Merge 2.8.x into main

.codeclimate.yml

@@ -1,6 +1,7 @@
 version: "2"
 exclude_patterns:
 - rasa/core/utils.py   # codeclimate has some encoding issues with this files because of emojis
+- rasa/utils/tensorflow/temp_keras_modules.py # largely copy paste and temporary
 - .*
 - .github/
 - CHANGELOG.mdx
@@ -15,7 +16,7 @@ exclude_patterns:
 - bandit.yml
 - binder/
 - changelog/
-- data/        
+- data/
 - docs/
 - examples/
 - poetry.lock

.coveragerc

@@ -6,5 +6,7 @@ exclude_lines =
     if __name__ == .__main__.:
     def create_argument_parser
     if typing.TYPE_CHECKING
+    # Ellipsis are used as placeholders in python 3 that will be overridden
+    \.\.\.
 [run]
 source = rasa

.github/workflows/ci-model-regression-on-schedule.yml

@@ -59,10 +59,9 @@ jobs:
 
       - name: Get TensorFlow version
         run: |-
-          # Read TF version, trim special characters ^~><+
-          TF_VERSION=$(grep "tensorflow =" pyproject.toml | sed 's/^[^"]*"\([^"]*\),.*/\1/' | sed 's/[\^~><=]//g')
-          # Keep the first 3 characters, e.g. we keep 2.3 if TF_VERSION is 2.3.4
-          TF_VERSION=${TF_VERSION::3}
+          # Read TF version from poetry.lock file
+          pip install toml
+          TF_VERSION=$(scripts/read_tenforflow_version.sh)
           echo "TensorFlow version: $TF_VERSION"
           echo TF_VERSION=$TF_VERSION >> $GITHUB_ENV
 
@@ -166,6 +165,12 @@ jobs:
               && echo "::set-output name=is_config_exists::false" && exit 0)
           fi
 
+          # Skip job if dataset is Hermit and config is BERT + DIET(seq) + ResponseSelector(t2t) or Sparse + BERT + DIET(seq) + ResponseSelector(t2t)
+          if [[ "${{ matrix.dataset }}" == "Hermit" && "${{ matrix.config }}" =~ "BERT + DIET(seq) + ResponseSelector(t2t)" ]]; then
+            echo "::warning::This ${{ matrix.dataset }} dataset / ${{ matrix.config }} config is currently being skipped due to OOM associated with the upgrade to TF 2.6." \
+              && echo "::set-output name=is_config_exists::false" && exit 0
+          fi
+
           # Skip job if a given type is not available for a given dataset
           if [[ -z "${DOMAIN_FILE}" && "${{ matrix.type }}" == "core" ]]; then
             echo "::warning::The ${{ matrix.dataset }} dataset doesn't include core type. Skipping the job." \
@@ -255,7 +260,6 @@ jobs:
         env:
           TFHUB_CACHE_DIR: ~/.tfhub_cache/
           OMP_NUM_THREADS: 1
-          TF_DETERMINISTIC_OPS: 1 # Run GPUs in a deterministic mode
         run: |-
           poetry run rasa --version
 

.github/workflows/ci-model-regression.yml

@@ -145,8 +145,9 @@ jobs:
 
       - name: Get TensorFlow version
         run: |-
-          # Read TF version, trim special characters ^~><+
-          TF_VERSION=$(grep "tensorflow =" pyproject.toml | sed 's/^[^"]*"\([^"]*\),.*/\1/' | sed 's/[\^~><=]//g')
+          # Read TF version from poetry.lock file
+          pip install toml
+          TF_VERSION=$(scripts/read_tenforflow_version.sh)
           # Keep the first 3 characters, e.g. we keep 2.3 if TF_VERSION is 2.3.4
           TF_VERSION=${TF_VERSION::3}
           echo "TensorFlow version: $TF_VERSION"
@@ -233,6 +234,11 @@ jobs:
           echo "::set-output name=is_config_exists::true"
           echo "::set-output name=is_external::${IS_EXTERNAL}"
 
