fix(executors): fix the mro issue

jina/executors/encoders/frameworks.py

@@ -7,11 +7,15 @@
 
 from . import BaseNumericEncoder
 from ..decorators import batching, as_ndarray
-from ..frameworks import BaseOnnxExecutor, BasePaddleExecutor, BaseTorchExecutor
+from ..frameworks import BaseOnnxExecutor, BasePaddleExecutor, BaseTorchExecutor, BaseTFExecutor, BaseFrameworkExecutor
 from ...helper import is_url
 
 
-class BaseOnnxEncoder(BaseOnnxExecutor):
+class BaseFrameworkEncoder(BaseFrameworkExecutor, BaseNumericEncoder):
+    pass
+
+
+class BaseOnnxEncoder(BaseOnnxExecutor, BaseFrameworkEncoder):
 
     def __init__(self, output_feature: str, model_path: str = None, *args, **kwargs):
         """
@@ -56,7 +60,55 @@ def _append_outputs(input_fn, outputs_name_to_append, output_fn):
         onnx.save(model, output_fn)
 
 
-class BaseCVPaddlehubEncoder(BasePaddleExecutor, BaseNumericEncoder):
+class BaseTorchEncoder(BaseTorchExecutor, BaseFrameworkEncoder):
+    """"
+    :class:`BaseTorchEncoder` implements the common part for :class:`ImageTorchEncoder` and :class:`VideoTorchEncoder`.
+
+    ..warning::
+        :class:`BaseTorchEncoder`  is not intented to be used to do the real encoding.
+    """
+
+    def __init__(self,
+                 model_name: str = '',
+                 channel_axis: int = 1,
+                 *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.model_name = model_name
+        self.channel_axis = channel_axis
+        self._default_channel_axis = 1
+
+    @batching
+    @as_ndarray
+    def encode(self, data: 'np.ndarray', *args, **kwargs) -> 'np.ndarray':
+        import numpy as np
+        if self.channel_axis != self._default_channel_axis:
+            data = np.moveaxis(data, self.channel_axis, self._default_channel_axis)
+        import torch
+        _input = torch.from_numpy(data.astype('float32'))
+        if self.on_gpu:
+            _input = _input.cuda()
+        _feature = self._get_features(_input).detach()
+        if self.on_gpu:
+            _feature = _feature.cpu()
+        _feature = _feature.numpy()
+        return self._get_pooling(_feature)
+
+    def _get_features(self, data):
+        raise NotImplementedError
+
+    def _get_pooling(self, feature_map):
+        return feature_map
+
+
+class BaseTFEncoder(BaseTFExecutor, BaseFrameworkEncoder):
+    pass
+
+
+class BasePaddlehubEncoder(BasePaddleExecutor, BaseFrameworkEncoder):
+    pass
+
+
+class BaseCVPaddlehubEncoder(BasePaddlehubEncoder):
     """
     :class:`BaseCVPaddlehubEncoder` implements the common parts for :class:`ImagePaddlehubEncoder` and
         :class:`VideoPaddlehubEncoder`.
@@ -66,8 +118,8 @@ class BaseCVPaddlehubEncoder(BasePaddleExecutor, BaseNumericEncoder):
     """
 
     def __init__(self,
-                 model_name: str,
-                 output_feature: str,
+                 model_name: str = None,
+                 output_feature: str = None,
                  pool_strategy: str = None,
                  channel_axis: int = -3,
                  *args,
@@ -119,41 +171,3 @@ def get_pooling(self, data: 'np.ndarray', axis=None) -> 'np.ndarray':
         return getattr(np, self.pool_strategy)(data, axis=_reduce_axis)
 
