You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@beam.apache.org by "AnandInguva (via GitHub)" <gi...@apache.org> on 2023/02/09 00:46:47 UTC
[GitHub] [beam] AnandInguva opened a new pull request, #25393: Add WatchFilePattern
AnandInguva opened a new pull request, #25393:
URL: https://github.com/apache/beam/pull/25393
WatchFilePattern Ptransform.
------------------------
Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:
- [ ] Mention the appropriate issue in your description (for example: `addresses #123`), if applicable. This will automatically add a link to the pull request in the issue. If you would like the issue to automatically close on merging the pull request, comment `fixes #<ISSUE NUMBER>` instead.
- [ ] Update `CHANGES.md` with noteworthy changes.
- [ ] If this contribution is large, please file an Apache [Individual Contributor License Agreement](https://www.apache.org/licenses/icla.pdf).
See the [Contributor Guide](https://beam.apache.org/contribute) for more tips on [how to make review process smoother](https://beam.apache.org/contribute/get-started-contributing/#make-the-reviewers-job-easier).
To check the build health, please visit [https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md](https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md)
GitHub Actions Tests Status (on master branch)
------------------------------------------------------------------------------------------------
[](https://github.com/apache/beam/actions?query=workflow%3A%22Build+python+source+distribution+and+wheels%22+branch%3Amaster+event%3Aschedule)
[](https://github.com/apache/beam/actions?query=workflow%3A%22Python+Tests%22+branch%3Amaster+event%3Aschedule)
[](https://github.com/apache/beam/actions?query=workflow%3A%22Java+Tests%22+branch%3Amaster+event%3Aschedule)
[](https://github.com/apache/beam/actions?query=workflow%3A%22Go+tests%22+branch%3Amaster+event%3Aschedule)
See [CI.md](https://github.com/apache/beam/blob/master/CI.md) for more information about GitHub Actions CI.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1101891253
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -0,0 +1,205 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+
+"""
+A pipeline that uses RunInference PTransform to perform image classification
+and uses WatchFilePattern as side input to the RunInference PTransform.
+WatchFilePattern is used to watch for a file updates matching the file_pattern
+based on timestamps and emits latest model metadata, which is used in
+RunInference API for the dynamic model updates without the need for stopping
+the beam pipeline.
+
+This pipeline follows the pattern from
+https://beam.apache.org/documentation/patterns/side-inputs/
+
+This pipeline expects a PubSub topic as source, which emits an image
+path(UTF-8 encoded) that is accessible by the pipeline.
+
+To run the example on DataflowRunner,
+
+python apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py # pylint: disable=line-too-long
+ --project=<your-project>
+ --re=<your-region>
+ --temp_location=<your-tmp-location>
+ --staging_location=<your-staging-location>
+ --runner=DataflowRunner
+ --streaming
+ --interval=10
+ --num_workers=5
+ --requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+ --topic=<pubusb_topic>
+ --file_pattern=<glob_pattern>
Review Comment:
I added some information that would be relevant to this pipeline.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1424375452
R : @damccorm
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1426306169
I addressed almost all comments. PTAL @damccorm . Thanks for the suggestions
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1103204066
##########
sdks/python/apache_beam/ml/inference/utils.py:
##########
@@ -125,17 +125,21 @@ def __init__(
"""
Watches a directory for updates to files matching a given file pattern.
- **Note**: Start timestamp will be defaulted to timestamp when pipeline
- was run. All the files matching file_pattern, that are uploaded before
- the pipeline started will be discarded.
-
Args:
- file_pattern: A glob pattern used to watch a directory for model
- updates.
+ file_pattern: The file path to read from as a local file path or a
+ GCS ``gs://`` path. The path can contain glob characters
+ (``*``, ``?``, and ``[...]`` sets).
interval: Interval at which to check for files matching file_pattern
in seconds.
stop_timestamp: Timestamp after which no more files will be checked.
+ Constraints:
+ 1. If the file is read and then there is an update to that file, this
+ transform will ignore that update. Always update a file with unique
+ name.
Review Comment:
Changed the wording
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1424629528
TODO to myself: Change the typo after https://github.com/apache/beam/pull/25399 is merged
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] damccorm commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "damccorm (via GitHub)" <gi...@apache.org>.
damccorm commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1101675328
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -0,0 +1,205 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+
+"""
+A pipeline that uses RunInference PTransform to perform image classification
+and uses WatchFilePattern as side input to the RunInference PTransform.
+WatchFilePattern is used to watch for a file updates matching the file_pattern
+based on timestamps and emits latest model metadata, which is used in
+RunInference API for the dynamic model updates without the need for stopping
+the beam pipeline.
+
+This pipeline follows the pattern from
+https://beam.apache.org/documentation/patterns/side-inputs/
+
+This pipeline expects a PubSub topic as source, which emits an image
+path(UTF-8 encoded) that is accessible by the pipeline.
+
+To run the example on DataflowRunner,
+
+python apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py # pylint: disable=line-too-long
+ --project=<your-project>
+ --re=<your-region>
+ --temp_location=<your-tmp-location>
+ --staging_location=<your-staging-location>
+ --runner=DataflowRunner
+ --streaming
+ --interval=10
+ --num_workers=5
+ --requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+ --topic=<pubusb_topic>
+ --file_pattern=<glob_pattern>
Review Comment:
```suggestion
--file_pattern=<glob_pattern>
```
Could you add details here about how this file pattern is used (for updates) in a comment?
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -0,0 +1,205 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+
+"""
+A pipeline that uses RunInference PTransform to perform image classification
+and uses WatchFilePattern as side input to the RunInference PTransform.
+WatchFilePattern is used to watch for a file updates matching the file_pattern
+based on timestamps and emits latest model metadata, which is used in
+RunInference API for the dynamic model updates without the need for stopping
+the beam pipeline.
