Merge pull request #241 from BrikerMan/develop

Release v0.5.4

.circleci/config.yml

@@ -0,0 +1,153 @@
+version: 2
+
+defualt_executer: &defualt_executer
+  machine:
+    image: circleci/classic:201808-01
+
+default_steps:
+  - switch_to_3_7: &switch_to_3_7
+      run:
+        name: "Switch to Python v3.7"
+        command: |
+          pyenv versions
+          pyenv global 3.7.0
+
+  - create_env: &create_env
+      run:
+        name: create venv
+        command: |
+          python3 -m venv venv
+
+  - save_3_7_cache: &save_3_7_cache
+      save_cache:
+        paths:
+          - ./venv
+        key: py3.7-{{ checksum "requirements.dev.txt" }}
+
+  - restore_3_7_cache: &restore_3_7_cache
+      restore_cache:
+        keys:
+          - py3.7-{{ checksum "requirements.dev.txt" }}
+
+  - install_dependencies: &install_dependencies
+      run:
+        name: install dependencies
+        command: |
+          ls .
+          source venv/bin/activate
+          pip install -r requirements.dev.txt
+
+  - run_tests: &run_tests
+      run:
+        name: run tests
+        no_output_timeout: 1800
+        command: >
+          . venv/bin/activate
+
+          TESTFILES=$(circleci tests glob "tests/**/*.py" | circleci tests split --split-by=timings)
+
+          echo $TESTFILES
+
+          mkdir -pv test-reports
+
+          pytest --doctest-modules --junitxml=test-reports/junit.xml --cov=kashgari --cov-report=xml --cov-report=html $TESTFILES
+
+jobs:
+  lint:
+    <<: *defualt_executer
+
+    steps:
+      - checkout
+      - *switch_to_3_7
+      - *create_env
+      - *restore_3_7_cache
+      - *install_dependencies
+      - *save_3_7_cache
+      - run:
+          name: run lint
+          command: |
+            source venv/bin/activate
+            flake8 kashgari
+
+  test_3.6:
+    <<: *defualt_executer
+    parallelism: 4
+    steps:
+      - checkout
+      - run:
+          name: "Switch to Python v3.6"
+          command: |
+            pyenv versions
+            pyenv global 3.6.5
+      - *create_env
+      - restore_cache:
+          keys:
+            - py3.6-{{ checksum "requirements.dev.txt" }}
+            - py3.6-
+      - *install_dependencies
+      - save_cache:
+          paths:
+            - ./venv
+          key: py3.6-{{ checksum "requirements.dev.txt" }}
+      - *run_tests
+      - store_test_results:
+          path: test-reports
+
+      - store_artifacts:
+          path: test-reports
+
+  test_3.7:
+    <<: *defualt_executer
+    parallelism: 4
+
+    steps:
+      - checkout
+      - *switch_to_3_7
+      - *create_env
+      - *restore_3_7_cache
+      - *install_dependencies
+      - *save_3_7_cache
+      - *run_tests
+      - store_test_results:
+          path: test-reports
+
+      - store_artifacts:
+          path: test-reports
+
+  document:
+    <<: *defualt_executer
+
+    steps:
+      - checkout
+      - *switch_to_3_7
+      - *create_env
+      - *restore_3_7_cache
+      - *install_dependencies
+      - *save_3_7_cache
+      - run:
+          name: Update Documents
+          command: |
+            ls
+            cd mkdocs
+            ls
+            mkdocs build
+
+workflows:
+  version: 2
+  lint-test-and-deploy:
+    jobs:
+      - lint
+      - test_3.6:
+          requires:
+            - lint
+      - test_3.7:
+          requires:
+            - lint
+      - document:
+          filters:
+            branches:
+              only:
+                - master
+          requires:
+            - test_3.6
+            - test_3.7

kashgari/__init__.py

@@ -12,6 +12,10 @@
 """
 import os
 os.environ['TF_KERAS'] = '1'
+os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
+
+import tensorflow as tf
+tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR)
 
 import keras_bert
 from kashgari.macros import TaskType, config

kashgari/embeddings/stacked_embedding.py

@@ -7,6 +7,8 @@
 # file: stacked_embedding.py
 # time: 2019-05-23 09:18
 
+import json
+import pydoc
 from typing import Union, Optional, Tuple, List, Dict
 
 import numpy as np
@@ -15,11 +17,10 @@
 
 import kashgari
 from kashgari.embeddings.base_embedding import Embedding
-from kashgari.processors.base_processor import BaseProcessor
 from kashgari.layers import L
+from kashgari.processors.base_processor import BaseProcessor
 
 
-# Todo: A better name for this class
 class StackedEmbedding(Embedding):
     """Embedding layer without pre-training, train embedding layer while training model"""
 