 
-class BaseTorchEncoder(BaseTorchExecutor):
-    """"
-    :class:`BaseTorchEncoder` implements the common part for :class:`ImageTorchEncoder` and :class:`VideoTorchEncoder`.
-
-    ..warning::
-        :class:`BaseTorchEncoder`  is not intented to be used to do the real encoding.
-    """
-
-    def __init__(self,
-                 model_name: str,
-                 channel_axis: int = 1,
-                 *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self.model_name = model_name
-        self.channel_axis = channel_axis
-        self._default_channel_axis = 1
-
-    @batching
-    @as_ndarray
-    def encode(self, data: 'np.ndarray', *args, **kwargs) -> 'np.ndarray':
-        import numpy as np
-        if self.channel_axis != self._default_channel_axis:
-            data = np.moveaxis(data, self.channel_axis, self._default_channel_axis)
-        import torch
-        _input = torch.from_numpy(data.astype('float32'))
-        if self.on_gpu:
-            _input = _input.cuda()
-        _feature = self._get_features(_input).detach()
-        if self.on_gpu:
-            _feature = _feature.cpu()
-        _feature = _feature.numpy()
-        return self._get_pooling(_feature)
-
-    def _get_features(self, data):
-        raise NotImplementedError
-
-    def _get_pooling(self, feature_map):
-        return feature_map

jina/executors/encoders/image/onnx.py

@@ -3,11 +3,12 @@
 
 import numpy as np
 
+from .. import BaseImageEncoder
 from ..frameworks import BaseOnnxEncoder
 from ...decorators import batching, as_ndarray
 
 
-class OnnxImageEncoder(BaseOnnxEncoder):
+class OnnxImageEncoder(BaseImageEncoder, BaseOnnxEncoder):
     """
     :class:`OnnxImageEncoder` encodes data from a ndarray, potentially B x (Channel x Height x Width) into a
         ndarray of `B x D`.

jina/executors/encoders/image/paddlehub.py

@@ -1,10 +1,11 @@
 __copyright__ = "Copyright (c) 2020 Jina AI Limited. All rights reserved."
 __license__ = "Apache-2.0"
 
+from .. import BaseImageEncoder
 from ..frameworks import BaseCVPaddlehubEncoder
 
 
-class ImagePaddlehubEncoder(BaseCVPaddlehubEncoder):
+class ImagePaddlehubEncoder(BaseImageEncoder, BaseCVPaddlehubEncoder):
     """
     :class:`ImagePaddlehubEncoder` encodes data from a ndarray, potentially B x (Channel x Height x Width) into a
         ndarray of `B x D`.

jina/executors/encoders/image/tfkeras.py

@@ -5,10 +5,10 @@
 
 from .. import BaseImageEncoder
 from ...decorators import batching, as_ndarray
-from ...frameworks import BaseTFExecutor
+from ..frameworks import BaseTFEncoder
 
 
-class KerasImageEncoder(BaseTFExecutor, BaseImageEncoder):
+class KerasImageEncoder(BaseImageEncoder, BaseTFEncoder):
     """
     :class:`KerasImageEncoder` encodes data from a ndarray, potentially B x (Channel x Height x Width) into a
         ndarray of `B x D`.

jina/executors/encoders/image/torchvision.py

@@ -3,10 +3,11 @@
 
 import numpy as np
 
+from .. import BaseImageEncoder
 from ..frameworks import BaseTorchEncoder
 
 
-class ImageTorchEncoder(BaseTorchEncoder):
+class ImageTorchEncoder(BaseImageEncoder, BaseTorchEncoder):
     """
     :class:`ImageTorchEncoder` encodes data from a ndarray, potentially B x (Channel x Height x Width) into a
         ndarray of `B x D`.

jina/executors/encoders/nlp/farm.py

@@ -5,10 +5,10 @@
 
 from .. import BaseTextEncoder
 from ...decorators import batching, as_ndarray
-from ...frameworks import BaseTorchExecutor
+from ..frameworks import BaseTorchEncoder
 
 
-class FarmTextEncoder(BaseTorchExecutor, BaseTextEncoder):
+class FarmTextEncoder(BaseTextEncoder, BaseTorchEncoder):
     """FARM-based text encoder: (Framework for Adapting Representation Models)
     https://github.com/deepset-ai/FARM
 

jina/executors/encoders/nlp/flair.py

@@ -7,10 +7,10 @@
 
 from .. import BaseTextEncoder
 from ...decorators import batching, as_ndarray
-from ...frameworks import BaseTorchExecutor
+from ..frameworks import BaseTorchEncoder
 