+
+This pipeline follows the pattern from
+https://beam.apache.org/documentation/patterns/side-inputs/
+
+This pipeline expects a PubSub topic as source, which emits an image
+path(UTF-8 encoded) that is accessible by the pipeline.
+
+To run the example on DataflowRunner,
+
+python apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py # pylint: disable=line-too-long
+ --project=<your-project>
+ --re=<your-region>
+ --temp_location=<your-tmp-location>
+ --staging_location=<your-staging-location>
+ --runner=DataflowRunner
+ --streaming
+ --interval=10
+ --num_workers=5
+ --requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+ --topic=<pubusb_topic>
+ --file_pattern=<glob_pattern>
+"""
+
+import argparse
+import io
+import logging
+import os
+from typing import Iterable
+from typing import Iterator
+from typing import Optional
+from typing import Tuple
+
+import apache_beam as beam
+import torch
+from apache_beam.io.filesystems import FileSystems
+from apache_beam.ml.inference.base import KeyedModelHandler
+from apache_beam.ml.inference.base import PredictionResult
+from apache_beam.ml.inference.base import RunInference
+from apache_beam.ml.inference.pytorch_inference import PytorchModelHandlerTensor
+from apache_beam.ml.inference.utils import WatchFilePattern
+from apache_beam.options.pipeline_options import PipelineOptions
+from apache_beam.options.pipeline_options import SetupOptions
+from apache_beam.runners.runner import PipelineResult
+from PIL import Image
+from torchvision import models
+from torchvision import transforms
+
+
+def read_image(image_file_name: str,
+ path_to_dir: Optional[str] = None) -> Tuple[str, Image.Image]:
+ if path_to_dir is not None:
+ image_file_name = os.path.join(path_to_dir, image_file_name)
+ with FileSystems().open(image_file_name, 'r') as file:
+ data = Image.open(io.BytesIO(file.read())).convert('RGB')
+ return image_file_name, data
+
+
+def preprocess_image(data: Image.Image) -> torch.Tensor:
+ image_size = (224, 224)
+ # Pre-trained PyTorch models expect input images normalized with the
+ # below values (see: https://pytorch.org/vision/stable/models.html)
+ normalize = transforms.Normalize(
+ mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+ transform = transforms.Compose([
+ transforms.Resize(image_size),
+ transforms.ToTensor(),
+ normalize,
+ ])
+ return transform(data)
+
+
+def filter_empty_lines(text: str) -> Iterator[str]:
+ if len(text.strip()) > 0:
+ yield text
+
+
+class PostProcessor(beam.DoFn):
+ """
+ Return filename, prediction and the model id used to perform the
+ prediction
+ """
+ def process(self, element: Tuple[str, PredictionResult]) -> Iterable[str]:
+ filename, prediction_result = element
+ prediction = torch.argmax(prediction_result.inference, dim=0)
+ yield filename, prediction, prediction_result.model_id
+
+
+def parse_known_args(argv):
+ """Parses args for the workflow."""
+ parser = argparse.ArgumentParser()
+ parser.add_argument(
+ '--topic',
+ dest='topic',
+ help='PubSub topic emitting absolute path to the images.'
+ 'Path must be accessible by the pipeline.')
+ parser.add_argument(
+ '--model_path',
Review Comment:
Nit: Maybe `initial_model_path`? Could we also update the description to say something like:
`Path to the initial model's state_dict. This will be used until the first model update occurs.`
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -0,0 +1,205 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+
+"""
+A pipeline that uses RunInference PTransform to perform image classification
+and uses WatchFilePattern as side input to the RunInference PTransform.
+WatchFilePattern is used to watch for a file updates matching the file_pattern
+based on timestamps and emits latest model metadata, which is used in
+RunInference API for the dynamic model updates without the need for stopping
+the beam pipeline.
+
+This pipeline follows the pattern from
+https://beam.apache.org/documentation/patterns/side-inputs/
+
+This pipeline expects a PubSub topic as source, which emits an image
+path(UTF-8 encoded) that is accessible by the pipeline.
+
+To run the example on DataflowRunner,
+
+python apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py # pylint: disable=line-too-long
+ --project=<your-project>
+ --re=<your-region>
+ --temp_location=<your-tmp-location>
+ --staging_location=<your-staging-location>
+ --runner=DataflowRunner
+ --streaming
+ --interval=10
+ --num_workers=5
+ --requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+ --topic=<pubusb_topic>
+ --file_pattern=<glob_pattern>
+"""
+
+import argparse
+import io
+import logging
+import os
+from typing import Iterable
+from typing import Iterator
+from typing import Optional
+from typing import Tuple
+
+import apache_beam as beam
+import torch
+from apache_beam.io.filesystems import FileSystems
+from apache_beam.ml.inference.base import KeyedModelHandler
+from apache_beam.ml.inference.base import PredictionResult
+from apache_beam.ml.inference.base import RunInference
+from apache_beam.ml.inference.pytorch_inference import PytorchModelHandlerTensor
+from apache_beam.ml.inference.utils import WatchFilePattern
+from apache_beam.options.pipeline_options import PipelineOptions
+from apache_beam.options.pipeline_options import SetupOptions
+from apache_beam.runners.runner import PipelineResult
+from PIL import Image
+from torchvision import models
+from torchvision import transforms
+
+
+def read_image(image_file_name: str,
+ path_to_dir: Optional[str] = None) -> Tuple[str, Image.Image]:
+ if path_to_dir is not None:
+ image_file_name = os.path.join(path_to_dir, image_file_name)
+ with FileSystems().open(image_file_name, 'r') as file:
+ data = Image.open(io.BytesIO(file.read())).convert('RGB')
+ return image_file_name, data
+
+
+def preprocess_image(data: Image.Image) -> torch.Tensor:
+ image_size = (224, 224)
+ # Pre-trained PyTorch models expect input images normalized with the
+ # below values (see: https://pytorch.org/vision/stable/models.html)
+ normalize = transforms.Normalize(
+ mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+ transform = transforms.Compose([
+ transforms.Resize(image_size),
+ transforms.ToTensor(),
+ normalize,
+ ])
+ return transform(data)
+
+
+def filter_empty_lines(text: str) -> Iterator[str]:
+ if len(text.strip()) > 0:
+ yield text
+
+
+class PostProcessor(beam.DoFn):
+ """
+ Return filename, prediction and the model id used to perform the
+ prediction
+ """
+ def process(self, element: Tuple[str, PredictionResult]) -> Iterable[str]:
+ filename, prediction_result = element
+ prediction = torch.argmax(prediction_result.inference, dim=0)
+ yield filename, prediction, prediction_result.model_id
+
+
+def parse_known_args(argv):
+ """Parses args for the workflow."""
