CI: case excluding CPU (#158)

* skip pip for a dry run
* cleaning downloads
* adding gpu ipynb
* unzip progress
* devel requirements

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: Borda

.actions/assistant.py

@@ -20,6 +20,8 @@
 PATH_REQ_DEFAULT = os.path.join(_PATH_ROOT, "requirements", "default.txt")
 PATH_SCRIPT_RENDER = os.path.join(_PATH_HERE, "_ipynb-render.sh")
 PATH_SCRIPT_TEST = os.path.join(_PATH_HERE, "_ipynb-test.sh")
+# https://askubuntu.com/questions/909918/how-to-show-unzip-progress
+UNZIP_PROGRESS_BAR = ' | awk \'BEGIN {ORS=" "} {if(NR%10==0)print "."}\''
 REPO_NAME = "lightning-tutorials"
 COLAB_REPO_LINK = "https://colab.research.google.com/github/PytorchLightning"
 BRANCH_DEFAULT = "main"
@@ -136,8 +138,9 @@ def get_running_torch_version():
 class AssistantCLI:
     """Collection of handy CLI commands."""
 
-    DEVICE_ACCELERATOR = os.environ.get("ACCELERATOR", "cpu").lower()
-    DATASET_FOLDER = os.environ.get("PATH_DATASETS", "_datasets").lower()
+    _LOCAL_ACCELERATOR = "cpu,gpu" if get_running_cuda_version() else "cpu"
+    DEVICE_ACCELERATOR = os.environ.get("ACCELERATOR", _LOCAL_ACCELERATOR).lower()
+    DATASETS_FOLDER = os.environ.get("PATH_DATASETS", "_datasets")
     DRY_RUN = bool(int(os.environ.get("DRY_RUN", 0)))
     _META_REQUIRED_FIELDS = ("title", "author", "license", "description")
     _SKIP_DIRS = (
@@ -152,7 +155,7 @@ class AssistantCLI:
     )
     _META_FILE_REGEX = ".meta.{yaml,yml}"
     _META_PIP_KEY = "pip__"
-    _META_ACCEL_DEFAULT = ("CPU",)
+    _META_ACCEL_DEFAULT = _LOCAL_ACCELERATOR.split(",")
 
     # Map directory names to tag names. Note that dashes will be replaced with spaces in rendered tags in the docs.
     _DIR_TO_TAG = {
@@ -268,16 +271,15 @@ def _parse_requirements(folder: str) -> Tuple[str, str]:
 
     @staticmethod
     def _bash_download_data(folder: str) -> List[str]:
-        """Generate sequence of commands fro optional downloading dataset specified in the meta file.
+        """Generate sequence of commands for optional downloading dataset specified in the meta file.
 
         Args:
             folder: path to the folder with python script, meta and artefacts
         """
-        cmd = ["HERE=$PWD", f"cd {AssistantCLI.DATASET_FOLDER}"]
         meta = AssistantCLI._load_meta(folder)
         datasets = meta.get("datasets", {})
         data_kaggle = datasets.get("kaggle", [])
-        cmd += [f"python -m kaggle competitions download -c {name}" for name in data_kaggle]
+        cmd = [f"python -m kaggle competitions download -c {name}" for name in data_kaggle]
         files = [f"{name}.zip" for name in data_kaggle]
         data_web = datasets.get("web", [])
         cmd += [f"wget {web} --progress=bar:force:noscroll --tries=3" for web in data_web]
@@ -287,11 +289,11 @@ def _bash_download_data(folder: str) -> List[str]:
             if ext not in AssistantCLI._EXT_ARCHIVE:
                 continue
             if ext in AssistantCLI._EXT_ARCHIVE_ZIP:
-                cmd += [f"mkdir -p {name}", f"unzip -o {fn} -d {name}"]
+                cmd += [f"unzip -o {fn} -d {AssistantCLI.DATASETS_FOLDER}/{name} {UNZIP_PROGRESS_BAR}"]
             else:
                 cmd += [f"tar -zxvf {fn} --overwrite"]
             cmd += [f"rm {fn}"]
-        cmd += ["ls -l", "cd $HERE"]
+        cmd += [f"tree -L 2 {AssistantCLI.DATASETS_FOLDER}"]
         return cmd
 