+          # Warn about job if dataset is Hermit and config is BERT + DIET(seq) + ResponseSelector(t2t) or Sparse + BERT + DIET(seq) + ResponseSelector(t2t)
+          if [[ "${{ matrix.dataset }}" == "Hermit" && "${{ matrix.config }}" =~ "BERT + DIET(seq) + ResponseSelector(t2t)" ]]; then
+            echo "::warning::This ${{ matrix.dataset }} dataset / ${{ matrix.config }} config is currently being skipped on scheduled tests due to OOM associated with the upgrade to TF 2.6. You may see OOM here."
+          fi
+
           if [[ "${IS_EXTERNAL}" == "true" ]]; then
             echo "DATASET_DIR=dataset_external" >> $GITHUB_ENV
           else
@@ -330,7 +336,6 @@ jobs:
         env:
           TFHUB_CACHE_DIR: ~/.tfhub_cache/
           OMP_NUM_THREADS: 1
-          TF_DETERMINISTIC_OPS: 1
         run: |-
           poetry run rasa --version
 

.github/workflows/continous-integration.yml

@@ -308,9 +308,26 @@ jobs:
     - name: Install ddtrace
       if: needs.changes.outputs.backend == 'true'
       run: poetry run pip install -U ddtrace
+
+    - name: Test Code 🔍 (monothread)
+      # This is a temporary workaround for OOM on Windows caused by increased memory
+      # consumption associated with the upgrade of our tensorflow dependency
+      # see https://github.com/RasaHQ/rasa/issues/9734 for more information.
+      # Once the memory issues have been addressed, all configurations should be run on
+      # multiple threads
+      if: needs.changes.outputs.backend == 'true' && (matrix.os == 'windows-latest' && matrix.test == 'test-nlu-predictors')
+      env:
+        JOBS: 1
+        PYTHONIOENCODING: "utf-8"
+        DD_ENV: ${{ matrix.test }}
+        DD_SERVICE: rasa
+        DD_ARGS: --ddtrace
+      run: |
+        make ${{ matrix.test }}
+        poetry run coverage xml -o ${{ github.workspace }}/tests/${{ matrix.test }}-coverage.xml --include="rasa/*"
 
-    - name: Test Code 🔍
-      if: needs.changes.outputs.backend == 'true'
+    - name: Test Code 🔍 (multithread)
+      if: needs.changes.outputs.backend == 'true' && !(matrix.os == 'windows-latest' && matrix.test == 'test-nlu-predictors')
       env:
         JOBS: 2
         PYTHONIOENCODING: "utf-8"