+ parser = argparse.ArgumentParser()
+ parser.add_argument(
+ '--topic',
+ dest='topic',
+ help='PubSub topic emitting absolute path to the images.'
+ 'Path must be accessible by the pipeline.')
+ parser.add_argument(
+ '--model_path',
+ dest='model_path',
+ default='gs://apache-beam-samples/run_inference/resnet152.pth',
+ help="Path to the model's state_dict.")
+ parser.add_argument(
+ '--file_pattern', help='Glob pattern to watch for an update.')
+ parser.add_argument(
+ '--interval',
+ default=10,
+ type=int,
+ help='Interval used to check for file updates.')
+
+ return parser.parse_known_args(argv)
+
+
+def run(
+ argv=None,
+ model_class=None,
+ model_params=None,
+ save_main_session=True,
+ device='CPU',
+ test_pipeline=None) -> PipelineResult:
+ """
+ Args:
+ argv: Command line arguments defined for this example.
+ model_class: Reference to the class definition of the model.
+ model_params: Parameters passed to the constructor of the model_class.
+ These will be used to instantiate the model object in the
+ RunInference PTransform.
+ save_main_session: Used for internal testing.
+ device: Device to be used on the Runner. Choices are (CPU, GPU).
+ test_pipeline: Used for internal testing.
+ """
+ known_args, pipeline_args = parse_known_args(argv)
+ pipeline_options = PipelineOptions(pipeline_args)
+ pipeline_options.view_as(SetupOptions).save_main_session = save_main_session
+
+ if not model_class:
+ model_class = models.resnet152
+ model_params = {'num_classes': 1000}
+
+ class PytorchModelHandlerTensorWithBatchSize(PytorchModelHandlerTensor):
+ def batch_elements_kwargs(self):
Review Comment:
Can we use the batch config options now?
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -0,0 +1,205 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+
+"""
+A pipeline that uses RunInference PTransform to perform image classification
+and uses WatchFilePattern as side input to the RunInference PTransform.
+WatchFilePattern is used to watch for a file updates matching the file_pattern
+based on timestamps and emits latest model metadata, which is used in
+RunInference API for the dynamic model updates without the need for stopping
+the beam pipeline.
+
+This pipeline follows the pattern from
+https://beam.apache.org/documentation/patterns/side-inputs/
+
+This pipeline expects a PubSub topic as source, which emits an image
+path(UTF-8 encoded) that is accessible by the pipeline.
+
+To run the example on DataflowRunner,
+
+python apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py # pylint: disable=line-too-long
+ --project=<your-project>
+ --re=<your-region>
+ --temp_location=<your-tmp-location>
+ --staging_location=<your-staging-location>
+ --runner=DataflowRunner
+ --streaming
+ --interval=10
+ --num_workers=5
+ --requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+ --topic=<pubusb_topic>
Review Comment:
```suggestion
--topic=<pubsub_topic>
```
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -0,0 +1,205 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+
+"""
+A pipeline that uses RunInference PTransform to perform image classification
+and uses WatchFilePattern as side input to the RunInference PTransform.
+WatchFilePattern is used to watch for a file updates matching the file_pattern
+based on timestamps and emits latest model metadata, which is used in
+RunInference API for the dynamic model updates without the need for stopping
+the beam pipeline.
+
+This pipeline follows the pattern from
+https://beam.apache.org/documentation/patterns/side-inputs/
+
+This pipeline expects a PubSub topic as source, which emits an image
+path(UTF-8 encoded) that is accessible by the pipeline.
+
+To run the example on DataflowRunner,
+
+python apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py # pylint: disable=line-too-long
+ --project=<your-project>
+ --re=<your-region>
+ --temp_location=<your-tmp-location>
+ --staging_location=<your-staging-location>
+ --runner=DataflowRunner
+ --streaming
+ --interval=10
+ --num_workers=5
+ --requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+ --topic=<pubusb_topic>
+ --file_pattern=<glob_pattern>
+"""
+
+import argparse
+import io
+import logging
+import os
+from typing import Iterable
+from typing import Iterator
+from typing import Optional
+from typing import Tuple
+
+import apache_beam as beam
+import torch
+from apache_beam.io.filesystems import FileSystems
+from apache_beam.ml.inference.base import KeyedModelHandler
+from apache_beam.ml.inference.base import PredictionResult
+from apache_beam.ml.inference.base import RunInference
+from apache_beam.ml.inference.pytorch_inference import PytorchModelHandlerTensor
+from apache_beam.ml.inference.utils import WatchFilePattern
+from apache_beam.options.pipeline_options import PipelineOptions
+from apache_beam.options.pipeline_options import SetupOptions
+from apache_beam.runners.runner import PipelineResult
+from PIL import Image
+from torchvision import models
+from torchvision import transforms
+
+
+def read_image(image_file_name: str,
+ path_to_dir: Optional[str] = None) -> Tuple[str, Image.Image]:
+ if path_to_dir is not None:
+ image_file_name = os.path.join(path_to_dir, image_file_name)
+ with FileSystems().open(image_file_name, 'r') as file:
+ data = Image.open(io.BytesIO(file.read())).convert('RGB')
+ return image_file_name, data
+
+
+def preprocess_image(data: Image.Image) -> torch.Tensor:
+ image_size = (224, 224)
+ # Pre-trained PyTorch models expect input images normalized with the
+ # below values (see: https://pytorch.org/vision/stable/models.html)
+ normalize = transforms.Normalize(
+ mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+ transform = transforms.Compose([
+ transforms.Resize(image_size),
+ transforms.ToTensor(),
+ normalize,
+ ])
+ return transform(data)
+
+
+def filter_empty_lines(text: str) -> Iterator[str]:
+ if len(text.strip()) > 0:
+ yield text
+
+
+class PostProcessor(beam.DoFn):
+ """
+ Return filename, prediction and the model id used to perform the
+ prediction
+ """
+ def process(self, element: Tuple[str, PredictionResult]) -> Iterable[str]:
+ filename, prediction_result = element
+ prediction = torch.argmax(prediction_result.inference, dim=0)
+ yield filename, prediction, prediction_result.model_id
+
+
+def parse_known_args(argv):
+ """Parses args for the workflow."""
