:
+            if plugin.plugin_object.dataset_type.lower() == args.dataset_type.lower():
+                if callable(getattr(plugin.plugin_object, 'get_generator', None)):
+                    generators = plugin.plugin_object.get_generator(args, transform_generator=transform_generator)
+                    break
+                else:
+                    print("Plugin ({}) does not contain a definition for get_generator and cannot be utilised".format(plugin.name))
+
+    if args.dataset_type == 'oid':
+        generators = {
+            'train_generator' : OpenImagesGenerator(
+                args.main_dir,
+                subset='train',
+                version=args.version,
+                labels_filter=args.labels_filter,
+                annotation_cache_dir=args.annotation_cache_dir,
+                transform_generator=transform_generator,
+                batch_size=args.batch_size
+            ),
 
-        if args.val_annotations:
-            validation_generator = CSVGenerator(
-                args.val_annotations,
-                args.classes,
+            'validation_generator' : OpenImagesGenerator(
+                args.main_dir,
+                subset='validation',
+                version=args.version,
+                labels_filter=args.labels_filter,
+                annotation_cache_dir=args.annotation_cache_dir,
                 batch_size=args.batch_size
             )
-        else:
-            validation_generator = None
-    elif args.dataset_type == 'oid':
-        train_generator = OpenImagesGenerator(
-            args.main_dir,
-            subset='train',
-            version=args.version,
-            labels_filter=args.labels_filter,
-            annotation_cache_dir=args.annotation_cache_dir,
-            transform_generator=transform_generator,
-            batch_size=args.batch_size
-        )
+        }
 
-        validation_generator = OpenImagesGenerator(
-            args.main_dir,
-            subset='validation',
-            version=args.version,
-            labels_filter=args.labels_filter,
-            annotation_cache_dir=args.annotation_cache_dir,
-            batch_size=args.batch_size
-        )
-    else:
+    if generators is None:
         raise ValueError('Invalid data type received: {}'.format(args.dataset_type))
 
-    return train_generator, validation_generator
+    return generators["train_generator"], generators["validation_generator"]
 
 
 def check_args(parsed_args):
@@ -241,6 +212,9 @@ def check_args(parsed_args):
             "Multi GPU training ({}) and resuming from snapshots ({}) is not supported.".format(parsed_args.multi_gpu,
                                                                                                 parsed_args.snapshot))
 
+    for plugin in PluginManagerSingleton.get().getAllPlugins():
+        plugin.plugin_object.check_args(parsed_args)
+
     return parsed_args
 
 
@@ -249,26 +223,19 @@ def parse_args(args):
     subparsers = parser.add_subparsers(help='Arguments for specific dataset types.', dest='dataset_type')
     subparsers.required = True
 
-    coco_parser = subparsers.add_parser('coco')
-    coco_parser.add_argument('coco_path', help='Path to dataset directory (ie. /tmp/COCO).')
-
-    pascal_parser = subparsers.add_parser('pascal')
-    pascal_parser.add_argument('pascal_path', help='Path to dataset directory (ie. /tmp/VOCdevkit).')
-
     def csv_list(string):
         return string.split(',')
 
+    # Invoke all loaded plugins for their subparsers.
+    for plugin in PluginManagerSingleton.get().getAllPlugins():
+        plugin.plugin_object.parser_args(subparsers)
+
     oid_parser = subparsers.add_parser('oid')
     oid_parser.add_argument('main_dir', help='Path to dataset directory.')
-    oid_parser.add_argument('--version',  help='The current dataset version is V3.', default='2017_11')
-    oid_parser.add_argument('--labels_filter',  help='A list of labels to filter.', type=csv_list, default=None)
+    oid_parser.add_argument('--version', help='The current dataset version is V3.', default='2017_11')
+    oid_parser.add_argument('--labels_filter', help='A list of labels to filter.', type=csv_list, default=None)
     oid_parser.add_argument('--annotation_cache_dir', help='Path to store annotation cache.', default='.')
 
-    csv_parser = subparsers.add_parser('csv')
-    csv_parser.add_argument('annotations', help='Path to CSV file containing annotations for training.')
-    csv_parser.add_argument('classes', help='Path to a CSV file containing class label mapping.')
-    csv_parser.add_argument('--val-annotations', help='Path to CSV file containing annotations for validation (optional).')
-
     group = parser.add_mutually_exclusive_group()
     group.add_argument('--snapshot',          help='Resume training from a snapshot.')
     group.add_argument('--imagenet-weights',  help='Initialize the model with pretrained imagenet weights. This is the default behaviour.', action='store_const', const=True, default=True)
@@ -287,8 +254,37 @@ def csv_list(string):
 
     return check_args(parser.parse_args(args))
 
+def load_plugins(plugin_path):
+    """
+    Responsible for initialising the plugin manager, setting the plugin directory to search, and loading all available
+    plugins - then activating them.
+
+    :param plugin_path: String/[Str] for plugin paths to check
+    :return: None
+    """
+
+    pm = PluginManagerSingleton.get()
+    pl = pm.getPluginLocator()
+    pl.setPluginInfoExtension("dataset")
+    pm.setPluginLocator(pl)
+
+    plugin_path = [plugin_path] if type(plugin_path) is str else plugin_path
+    pm.setPluginPlaces(plugin_path)
+
+    # Load all plugins
+    pm.collectPlugins()
+
+    for k, n in enumerate(pm.getAllPlugins()):
+        pm.activatePluginByName(n.name)
+        print("Loaded: {}".format(n.name))
 
 def main(args=None):
+    #Load plugins first, as their procedures are needed for parsing args.
+    print("Loading plugins...")
+    # Load Plugins
+    load_plugins(['plugins'])
+    print("Loaded plugins.")
+
     # parse arguments
     if args is None:
         args = sys.argv[1:]