CHANGELOG.mdx

@@ -16,6 +16,100 @@ https://github.com/RasaHQ/rasa/tree/main/changelog/ . -->
 
 <!-- TOWNCRIER -->
 
+## [2.8.12] - 2021-10-21
+### Bugfixes
+- [#9771](https://github.com/rasahq/rasa/issues/9771): Fixed a bug where `rasa test --fail-on-prediction-errors` would raise a 
+  `WrongPredictionException` for entities which were actually predicted correctly.
+
+  This happened in two ways:
+  1. if for a user message some entities were extracted multiple times (by multiple entity
+  extractors) but listed only once in the test story,
+  2. if the order in which entities from a message were extracted didn't match the order
+  in which they were listed in the test story.
+
+### Improved Documentation
+- [#9691](https://github.com/rasahq/rasa/issues/9691): Improve the documentation for training `TEDPolicy` with data augmentation.
+
+
+## [2.8.11] - 2021-10-20
+### Bugfixes
+- [#9858](https://github.com/rasahq/rasa/issues/9858): Updates dependency on `sanic-jwt` (1.5.0 -> ">=1.6.0, <1.7.0")
+
+  This removes the need to pin the version of `pyjwt` as the newer version of `sanic-jwt`
+  manages this properly.
+
+
+## [2.8.10] - 2021-10-14
+### Bugfixes
+- [#5657](https://github.com/rasahq/rasa/issues/5657): Add List handling in the `send_custom_json` method on `channels/facebook.py`.
+  Bellow are some examples that could cause en error before.
+
+  Example 1: when the whole json is a List
+  ```
+  [
+      {
+          "blocks": {
+              "type": "progression_bar",
+              "text": {"text": "progression 1", "level": "1"},
+          }
+      },
+      {"sender": {"id": "example_id"}},
+  ]
+  ```
+
+  Example 2: instead of being a Dict, *blocks* is a List when there are 2 *type*
+  keys under it
+  ```
+  {
+      "blocks": [
+          {"type": "title", "text": {"text": "Conversation progress"}},
+          {
+              "type": "progression_bar",
+              "text": {"text": "Look how far we are...", "level": "1"},
+          },
+      ]
+  }
+  ```
+- [#7676](https://github.com/rasahq/rasa/issues/7676): Fixed bug when using wit.ai training data to train.
+  Training failed with an error similarly to this:
+
+  ```bash
+    File "./venv/lib/python3.8/site-packages/rasa/nlu/classifiers/diet_classifier.py", line 803, in train
+      self.check_correct_entity_annotations(training_data)
+    File "./venv/lib/python3.8/site-packages/rasa/nlu/extractors/extractor.py", line 418, in check_correct_entity_annotations
+      entities_repr = [
+    File "./venv/lib/python3.8/site-packages/rasa/nlu/extractors/extractor.py", line 422, in <listcomp>
+      entity[ENTITY_ATTRIBUTE_VALUE],
+  KeyError: 'value'
+  ```
+- [#9851](https://github.com/rasahq/rasa/issues/9851): Fix CVE-2021-41127
+
+
+## [2.8.9] - 2021-10-08
+### Improvements
+- [#7619](https://github.com/rasahq/rasa/issues/7619): Bump TensorFlow version to 2.6.
+
+  This update brings some security benefits (see TensorFlow 
+  [release notes](https://github.com/tensorflow/tensorflow/releases/tag/v2.6.0)
+  for details). However, internal experiments suggest that it is also associated with
+  increased train and inference time, as well as increased memory usage.
+
+  You can read more about why we decided to update TensorFlow, and what the expected 
+  impact is [here](https://rasa.com/blog/let-s-talk-about-tensorflow-2-6/).
+
+  If you experience a significant increase in train time, inference time, and/or memory 
+  usage, please let us know in the [forum](https://forum.rasa.com/t/feedback-upgrading-to-tensorflow-2-6/48331).
+
+  Users can no longer set `TF_DETERMINISTIC_OPS=1` if they are using GPU(s) because a 
+  `tf.errors.UnimplementedError` will be thrown by TensorFlow (read more
+  [here](https://github.com/tensorflow/tensorflow/releases/tag/v2.6.0)).
+
+  :::caution
+  This **breaks backward compatibility of previously trained models**.
+  It is not possible to load models trained with previous versions of Rasa Open Source. Please re-train
+  your assistant before trying to use this version.
+
+  :::
 ## [2.8.8] - 2021-10-06
 
 

Makefile

@@ -205,7 +205,7 @@ test-performance: test-marker
 test-marker: clean
     # OMP_NUM_THREADS can improve overall performance using one thread by process (on tensorflow), avoiding overload
 	# TF_CPP_MIN_LOG_LEVEL=2 sets C code log level for tensorflow to error suppressing lower log events
-	OMP_NUM_THREADS=1 TF_CPP_MIN_LOG_LEVEL=2 poetry run coverage run -m pytest tests -n $(JOBS) -m "$(PYTEST_MARKER)" --ignore $(INTEGRATION_TEST_FOLDER) $(DD_ARGS)
+	OMP_NUM_THREADS=1 TF_CPP_MIN_LOG_LEVEL=2 poetry run pytest tests -n $(JOBS) -m "$(PYTEST_MARKER)" --cov rasa --ignore $(INTEGRATION_TEST_FOLDER) $(DD_ARGS)
 