+ parser = argparse.ArgumentParser()
+ parser.add_argument(
+ '--topic',
+ dest='topic',
+ help='PubSub topic emitting absolute path to the images.'
+ 'Path must be accessible by the pipeline.')
+ parser.add_argument(
+ '--model_path',
+ dest='model_path',
+ default='gs://apache-beam-samples/run_inference/resnet152.pth',
+ help="Path to the model's state_dict.")
+ parser.add_argument(
+ '--file_pattern', help='Glob pattern to watch for an update.')
+ parser.add_argument(
+ '--interval',
+ default=10,
+ type=int,
+ help='Interval used to check for file updates.')
+
+ return parser.parse_known_args(argv)
+
+
+def run(
+ argv=None,
+ model_class=None,
+ model_params=None,
+ save_main_session=True,
+ device='CPU',
+ test_pipeline=None) -> PipelineResult:
+ """
+ Args:
+ argv: Command line arguments defined for this example.
+ model_class: Reference to the class definition of the model.
+ model_params: Parameters passed to the constructor of the model_class.
+ These will be used to instantiate the model object in the
+ RunInference PTransform.
+ save_main_session: Used for internal testing.
+ device: Device to be used on the Runner. Choices are (CPU, GPU).
+ test_pipeline: Used for internal testing.
+ """
+ known_args, pipeline_args = parse_known_args(argv)
+ pipeline_options = PipelineOptions(pipeline_args)
+ pipeline_options.view_as(SetupOptions).save_main_session = save_main_session
+
+ if not model_class:
+ model_class = models.resnet152
+ model_params = {'num_classes': 1000}
+
+ class PytorchModelHandlerTensorWithBatchSize(PytorchModelHandlerTensor):
+ def batch_elements_kwargs(self):
+ return {'min_batch_size': 10, 'max_batch_size': 100}
+
+ # In this example we pass keyed inputs to RunInference transform.
+ # Therefore, we use KeyedModelHandler wrapper over PytorchModelHandler.
+ model_handler = KeyedModelHandler(
+ PytorchModelHandlerTensorWithBatchSize(
+ state_dict_path=known_args.model_path,
+ model_class=model_class,
+ model_params=model_params,
+ device=device))
+
+ pipeline = test_pipeline
+ if not test_pipeline:
+ pipeline = beam.Pipeline(options=pipeline_options)
+
+ side_input = pipeline | WatchFilePattern(
+ interval=known_args.interval, file_pattern=known_args.file_pattern)
Review Comment:
This is awesome, just 2 lines and an additional arg to RunInference to do this!
##########
sdks/python/apache_beam/ml/inference/utils.py:
##########
@@ -46,3 +61,101 @@ def _convert_to_result(
y in zip(batch, predictions_per_tensor)
]
return [PredictionResult(x, y, model_id) for x, y in zip(batch, predictions)]
+
+
+class _ConvertIterToSingleton(beam.DoFn):
+ """
+ Internal only; No backwards compatibility.
+
+ The MatchContinuously transform examines all files present in a given
+ directory and returns those that have timestamps older than the
+ pipeline's start time. This can produce an Iterable rather than a
+ Singleton. This class only returns the file path when it is first
+ encountered, and it is cached as part of the side input caching mechanism.
+ If the path is seen again, it will not return anything.
+ By doing this, we can ensure that the output of this transform can be wrapped
+ with beam.pvalue.AsSingleton().
+ """
+ COUNT_STATE = CombiningValueStateSpec('count', combine_fn=sum)
+
+ def process(self, element, count_state=beam.DoFn.StateParam(COUNT_STATE)):
+ counter = count_state.read()
+ if counter == 0:
+ count_state.add(1)
+ yield element[1]
+
+
+class _GetLatestFileByTimeStamp(beam.DoFn):
+ """
+ Internal only; No backwards compatibility.
+
+ This DoFn checks the timestamps of files against the time that the pipeline
+ began running. It returns the files that were modified after the pipeline
+ started. If no such files are found, it returns a default file as fallback.
+ """
+ TIME_STATE = CombiningValueStateSpec(
+ 'count', combine_fn=partial(max, default=_START_TIME_STAMP))
Review Comment:
```suggestion
'max', combine_fn=partial(max, default=_START_TIME_STAMP))
```
Nit
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -0,0 +1,205 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+
+"""
+A pipeline that uses RunInference PTransform to perform image classification
+and uses WatchFilePattern as side input to the RunInference PTransform.
+WatchFilePattern is used to watch for a file updates matching the file_pattern
+based on timestamps and emits latest model metadata, which is used in
+RunInference API for the dynamic model updates without the need for stopping
+the beam pipeline.
+
+This pipeline follows the pattern from
+https://beam.apache.org/documentation/patterns/side-inputs/
+
+This pipeline expects a PubSub topic as source, which emits an image
+path(UTF-8 encoded) that is accessible by the pipeline.