keras_retinanet/plugins/coco/__init__.py

@@ -0,0 +1,23 @@
+;Copyright 2017-2018 Ashley Williamson (https://inp.io)
+;
+;Licensed under the Apache License, Version 2.0 (the "License");
+;you may not use this file except in compliance with the License.
+;You may obtain a copy of the License at
+;
+;    http://www.apache.org/licenses/LICENSE-2.0
+;
+;Unless required by applicable law or agreed to in writing, software
+;distributed under the License is distributed on an "AS IS" BASIS,
+;WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+;See the License for the specific language governing permissions and
+;limitations under the License.
+
+[Core]
+Name = Coco Dataset
+Module = coco
+
+[Documentation]
+Description = RetinaNet Definition for COCO Dataset
+Author = Ashley Williamson
+Version = 1.0
+Website = https://inp.io

keras_retinanet/plugins/coco/coco.dataset

@@ -0,0 +1,54 @@
+"""
+Copyright 2017-2018 Ashley Williamson (https://inp.io)
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+"""
+
+import keras_retinanet.utils.plugin as plugins
+# See https://yapsy.readthedocs.io/en/latest/Advices.html#plugin-class-detection-caveat
+# Caveat surrounding import. Must us 'as' rather than directly importing DatasetPlugin
+
+from keras_retinanet.preprocessing.coco import CocoGenerator
+
+
+class CocoPlugin(plugins.DatasetPlugin):
+    def __init__(self):
+        super(CocoPlugin, self).__init__()
+
+        self.dataset_type = "coco"
+
+    def parser_args(self, subparsers):
+        coco_parser = subparsers.add_parser(self.dataset_type)
+        coco_parser.add_argument('coco_path', help='Path to dataset directory (ie. /tmp/COCO).')
+
+        return coco_parser
+
+    def get_generator(self, args, transform_generator=None):
+        train_generator = CocoGenerator(
+            args.coco_path,
+            'train2017',
+            transform_generator=transform_generator,
+            batch_size=args.batch_size
+        )
+
+        validation_generator = CocoGenerator(
+            args.coco_path,
+            'val2017',
+            batch_size=args.batch_size
+        )
+
+        return {
+            "train_generator": train_generator,
+            "validation_generator": validation_generator
+        }
+

keras_retinanet/plugins/coco/coco.py

@@ -0,0 +1,23 @@
+;Copyright 2017-2018 Ashley Williamson (https://inp.io)
+;
+;Licensed under the Apache License, Version 2.0 (the "License");
+;you may not use this file except in compliance with the License.
+;You may obtain a copy of the License at
+;
+;    http://www.apache.org/licenses/LICENSE-2.0
+;
+;Unless required by applicable law or agreed to in writing, software
+;distributed under the License is distributed on an "AS IS" BASIS,
+;WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+;See the License for the specific language governing permissions and
+;limitations under the License.
+
+[Core]
+Name = CSV Dataset
+Module = csv
+
+[Documentation]
+Description = RetinaNet Definition for CSV Dataset
+Author = Ashley Williamson
+Version = 1.0
+Website = https://inp.io

keras_retinanet/plugins/csv/__init__.py

@@ -0,0 +1,59 @@
+"""
+Copyright 2017-2018 Ashley Williamson (https://inp.io)
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+"""
+
+import keras_retinanet.utils.plugin as plugins
+# See https://yapsy.readthedocs.io/en/latest/Advices.html#plugin-class-detection-caveat
+# Caveat surrounding import. Must us 'as' rather than directly importing DatasetPlugin
+
+from keras_retinanet.preprocessing.csv_generator import CSVGenerator
+
+
+class CSVPlugin(plugins.DatasetPlugin):
+    def __init__(self):
+        super(CSVPlugin, self).__init__()
+
+        self.dataset_type = "csv"
+
+    def parser_args(self, subparsers):
+        csv_parser = subparsers.add_parser(self.dataset_type)
+        csv_parser.add_argument('annotations', help='Path to CSV file containing annotations for training.')
+        csv_parser.add_argument('classes', help='Path to a CSV file containing class label mapping.')
+        csv_parser.add_argument('--val-annotations',
+                                help='Path to CSV file containing annotations for validation (optional).')
+
+        return csv_parser
+
+    def get_generator(self, args, transform_generator=None):
+        train_generator = CSVGenerator(
+            args.annotations,
+            args.classes,
+            transform_generator=transform_generator,
+            batch_size=args.batch_size
+        )
+
+        if args.val_annotations:
+            validation_generator = CSVGenerator(
+                args.val_annotations,
+                args.classes,
+                batch_size=args.batch_size
+            )
+        else:
+            validation_generator = None
+
+        return {
+            "train_generator": train_generator,
+            "validation_generator": validation_generator
+        }

keras_retinanet/plugins/csv/csv.dataset

@@ -0,0 +1,23 @@
+;Copyright 2017-2018 Ashley Williamson (https://inp.io)
+;
+;Licensed under the Apache License, Version 2.0 (the "License");
+;you may not use this file except in compliance with the License.
+;You may obtain a copy of the License at
+;
+;    http://www.apache.org/licenses/LICENSE-2.0
+;
+;Unless required by applicable law or agreed to in writing, software
+;distributed under the License is distributed on an "AS IS" BASIS,
+;WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+;See the License for the specific language governing permissions and
+;limitations under the License.
+
+[Core]
+Name = VOC Dataset
+Module = voc
+
+[Documentation]
+Description = RetinaNet Definition for VOC Dataset
+Author = Ashley Williamson
+Version = 1.0
+Website = https://inp.io