Rename nlp example (#2221)

* rename nlp case

* rename to e2e_nlp

* Add HuggingFace and Skypilot deployment steps

.github/workflows/update-templates-to-examples.yml

@@ -169,7 +169,7 @@ jobs:
               issue_number: ${{ github.event.pull_request.number }},
               owner: context.repo.owner,
               repo: context.repo.repo,
-              body: 'NLP template updates in `examples/nlp-case` have been pushed.'
+              body: 'NLP template updates in `examples/e2e_nlp` have been pushed.'
             })
   update-starter-template-to-examples:
     name: update-starter-template-to-examples

examples/nlp-case/.copier-answers.yml → examples/e2e_nlp/.copier-answers.yml

@@ -1,12 +1,12 @@
 # Changes here will be overwritten by Copier
-_commit: 0.45.0
-_src_path: gh:zenml-io/template-nlp
+_commit: 0.45.0-17-gbd996ed
+_src_path: template-nlp
 accelerator: cpu
 cloud_of_choice: aws
 dataset: airline_reviews
 deploy_locally: true
-deploy_to_huggingface: false
-deploy_to_skypilot: false
+deploy_to_huggingface: true
+deploy_to_skypilot: true
 email: ''
 full_name: ZenML GmbH
 metric_compare_promotion: true