Review Comment:
Could you add more detail on how this is emitted? Basically, something like "if you want to use this, publish messages to the topic specified in your args, formatted as a gcs string..."
##########
sdks/python/apache_beam/ml/inference/utils.py:
##########
@@ -46,3 +61,101 @@ def _convert_to_result(
y in zip(batch, predictions_per_tensor)
]
return [PredictionResult(x, y, model_id) for x, y in zip(batch, predictions)]
+
+
+class _ConvertIterToSingleton(beam.DoFn):
+ """
+ Internal only; No backwards compatibility.
+
+ The MatchContinuously transform examines all files present in a given
+ directory and returns those that have timestamps older than the
+ pipeline's start time. This can produce an Iterable rather than a
+ Singleton. This class only returns the file path when it is first
+ encountered, and it is cached as part of the side input caching mechanism.
+ If the path is seen again, it will not return anything.
+ By doing this, we can ensure that the output of this transform can be wrapped
+ with beam.pvalue.AsSingleton().
+ """
+ COUNT_STATE = CombiningValueStateSpec('count', combine_fn=sum)
+
+ def process(self, element, count_state=beam.DoFn.StateParam(COUNT_STATE)):
+ counter = count_state.read()
+ if counter == 0:
+ count_state.add(1)
+ yield element[1]
+
+
+class _GetLatestFileByTimeStamp(beam.DoFn):
+ """
+ Internal only; No backwards compatibility.
+
+ This DoFn checks the timestamps of files against the time that the pipeline
+ began running. It returns the files that were modified after the pipeline
+ started. If no such files are found, it returns a default file as fallback.
+ """
+ TIME_STATE = CombiningValueStateSpec(
+ 'count', combine_fn=partial(max, default=_START_TIME_STAMP))
+
+ def process(
+ self, element, time_state=beam.DoFn.StateParam(TIME_STATE)
+ ) -> List[Tuple[str, ModelMetdata]]:
+ _, file_metadata = element
+ new_ts = file_metadata.last_updated_in_seconds
+ old_ts = time_state.read()
+ if new_ts > old_ts:
+ # time_state.clear()
Review Comment:
```suggestion
```
Nit:
##########
sdks/python/apache_beam/ml/inference/utils.py:
##########
@@ -46,3 +61,101 @@ def _convert_to_result(
y in zip(batch, predictions_per_tensor)
]
return [PredictionResult(x, y, model_id) for x, y in zip(batch, predictions)]
+
+
+class _ConvertIterToSingleton(beam.DoFn):
+ """
+ Internal only; No backwards compatibility.
+
+ The MatchContinuously transform examines all files present in a given
+ directory and returns those that have timestamps older than the
+ pipeline's start time. This can produce an Iterable rather than a
+ Singleton. This class only returns the file path when it is first
+ encountered, and it is cached as part of the side input caching mechanism.
+ If the path is seen again, it will not return anything.
+ By doing this, we can ensure that the output of this transform can be wrapped
+ with beam.pvalue.AsSingleton().
+ """
+ COUNT_STATE = CombiningValueStateSpec('count', combine_fn=sum)
+
+ def process(self, element, count_state=beam.DoFn.StateParam(COUNT_STATE)):
+ counter = count_state.read()
+ if counter == 0:
+ count_state.add(1)
+ yield element[1]
+
+
+class _GetLatestFileByTimeStamp(beam.DoFn):
+ """
+ Internal only; No backwards compatibility.
+
+ This DoFn checks the timestamps of files against the time that the pipeline
+ began running. It returns the files that were modified after the pipeline
+ started. If no such files are found, it returns a default file as fallback.
+ """
+ TIME_STATE = CombiningValueStateSpec(
+ 'count', combine_fn=partial(max, default=_START_TIME_STAMP))
+
+ def process(
+ self, element, time_state=beam.DoFn.StateParam(TIME_STATE)
+ ) -> List[Tuple[str, ModelMetdata]]:
+ _, file_metadata = element
+ new_ts = file_metadata.last_updated_in_seconds
+ old_ts = time_state.read()
+ if new_ts > old_ts:
+ # time_state.clear()
+ time_state.add(new_ts)
+ model_path = file_metadata.path
+ else:
+ model_path = ''
+
+ model_name = os.path.splitext(os.path.basename(model_path))[0]
+ return [
+ (model_path, ModelMetdata(model_id=model_path, model_name=model_name))
+ ]
+
+
+class WatchFilePattern(beam.PTransform):
+ def __init__(
+ self,
+ file_pattern,
+ interval=360,
+ stop_timestamp=MAX_TIMESTAMP,
+ ):
+ """
+ Watches a directory for updates to files matching a given file pattern.
Review Comment:
We probably need a note in here about not having an empty file_pattern since we're using `EmptyMatchTreatment.DISALLOW`
##########
sdks/python/apache_beam/ml/inference/utils_test.py:
##########
@@ -0,0 +1,95 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+# pytype: skip-file
+
+import unittest
+
+import apache_beam as beam
+from apache_beam.io.filesystem import FileMetadata
+from apache_beam.ml.inference import utils
+from apache_beam.testing.test_pipeline import TestPipeline
+from apache_beam.testing.util import assert_that
+from apache_beam.testing.util import equal_to
+
+
+class WatchFilePatternTest(unittest.TestCase):
+ def test_latest_file_by_timestamp_default_value(self):
+ # match continuously returns the files in sorted timestamp order.