 generate-pending-changelog:
 	poetry run python -c "from scripts import release; release.generate_changelog('major.minor.patch')"

data/graph_schemas/config_pretrained_embeddings_mitie_predict_schema.yml

@@ -10,9 +10,9 @@ nodes:
     is_target: false
     is_input: false
     resource: null
-  run_MitieNLP0:
+  provide_MitieNLP0:
     needs: {}
-    uses: rasa.nlu.utils.mitie_utils.MitieNLPGraphComponent
+    uses: rasa.nlu.utils.mitie_utils.MitieNLP
     constructor_name: load
     fn: provide
     config:
@@ -24,7 +24,7 @@ nodes:
   run_MitieTokenizer1:
     needs:
       messages: nlu_message_converter
-    uses: rasa.nlu.tokenizers.mitie_tokenizer.MitieTokenizerGraphComponent
+    uses: rasa.nlu.tokenizers.mitie_tokenizer.MitieTokenizer
     constructor_name: load
     fn: process
     config: {}
@@ -35,8 +35,8 @@ nodes:
   run_MitieEntityExtractor2:
     needs:
       messages: run_MitieTokenizer1
-      model: run_MitieNLP0
-    uses: rasa.nlu.extractors.mitie_entity_extractor.MitieEntityExtractorGraphComponent
+      model: provide_MitieNLP0
+    uses: rasa.nlu.extractors.mitie_entity_extractor.MitieEntityExtractor
     constructor_name: load
     fn: process
     config: {}
@@ -48,7 +48,7 @@ nodes:
   run_EntitySynonymMapper3:
     needs:
       messages: run_MitieEntityExtractor2
-    uses: rasa.nlu.extractors.entity_synonyms.EntitySynonymMapperGraphComponent
+    uses: rasa.nlu.extractors.entity_synonyms.EntitySynonymMapper
     constructor_name: load
     fn: process
     config: {}
@@ -60,7 +60,7 @@ nodes:
   run_RegexFeaturizer4:
     needs:
       messages: run_EntitySynonymMapper3
-    uses: rasa.nlu.featurizers.sparse_featurizer.regex_featurizer.RegexFeaturizerGraphComponent
+    uses: rasa.nlu.featurizers.sparse_featurizer.regex_featurizer.RegexFeaturizer
     constructor_name: load
     fn: process
     config: {}
@@ -72,8 +72,8 @@ nodes:
   run_MitieFeaturizer5:
     needs:
       messages: run_RegexFeaturizer4
-      model: run_MitieNLP0
-    uses: rasa.nlu.featurizers.dense_featurizer.mitie_featurizer.MitieFeaturizerGraphComponent
+      model: provide_MitieNLP0
+    uses: rasa.nlu.featurizers.dense_featurizer.mitie_featurizer.MitieFeaturizer
     constructor_name: load
     fn: process
     config: {}
@@ -84,7 +84,7 @@ nodes:
   run_SklearnIntentClassifier6:
     needs:
       messages: run_MitieFeaturizer5
-    uses: rasa.nlu.classifiers.sklearn_intent_classifier.SklearnIntentClassifierGraphComponent
+    uses: rasa.nlu.classifiers.sklearn_intent_classifier.SklearnIntentClassifier
     constructor_name: load
     fn: process
     config: {}
@@ -93,10 +93,10 @@ nodes:
     is_input: false
     resource:
       name: train_SklearnIntentClassifier6
-  run_RegexMessageHandlerGraphComponent:
+  run_RegexMessageHandler:
     needs:
       messages: run_SklearnIntentClassifier6
-    uses: rasa.nlu.classifiers.regex_message_handler.RegexMessageHandlerGraphComponent
+    uses: rasa.nlu.classifiers.regex_message_handler.RegexMessageHandler
     constructor_name: load
     fn: process
     config: {}