examples/nlp-case/README.md → examples/e2e_nlp/README.md

@@ -1,11 +1,14 @@
 # ZenML NLP project
 
 This is a comprehensive supervised ML project built with the
-ZenML framework and its integration. The project trains one or more
-scikit-learn classification models to make predictions on the tabular
-classification datasets provided by the scikit-learn library. The project was
-generated from the [NLP ZenML project template](https://github.com/zenml-io/template-nlp).
+ZenML framework and its integration. The project will be 
+a comprehensive starting point for anyone looking to build
+and deploy NLP models using the ZenML framework by streamlining
+the process of training, promoting, and deploying NLP models
+with a focus on reproducibility, scalability, and ease of use.
+The project was generated from the [NLP ZenML project template](https://github.com/zenml-io/template-nlp).
 with the following properties:
+
 - Project name: ZenML NLP project
 - Technical Name: nlp_use_case
 - Version: `0.0.1`
@@ -16,6 +19,8 @@ Settings of your project are:
 - Accelerator: `cpu`
 - Trained model promotion to `staging` based on accuracy metric vs currently deployed model
 - Local deployment enabled
+- Deployment to HuggingFace Hub enabled
+- Deployment to SkyPilot enabled
 - Dataset: `airline_reviews`
 - Model: `distilbert-base-uncased`
 - Notifications about failures enabled
@@ -181,20 +186,3 @@ The project loosely follows [the recommended ZenML project structure](https://do
 ├── config.yaml               # ZenML configuration file
 └── run.py                    # CLI tool to run pipelines on ZenML Stack
 ```
-
-
-
-
-
-
-
-
-
-
-
-
-integrations
-├── README.md                 # this file
-├── requirements.txt          # extra Python dependencies 
-└── run.py                    # CLI tool to run pipelines on ZenML Stack
-```

examples/nlp-case/pipelines/deploying.py → examples/e2e_nlp/pipelines/deploying.py

@@ -19,6 +19,8 @@
 
 from steps import (
     deploy_locally,
+    deploy_to_huggingface,
+    deploy_to_skypilot,
     notify_on_failure,
     notify_on_success,
     save_model_to_deploy,
@@ -87,7 +89,16 @@ def nlp_use_case_deploy_pipeline(
         tokenizer_name_or_path=tokenizer_name_or_path,
         after=["save_model_to_deploy"],
     )
-    last_step_name = "deploy_locally"
+    ########## Deploy to HuggingFace ##########
+    deploy_to_huggingface(
+        repo_name=repo_name,
+        after=["save_model_to_deploy"],
+    )
+    ########## Deploy to Skypilot ##########
+    deploy_to_skypilot(
+        after=["save_model_to_deploy"],
+    )
+    last_step_name = "deploy_to_skypilot"
 
     notify_on_success(after=[last_step_name])
     ### YOUR CODE ENDS HERE ###

examples/nlp-case/steps/__init__.py → examples/e2e_nlp/steps/__init__.py

@@ -36,4 +36,6 @@
 from .deploying import (
     save_model_to_deploy,
     deploy_locally,
+    deploy_to_huggingface,
+    deploy_to_skypilot,
 )

examples/nlp-case/steps/deploying/__init__.py → examples/e2e_nlp/steps/deploying/__init__.py

@@ -19,3 +19,9 @@
 from .save_model import save_model_to_deploy
 
 from .local_deployment import deploy_locally
+
+
+from .huggingface_deployment import deploy_to_huggingface
+
+
+from .skypilot_deployment import deploy_to_skypilot

examples/e2e_nlp/steps/deploying/huggingface_deployment.py

@@ -0,0 +1,63 @@
+# Apache Software License 2.0
+#
+# Copyright (c) ZenML GmbH 2024. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+import os
+
+from huggingface_hub import HfApi
+
+from zenml import step
+from zenml.client import Client
+from zenml.logger import get_logger
+
+# Initialize logger
+logger = get_logger(__name__)
+
+
+@step()
+def deploy_to_huggingface(
+    repo_name: str,
+):
+    """
+    This step deploy the model to huggingface.
+
+    Args:
+        repo_name: The name of the repo to create/use on huggingface.
+    """
+    ### ADD YOUR OWN CODE HERE - THIS IS JUST AN EXAMPLE ###
+    secret = Client().get_secret("huggingface_creds")
+    assert secret, "No secret found with name 'huggingface_creds'. Please create one that includes your `username` and `token`."
+    token = secret.secret_values["token"]
+    api = HfApi(token=token)
+    hf_repo = api.create_repo(
+        repo_id=repo_name, repo_type="space", space_sdk="gradio", exist_ok=True
+    )
+    zenml_repo_root = Client().root
+    if not zenml_repo_root:
+        logger.warning(
+            "You're running the `deploy_to_huggingface` step outside of a ZenML repo. "
+            "Since the deployment step to huggingface is all about pushing the repo to huggingface, "
+            "this step will not work outside of a ZenML repo where the gradio folder is present."
+        )
+        raise
+    gradio_folder_path = os.path.join(zenml_repo_root, "gradio")
+    space = api.upload_folder(
+        folder_path=gradio_folder_path,
+        repo_id=hf_repo.repo_id,
+        repo_type="space",
+    )
+    logger.info(f"Space created: {space}")
+    ### YOUR CODE ENDS HERE ###

examples/e2e_nlp/steps/deploying/skypilot_deployment.py

@@ -0,0 +1,53 @@
+# Apache Software License 2.0
+#
+# Copyright (c) ZenML GmbH 2024. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+
+import os
+import re
+
+import sky
+
+from zenml import step
+from zenml.client import Client
+from zenml.logger import get_logger
+
+# Initialize logger
+logger = get_logger(__name__)
+
+
+@step()
+def deploy_to_skypilot():
+    """
+    This step deploy the model to a VM using SkyPilot.
+
+    This step requires `skypilot` to be installed.
+    aswell as a configured cloud account locally (e.g. AWS, GCP, Azure).
+    """
+    ### ADD YOUR OWN CODE HERE - THIS IS JUST AN EXAMPLE ###
+    zenml_repo_root = Client().root
+    if not zenml_repo_root:
+        logger.warning(
+            "You're running the `deploy_to_huggingface` step outside of a ZenML repo. "
+            "Since the deployment step to huggingface is all about pushing the repo to huggingface, "
+            "this step will only work within a ZenML repo where the gradio folder is present."
+        )
+        raise
+    gradio_task_yaml = os.path.join(zenml_repo_root, "gradio", "serve.yaml")
+    task = sky.Task.from_yaml(gradio_task_yaml)
+    cluster_name = re.sub(r"[^a-zA-Z0-9]+", "-", "nlp_use_case-cluster")
+    sky.launch(task, cluster_name=cluster_name, detach_setup=True)
+    ### YOUR CODE ENDS HERE ###