+ main_input_pcoll = [
+ FileMetadata(
+ 'path1.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP - 20),
+ FileMetadata(
+ 'path2.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP - 10)
+ ]
+ with TestPipeline() as p:
+ files_pc = (
+ p
+ | beam.Create(main_input_pcoll)
+ | beam.Map(lambda x: (x.path, x))
+ | beam.ParDo(utils._GetLatestFileByTimeStamp())
+ | beam.Map(lambda x: x[0]))
+ assert_that(files_pc, equal_to(['', '']))
+
+ def test_latest_file_with_timestamp_after_pipeline_construction_time(self):
+ main_input_pcoll = [
+ FileMetadata(
+ 'path1.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP + 10)
+ ]
+ with TestPipeline() as p:
+ files_pc = (
+ p
+ | beam.Create(main_input_pcoll)
+ | beam.Map(lambda x: (x.path, x))
+ | beam.ParDo(utils._GetLatestFileByTimeStamp())
+ | beam.Map(lambda x: x[0]))
+ assert_that(files_pc, equal_to(['path1.py']))
+
+ def test_emitting_singleton_output(self):
+ # match continuously returns the files in sorted timestamp order.
+ main_input_pcoll = [
+ FileMetadata(
+ 'path1.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP - 20),
+ # returns default
+ FileMetadata(
+ 'path2.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP - 10),
+ # returns default
+ FileMetadata(
+ 'path3.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP + 10)
Review Comment:
Could you add one more `FileMetadata` updated at `utils._START_TIME_STAMP + 20` just to exercise the multiple updates functionality (I don't see any reason it wouldn't equal `['', 'path3.py', 'path4.py']` but would like the extra validation + guardrails if we ever refactor
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1101891603
##########
sdks/python/apache_beam/ml/inference/utils.py:
##########
@@ -46,3 +61,101 @@ def _convert_to_result(
y in zip(batch, predictions_per_tensor)
]
return [PredictionResult(x, y, model_id) for x, y in zip(batch, predictions)]
+
+
+class _ConvertIterToSingleton(beam.DoFn):
+ """
+ Internal only; No backwards compatibility.
+
+ The MatchContinuously transform examines all files present in a given
+ directory and returns those that have timestamps older than the
+ pipeline's start time. This can produce an Iterable rather than a
+ Singleton. This class only returns the file path when it is first
+ encountered, and it is cached as part of the side input caching mechanism.
+ If the path is seen again, it will not return anything.
+ By doing this, we can ensure that the output of this transform can be wrapped
+ with beam.pvalue.AsSingleton().
+ """
+ COUNT_STATE = CombiningValueStateSpec('count', combine_fn=sum)
+
+ def process(self, element, count_state=beam.DoFn.StateParam(COUNT_STATE)):
+ counter = count_state.read()
+ if counter == 0:
+ count_state.add(1)
+ yield element[1]
+
+
+class _GetLatestFileByTimeStamp(beam.DoFn):
+ """
+ Internal only; No backwards compatibility.
+
+ This DoFn checks the timestamps of files against the time that the pipeline
+ began running. It returns the files that were modified after the pipeline
+ started. If no such files are found, it returns a default file as fallback.
+ """
+ TIME_STATE = CombiningValueStateSpec(
+ 'count', combine_fn=partial(max, default=_START_TIME_STAMP))
+
+ def process(
+ self, element, time_state=beam.DoFn.StateParam(TIME_STATE)
+ ) -> List[Tuple[str, ModelMetdata]]:
+ _, file_metadata = element
+ new_ts = file_metadata.last_updated_in_seconds
+ old_ts = time_state.read()
+ if new_ts > old_ts:
+ # time_state.clear()
+ time_state.add(new_ts)
+ model_path = file_metadata.path
+ else:
+ model_path = ''
+
+ model_name = os.path.splitext(os.path.basename(model_path))[0]
+ return [
+ (model_path, ModelMetdata(model_id=model_path, model_name=model_name))
+ ]
+
+
+class WatchFilePattern(beam.PTransform):
+ def __init__(
+ self,
+ file_pattern,
+ interval=360,
+ stop_timestamp=MAX_TIMESTAMP,
+ ):
+ """
+ Watches a directory for updates to files matching a given file pattern.
Review Comment:
Updated doc string
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1101825988
##########
sdks/python/apache_beam/ml/inference/utils_test.py:
##########
@@ -0,0 +1,95 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+# pytype: skip-file
+
+import unittest
+
+import apache_beam as beam
+from apache_beam.io.filesystem import FileMetadata
+from apache_beam.ml.inference import utils
+from apache_beam.testing.test_pipeline import TestPipeline
+from apache_beam.testing.util import assert_that
+from apache_beam.testing.util import equal_to
+
+
+class WatchFilePatternTest(unittest.TestCase):
+ def test_latest_file_by_timestamp_default_value(self):
+ # match continuously returns the files in sorted timestamp order.
+ main_input_pcoll = [
+ FileMetadata(
+ 'path1.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP - 20),
+ FileMetadata(
+ 'path2.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP - 10)
+ ]
+ with TestPipeline() as p:
+ files_pc = (
+ p
+ | beam.Create(main_input_pcoll)
+ | beam.Map(lambda x: (x.path, x))
+ | beam.ParDo(utils._GetLatestFileByTimeStamp())
+ | beam.Map(lambda x: x[0]))
+ assert_that(files_pc, equal_to(['', '']))
+
+ def test_latest_file_with_timestamp_after_pipeline_construction_time(self):
+ main_input_pcoll = [
+ FileMetadata(
+ 'path1.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP + 10)
+ ]
+ with TestPipeline() as p:
+ files_pc = (
+ p
+ | beam.Create(main_input_pcoll)
+ | beam.Map(lambda x: (x.path, x))
+ | beam.ParDo(utils._GetLatestFileByTimeStamp())
+ | beam.Map(lambda x: x[0]))
+ assert_that(files_pc, equal_to(['path1.py']))
+
+ def test_emitting_singleton_output(self):
+ # match continuously returns the files in sorted timestamp order.