data/graph_schemas/config_pretrained_embeddings_mitie_train_schema.yml

@@ -36,9 +36,9 @@ nodes:
     is_target: false
     is_input: true
     resource: null
-  run_MitieNLP0:
+  provide_MitieNLP0:
     needs: {}
-    uses: rasa.nlu.utils.mitie_utils.MitieNLPGraphComponent
+    uses: rasa.nlu.utils.mitie_utils.MitieNLP
     constructor_name: create
     fn: provide
     config:
@@ -50,7 +50,7 @@ nodes:
   run_MitieTokenizer1:
     needs:
       training_data: nlu_training_data_provider
-    uses: rasa.nlu.tokenizers.mitie_tokenizer.MitieTokenizerGraphComponent
+    uses: rasa.nlu.tokenizers.mitie_tokenizer.MitieTokenizer
     constructor_name: load
     fn: process_training_data
     config: {}
@@ -61,8 +61,8 @@ nodes:
   train_MitieEntityExtractor2:
     needs:
       training_data: run_MitieTokenizer1
-      model: run_MitieNLP0
-    uses: rasa.nlu.extractors.mitie_entity_extractor.MitieEntityExtractorGraphComponent
+      model: provide_MitieNLP0
+    uses: rasa.nlu.extractors.mitie_entity_extractor.MitieEntityExtractor
     constructor_name: create
     fn: train
     config: {}
@@ -73,7 +73,7 @@ nodes:
   train_EntitySynonymMapper3:
     needs:
       training_data: run_MitieTokenizer1
-    uses: rasa.nlu.extractors.entity_synonyms.EntitySynonymMapperGraphComponent
+    uses: rasa.nlu.extractors.entity_synonyms.EntitySynonymMapper
     constructor_name: create
     fn: train
     config: {}
@@ -84,7 +84,7 @@ nodes:
   train_RegexFeaturizer4:
     needs:
       training_data: run_MitieTokenizer1
-    uses: rasa.nlu.featurizers.sparse_featurizer.regex_featurizer.RegexFeaturizerGraphComponent
+    uses: rasa.nlu.featurizers.sparse_featurizer.regex_featurizer.RegexFeaturizer
     constructor_name: create
     fn: train
     config: {}
@@ -96,7 +96,7 @@ nodes:
     needs:
       training_data: run_MitieTokenizer1
       resource: train_RegexFeaturizer4
-    uses: rasa.nlu.featurizers.sparse_featurizer.regex_featurizer.RegexFeaturizerGraphComponent
+    uses: rasa.nlu.featurizers.sparse_featurizer.regex_featurizer.RegexFeaturizer
     constructor_name: load
     fn: process_training_data
     config: {}
@@ -107,8 +107,8 @@ nodes:
   run_MitieFeaturizer5:
     needs:
       training_data: run_RegexFeaturizer4
-      model: run_MitieNLP0
-    uses: rasa.nlu.featurizers.dense_featurizer.mitie_featurizer.MitieFeaturizerGraphComponent
+      model: provide_MitieNLP0
+    uses: rasa.nlu.featurizers.dense_featurizer.mitie_featurizer.MitieFeaturizer
     constructor_name: load
     fn: process_training_data
     config: {}
@@ -119,7 +119,7 @@ nodes:
   train_SklearnIntentClassifier6:
     needs:
       training_data: run_MitieFeaturizer5
-    uses: rasa.nlu.classifiers.sklearn_intent_classifier.SklearnIntentClassifierGraphComponent
+    uses: rasa.nlu.classifiers.sklearn_intent_classifier.SklearnIntentClassifier
     constructor_name: create
     fn: train
     config: {}