@@ -28,12 +29,30 @@ def _load_saved_instance(cls,
                              config_dict: Dict,
                              model_path: str,
                              tf_model: keras.Model):
-        pass
+        embeddings = []
+        for embed_info in config_dict['embeddings']:
+            embed_class = pydoc.locate(f"{embed_info['module']}.{embed_info['class_name']}")
+            embedding: Embedding = embed_class._load_saved_instance(embed_info,
+                                                                    model_path,
+                                                                    tf_model)
+            embeddings.append(embedding)
+        instance = cls(embeddings=embeddings,
+                       from_saved_model=True)
+        print('----')
+        print(instance.embeddings)
+
+        embed_model_json_str = json.dumps(config_dict['embed_model'])
+        instance.embed_model = keras.models.model_from_json(embed_model_json_str,
+                                                            custom_objects=kashgari.custom_objects)
+        # Load Weights from model
+        for layer in instance.embed_model.layers:
+            layer.set_weights(tf_model.get_layer(layer.name).get_weights())
+        return instance
 
     def info(self):
         info = super(StackedEmbedding, self).info()
         info['embeddings'] = [embed.info() for embed in self.embeddings]
-        info['config'] = []
+        info['config'] = {}
         return info
 
     def __init__(self,
@@ -58,9 +77,10 @@ def __init__(self,
                                                processor=processor,
                                                from_saved_model=from_saved_model)
 
+        self.embeddings = embeddings
+        self.processor = embeddings[0].processor
+
         if not from_saved_model:
-            self.embeddings = embeddings
-            self.processor = embeddings[0].processor
             self._build_model()
 
     def _build_model(self, **kwargs):
@@ -71,8 +91,6 @@ def _build_model(self, **kwargs):
 
             for embed in self.embeddings:
                 inputs += embed.embed_model.inputs
-                print(embed.embed_model.input)
-                print(embed.embed_model.inputs)
 
             # inputs = [embed.embed_model.inputs for embed in self.embeddings]
             outputs = layer_concatenate([embed.embed_model.output for embed in self.embeddings])

kashgari/tasks/base_model.py

@@ -7,17 +7,19 @@
 # file: base_model.py
 # time: 2019-05-22 11:21
 
-
+import os
 import json
 import logging
-import os
+import warnings
 import pathlib
 from typing import Dict, Any, List, Optional, Union, Tuple
 
 import numpy as np
 import tensorflow as tf
 from tensorflow import keras
 
+import kashgari
+
 from kashgari import utils
 from kashgari.embeddings import BareEmbedding
 from kashgari.embeddings.base_embedding import Embedding
@@ -44,7 +46,7 @@ def info(self):
             'class_name': self.__class__.__name__,
             'module': self.__class__.__module__,
             'tf_version': tf.__version__,
-            'kashgari_version': tf.__version__
+            'kashgari_version': kashgari.__version__
         }
 
     @property
@@ -59,6 +61,17 @@ def token2idx(self) -> Dict[str, int]:
     def label2idx(self) -> Dict[str, int]:
         return self.embedding.label2idx
 
+    @property
+    def pre_processor(self):
+        warnings.warn("The 'pre_processor' property is deprecated, "
+                      "use 'processor' instead", DeprecationWarning, 2)
+        """Deprecated. Use `self.processor` instead."""
+        return self.embedding.processor
+
+    @property
+    def processor(self):
+        return self.embedding.processor
+
     def __init__(self,
                  embedding: Optional[Embedding] = None,
                  hyper_parameters: Optional[Dict[str, Dict[str, Any]]] = None):
@@ -87,7 +100,6 @@ def __init__(self,
         self.tf_model: keras.Model = None
         self.hyper_parameters = self.get_default_hyper_parameters()
         self.model_info = {}
-        self.pre_processor = self.embedding.processor
 
         if hyper_parameters:
             self.hyper_parameters.update(hyper_parameters)

kashgari/tasks/classification/base_model.py

@@ -179,12 +179,12 @@ def evaluate(self,
                 logging.debug('y      : {}'.format(y_data[index]))
                 logging.debug('y_pred : {}'.format(y_pred[index]))
 
-        if self.pre_processor.multi_label:
-            y_pred_b = self.pre_processor.multi_label_binarizer.fit_transform(y_pred)
-            y_true_b = self.pre_processor.multi_label_binarizer.fit_transform(y_data)
+        if self.processor.multi_label:
+            y_pred_b = self.processor.multi_label_binarizer.fit_transform(y_pred)
+            y_true_b = self.processor.multi_label_binarizer.fit_transform(y_data)
             report = metrics.classification_report(y_pred_b,
                                                    y_true_b,
-                                                   target_names=self.pre_processor.multi_label_binarizer.classes_,
+                                                   target_names=self.processor.multi_label_binarizer.classes_,
                                                    output_dict=output_dict,
                                                    digits=digits)
         else:

kashgari/tasks/classification/dpcnn_model.py

@@ -136,7 +136,7 @@ def resnet_block(self, inputs, filters: int, kernel_size: int = 3,
         return tensor_out
 
     def build_model_arc(self):
-        output_dim = len(self.pre_processor.label2idx)
+        output_dim = len(self.processor.label2idx)
         config = self.hyper_parameters
         embed_model = self.embedding.embed_model