+ main_input_pcoll = [
+ FileMetadata(
+ 'path1.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP - 20),
+ # returns default
+ FileMetadata(
+ 'path2.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP - 10),
+ # returns default
+ FileMetadata(
+ 'path3.py',
+ 10,
+ last_updated_in_seconds=utils._START_TIME_STAMP + 10)
Review Comment:
Added
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] damccorm commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "damccorm (via GitHub)" <gi...@apache.org>.
damccorm commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1103215023
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -0,0 +1,218 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements. See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You 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.
+#
+
+"""
+A pipeline that uses RunInference PTransform to perform image classification
+and uses WatchFilePattern as side input to the RunInference PTransform.
+WatchFilePattern is used to watch for a file updates matching the file_pattern
+based on timestamps and emits latest model metadata, which is used in
+RunInference API for the dynamic model updates without the need for stopping
+the beam pipeline.
+
+This pipeline follows the pattern from
+https://beam.apache.org/documentation/patterns/side-inputs/
+
+To use the PubSub reading from a topic in the pipeline as source, you can
+publish a path to the model(resnet152 used in the pipeline from
+torchvision.models.resnet152) to the PubSub topic. Then pass that
+topic via command line arg --topic. The published path(str) should be
+UTF-8 encoded.
+
+To run the example on DataflowRunner,
+
+python apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py # pylint: disable=line-too-long
+ --project=<your-project>
+ --re=<your-region>
+ --temp_location=<your-tmp-location>
+ --staging_location=<your-staging-location>
+ --runner=DataflowRunner
+ --streaming
+ --interval=10
+ --num_workers=5
+ --requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+ --topic=<pubusb_topic>
Review Comment:
```suggestion
--topic=<pubsub_topic>
```
Looks like this one got missed
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] github-actions[bot] commented on pull request #25393: Add WatchFilePattern
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1424669833
Assigning reviewers. If you would like to opt out of this review, comment `assign to next reviewer`:
R: @jrmccluskey for label python.
R: @ahmedabu98 for label io.
Available commands:
- `stop reviewer notifications` - opt out of the automated review tooling
- `remind me after tests pass` - tag the comment author after tests pass
- `waiting on author` - shift the attention set back to the author (any comment or push by the author will return the attention set to the reviewers)
The PR bot will only process comments in the main thread (not review comments).
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] damccorm commented on pull request #25393: Add WatchFilePattern
Posted by "damccorm (via GitHub)" <gi...@apache.org>.
damccorm commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1424749580
> TODO to myself: Change the typo after #25399 is merged
Merged - sorry for the delay, had to retry some precommit checks
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1103204646
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -26,8 +26,11 @@
This pipeline follows the pattern from
https://beam.apache.org/documentation/patterns/side-inputs/
-This pipeline expects a PubSub topic as source, which emits an image
-path(UTF-8 encoded) that is accessible by the pipeline.
+To use the PubSub reading from a topic in the pipeline as source, you can
+publish a path to the model(resnet152 used in the pipeline from
+torchvision.models.resnet152) to the PubSub topic. Then pass that
+topic via command line arg --topic. The published path(str) should be
+UTF-8 encoded.
Review Comment:
Yes but in WatchTransform doc string. corrected it
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] codecov[bot] commented on pull request #25393: Add WatchFilePattern
Posted by "codecov[bot] (via GitHub)" <gi...@apache.org>.
codecov[bot] commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1423464195
# [Codecov](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=h1&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) Report
> Merging [#25393](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) (9394f03) into [master](https://codecov.io/gh/apache/beam/commit/9fcb3a5b48a6db0dcf57f454d1d1eca10cf1c41b?el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) (9fcb3a5) will **decrease** coverage by `0.05%`.
> The diff coverage is `32.45%`.
```diff
@@ Coverage Diff @@
## master #25393 +/- ##
==========================================
- Coverage 72.91% 72.86% -0.05%
==========================================
Files 746 747 +1
Lines 99267 99381 +114
==========================================
+ Hits 72380 72415 +35
- Misses 25522 25601 +79
Partials 1365 1365
```
| Flag | Coverage Δ | |
|---|---|---|
| python | `82.30% <32.45%> (-0.09%)` | :arrow_down: |
Flags with carried forward coverage won't be shown. [Click here](https://docs.codecov.io/docs/carryforward-flags?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#carryforward-flags-in-the-pull-request-comment) to find out more.
| [Impacted Files](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) | Coverage Δ | |
|---|---|---|
| [...e/pytorch\_image\_classification\_with\_side\_inputs.py](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c2Rrcy9weXRob24vYXBhY2hlX2JlYW0vZXhhbXBsZXMvaW5mZXJlbmNlL3B5dG9yY2hfaW1hZ2VfY2xhc3NpZmljYXRpb25fd2l0aF9zaWRlX2lucHV0cy5weQ==) | `0.00% <0.00%> (ø)` | |
| [sdks/python/apache\_beam/ml/inference/utils.py](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c2Rrcy9weXRob24vYXBhY2hlX2JlYW0vbWwvaW5mZXJlbmNlL3V0aWxzLnB5) | `92.30% <90.00%> (-7.70%)` | :arrow_down: |
| [sdks/python/apache\_beam/io/fileio.py](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c2Rrcy9weXRob24vYXBhY2hlX2JlYW0vaW8vZmlsZWlvLnB5) | `96.11% <100.00%> (+0.01%)` | :arrow_up: |
| [...eam/runners/portability/fn\_api\_runner/execution.py](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c2Rrcy9weXRob24vYXBhY2hlX2JlYW0vcnVubmVycy9wb3J0YWJpbGl0eS9mbl9hcGlfcnVubmVyL2V4ZWN1dGlvbi5weQ==) | `92.49% <0.00%> (-0.64%)` | :arrow_down: |
| [...hon/apache\_beam/runners/worker/bundle\_processor.py](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c2Rrcy9weXRob24vYXBhY2hlX2JlYW0vcnVubmVycy93b3JrZXIvYnVuZGxlX3Byb2Nlc3Nvci5weQ==) | `93.37% <0.00%> (-0.13%)` | :arrow_down: |
| [...ks/python/apache\_beam/runners/worker/sdk\_worker.py](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c2Rrcy9weXRob24vYXBhY2hlX2JlYW0vcnVubmVycy93b3JrZXIvc2RrX3dvcmtlci5weQ==) | `89.24% <0.00%> (+0.16%)` | :arrow_up: |
| [...python/apache\_beam/runners/worker/worker\_status.py](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c2Rrcy9weXRob24vYXBhY2hlX2JlYW0vcnVubmVycy93b3JrZXIvd29ya2VyX3N0YXR1cy5weQ==) | `75.33% <0.00%> (+0.66%)` | :arrow_up: |
| [.../apache\_beam/runners/interactive/dataproc/types.py](https://codecov.io/gh/apache/beam/pull/25393?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-c2Rrcy9weXRob24vYXBhY2hlX2JlYW0vcnVubmVycy9pbnRlcmFjdGl2ZS9kYXRhcHJvYy90eXBlcy5weQ==) | `96.55% <0.00%> (+3.44%)` | :arrow_up: |
:mega: We’re building smart automated test selection to slash your CI/CD build times. [Learn more](https://about.codecov.io/iterative-testing/?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1103204387
##########
sdks/python/apache_beam/ml/inference/utils.py:
##########
@@ -94,7 +94,7 @@ class _GetLatestFileByTimeStamp(beam.DoFn):
started. If no such files are found, it returns a default file as fallback.