 
-class FlairTextEncoder(BaseTorchExecutor, BaseTextEncoder):
+class FlairTextEncoder(BaseTextEncoder, BaseTorchEncoder):
     """
     :class:`FlairTextEncoder` encodes data from an array of string in size `B` into a ndarray in size `B x D`.
     Internally, :class:`FlairTextEncoder` wraps the DocumentPoolEmbeddings from Flair.

jina/executors/encoders/nlp/paddlehub.py

@@ -5,10 +5,10 @@
 
 from .. import BaseTextEncoder
 from ...decorators import batching, as_ndarray
-from ...frameworks import BasePaddleExecutor
+from ..frameworks import BasePaddlehubEncoder
 
 
-class TextPaddlehubEncoder(BasePaddleExecutor, BaseTextEncoder):
+class TextPaddlehubEncoder(BaseTextEncoder, BasePaddlehubEncoder):
     """
     :class:`TextPaddlehubEncoder` encodes data from an array of string in size `B` into a ndarray in size `B x D`.
     Internally, :class:`TextPaddlehubEncoder` wraps the Ernie module from paddlehub.

jina/executors/encoders/nlp/transformer.py

@@ -7,10 +7,10 @@
 
 from ..helper import reduce_mean, reduce_max, reduce_min, reduce_cls
 from ...decorators import batching, as_ndarray
-from ...frameworks import BaseFrameworkExecutor, BaseTorchExecutor, BaseTFExecutor
+from ..frameworks import BaseFrameworkEncoder, BaseTFEncoder, BaseTorchEncoder
 
 
-class BaseTransformerEncoder(BaseFrameworkExecutor):
+class BaseTransformerEncoder(BaseFrameworkEncoder):
     """
     :class:`TransformerTextEncoder` encodes data from an array of string in size `B` into an ndarray in size `B x D`.
     """
@@ -174,7 +174,7 @@ def get_tensor_func(self):
         raise NotImplementedError
 
 
-class TransformerTFEncoder(BaseTransformerEncoder, BaseTFExecutor):
+class TransformerTFEncoder(BaseTransformerEncoder, BaseTFEncoder):
     """
     Internally, TransformerTFEncoder wraps the tensorflow-version of transformers from huggingface.
     """
@@ -209,7 +209,7 @@ def get_tensor_func(self):
         return tf.constant
 
 
-class TransformerTorchEncoder(BaseTransformerEncoder, BaseTorchExecutor):
+class TransformerTorchEncoder(BaseTransformerEncoder, BaseTorchEncoder):
     """
     Internally, TransformerTorchEncoder wraps the pytorch-version of transformers from huggingface.
     """

jina/executors/encoders/video/paddlehub.py

@@ -1,10 +1,12 @@
 __copyright__ = "Copyright (c) 2020 Jina AI Limited. All rights reserved."
 __license__ = "Apache-2.0"
 
+
+from .. import BaseVideoEncoder
 from ..frameworks import BaseCVPaddlehubEncoder
 
 
-class VideoPaddlehubEncoder(BaseCVPaddlehubEncoder):
+class VideoPaddlehubEncoder(BaseVideoEncoder, BaseCVPaddlehubEncoder):
     """
     :class:`VideoPaddlehubEncoder` encodes data from a ndarray, potentially B x T x (Channel x Height x Width) into a
         ndarray of `B x D`.

jina/executors/encoders/video/torchvision.py

@@ -1,10 +1,11 @@
 __copyright__ = "Copyright (c) 2020 Jina AI Limited. All rights reserved."
 __license__ = "Apache-2.0"
 
+from .. import BaseVideoEncoder
 from ..frameworks import BaseTorchEncoder
 
 
-class VideoTorchEncoder(BaseTorchEncoder):
+class VideoTorchEncoder(BaseVideoEncoder, BaseTorchEncoder):
     """
     :class:`VideoTorchEncoder` encodes data from a ndarray, potentially B x T x (Channel x Height x Width) into an
         ndarray of `B x D`.