"""
TIME_STATE = CombiningValueStateSpec(
- 'count', combine_fn=partial(max, default=_START_TIME_STAMP))
+ 'max', combine_fn=partial(max, default=_START_TIME_STAMP))
Review Comment:
Lint is passing now. It is not related to this change.
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -26,8 +26,11 @@
This pipeline follows the pattern from
https://beam.apache.org/documentation/patterns/side-inputs/
-This pipeline expects a PubSub topic as source, which emits an image
-path(UTF-8 encoded) that is accessible by the pipeline.
+To use the PubSub reading from a topic in the pipeline as source, you can
+publish a path to the model(resnet152 used in the pipeline from
+torchvision.models.resnet152) to the PubSub topic. Then pass that
+topic via command line arg --topic. The published path(str) should be
+UTF-8 encoded.
Review Comment:
Yes, corrected it
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] damccorm merged pull request #25393: Add WatchFilePattern
Posted by "damccorm (via GitHub)" <gi...@apache.org>.
damccorm merged PR #25393:
URL: https://github.com/apache/beam/pull/25393
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1426232326
Run Whitespace PreCommit
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1426232132
Run PythonDocs PreCommit
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] AnandInguva commented on pull request #25393: Add WatchFilePattern
Posted by "AnandInguva (via GitHub)" <gi...@apache.org>.
AnandInguva commented on PR #25393:
URL: https://github.com/apache/beam/pull/25393#issuecomment-1424338083
Run PythonLint PreCommit
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [beam] damccorm commented on a diff in pull request #25393: Add WatchFilePattern
Posted by "damccorm (via GitHub)" <gi...@apache.org>.
damccorm commented on code in PR #25393:
URL: https://github.com/apache/beam/pull/25393#discussion_r1101933389
##########
sdks/python/apache_beam/ml/inference/utils.py:
##########
@@ -94,7 +94,7 @@ class _GetLatestFileByTimeStamp(beam.DoFn):
started. If no such files are found, it returns a default file as fallback.
"""
TIME_STATE = CombiningValueStateSpec(
- 'count', combine_fn=partial(max, default=_START_TIME_STAMP))
+ 'max', combine_fn=partial(max, default=_START_TIME_STAMP))
Review Comment:
Seems like the linter doesn't like this change I suggested (count -> max) - if that's the case, feel free to switch back
##########
sdks/python/apache_beam/ml/inference/utils.py:
##########
@@ -125,17 +125,21 @@ def __init__(
"""
Watches a directory for updates to files matching a given file pattern.
- **Note**: Start timestamp will be defaulted to timestamp when pipeline
- was run. All the files matching file_pattern, that are uploaded before
- the pipeline started will be discarded.
-
Args:
- file_pattern: A glob pattern used to watch a directory for model
- updates.
+ file_pattern: The file path to read from as a local file path or a
+ GCS ``gs://`` path. The path can contain glob characters
+ (``*``, ``?``, and ``[...]`` sets).
interval: Interval at which to check for files matching file_pattern
in seconds.
stop_timestamp: Timestamp after which no more files will be checked.
+ Constraints:
+ 1. If the file is read and then there is an update to that file, this
+ transform will ignore that update. Always update a file with unique
+ name.
Review Comment:
This doesn't quite convey that the following won't work:
```
- Create model with name A (model A is now used)
- Create model with name B (model B is now used)
- Delete file with name A
- Create new model with name A (model A is still not used because of stateful DoFn)
```
Could we update to something like: `Any previously used filenames cannot be reused. If a file is added or updated to a previously used filename, this transform will ignore that update. To trigger a model update, always upload a file with unique name.`
(tried to keep as much of your wording as possible, feel free to reword if you don't like my phrasing).
##########
sdks/python/apache_beam/examples/inference/pytorch_image_classification_with_side_inputs.py:
##########
@@ -26,8 +26,11 @@
This pipeline follows the pattern from
https://beam.apache.org/documentation/patterns/side-inputs/
-This pipeline expects a PubSub topic as source, which emits an image
-path(UTF-8 encoded) that is accessible by the pipeline.
+To use the PubSub reading from a topic in the pipeline as source, you can
+publish a path to the model(resnet152 used in the pipeline from
+torchvision.models.resnet152) to the PubSub topic. Then pass that
+topic via command line arg --topic. The published path(str) should be
+UTF-8 encoded.
Review Comment:
Looks like docs precommit doesn't like some of your indentation here
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@beam.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org