     @staticmethod
@@ -310,26 +312,29 @@ def bash_render(folder: str, output_file: str = PATH_SCRIPT_RENDER) -> Optional[
             cmd += AssistantCLI._bash_download_data(folder)
         ipynb_file, meta_file, thumb_file = AssistantCLI._valid_folder(folder, ext=".ipynb")
         pub_ipynb = os.path.join(DIR_NOTEBOOKS, f"{folder}.ipynb")
+        pub_meta = pub_ipynb.replace(".ipynb", ".yaml")
         pub_dir = os.path.dirname(pub_ipynb)
         thumb_ext = os.path.splitext(thumb_file)[-1] if thumb_file else "."
         pub_thumb = os.path.join(DIR_NOTEBOOKS, f"{folder}{thumb_ext}") if thumb_file else ""
         cmd.append(f"mkdir -p {pub_dir}")
-        pip_req, pip_args = AssistantCLI._parse_requirements(folder)
-        cmd += [f"pip install {pip_req} {pip_args}", "pip list"]
         if AssistantCLI.DRY_RUN:
             # dry run does not execute the notebooks just takes them as they are
             cmd.append(f"cp {ipynb_file} {pub_ipynb}")
+            # copy and add meta config
+            cmd += [f"cp {meta_file} {pub_meta}", f"cat {pub_meta}", f"git add {pub_meta}"]
         else:
+            pip_req, pip_args = AssistantCLI._parse_requirements(folder)
+            cmd += [f"pip install {pip_req} --quiet {pip_args}", "pip list"]
             cmd.append(f"# available: {AssistantCLI.DEVICE_ACCELERATOR}\n")
             if AssistantCLI._valid_accelerator(folder):
                 cmd.append(f"python -m papermill {ipynb_file} {pub_ipynb} --kernel python")
             else:
                 warn("Invalid notebook's accelerator for this device. So no outputs will be generated.", RuntimeWarning)
                 cmd.append(f"cp {ipynb_file} {pub_ipynb}")
-        # Export the actual packages used in runtime
-        cmd.append(f"meta_file=$(python .actions/assistant.py update-env-details {folder})")
-        # copy and add to version the enriched meta config
-        cmd += ["echo $meta_file", "cat $meta_file", "git add $meta_file"]
+            # Export the actual packages used in runtime
+            cmd.append(f"meta_file=$(python .actions/assistant.py update-env-details {folder})")
+            # copy and add to version the enriched meta config
+            cmd += ["echo $meta_file", "cat $meta_file", "git add $meta_file"]
         # if thumb image is linked to the notebook, copy and version it too
         if thumb_file:
             cmd += [f"cp {thumb_file} {pub_thumb}", f"git add {pub_thumb}"]
@@ -353,26 +358,36 @@ def bash_test(folder: str, output_file: str = PATH_SCRIPT_TEST) -> Optional[str]
         """
         cmd = list(AssistantCLI._BASH_SCRIPT_BASE) + [f"# Testing: {folder}"]
         cmd += AssistantCLI._bash_download_data(folder)
-        ipynb_file, _, _ = AssistantCLI._valid_folder(folder, ext=".ipynb")
+        ipynb_file, meta_file, _ = AssistantCLI._valid_folder(folder, ext=".ipynb")
 
         # prepare isolated environment with inheriting the global packages
+        path_venv = os.path.join(folder, "venv")
         cmd += [
-            f"python -m virtualenv --system-site-packages {os.path.join(folder, 'venv')}",
-            f"source {os.path.join(folder, 'venv', 'bin', 'activate')}",
+            f"python -m virtualenv --system-site-packages {path_venv}",
+            f"source {os.path.join(path_venv, 'bin', 'activate')}",
             "pip --version",
         ]
-        # and install specific packages
-        pip_req, pip_args = AssistantCLI._parse_requirements(folder)
-        cmd += [f"pip install {pip_req} {pip_args}", "pip list"]
-        # Export the actual packages used in runtime
-        cmd.append(f"meta_file=$(python .actions/assistant.py update-env-details {folder} --base_path .)")
-        # show created meta config
-        cmd += ["echo $meta_file", "cat $meta_file"]
 
         cmd.append(f"# available: {AssistantCLI.DEVICE_ACCELERATOR}")
         if AssistantCLI._valid_accelerator(folder):
+            # and install specific packages
+            pip_req, pip_args = AssistantCLI._parse_requirements(folder)
+            cmd += [f"pip install {pip_req} --quiet {pip_args}", "pip list"]
+            # Export the actual packages used in runtime
+            cmd.append(f"meta_file=$(python .actions/assistant.py update-env-details {folder} --base_path .)")
+            # show created meta config
+            cmd += ["echo $meta_file", "cat $meta_file"]
             cmd.append(f"python -m pytest {ipynb_file} -v --nbval --nbval-cell-timeout=300")
         else:
+            pub_ipynb = os.path.join(DIR_NOTEBOOKS, f"{folder}.ipynb")
+            pub_meta = pub_ipynb.replace(".ipynb", ".yaml")
+            # copy and add meta config
+            cmd += [
+                f"mkdir -p {os.path.dirname(pub_meta)}",
+                f"cp {meta_file} {pub_meta}",
+                f"cat {pub_meta}",
+                f"git add {pub_meta}",
+            ]
             warn("Invalid notebook's accelerator for this device. So no tests will be run!!!", RuntimeWarning)
         # deactivate and clean local environment
         cmd += ["deactivate", f"rm -rf {os.path.join(folder, 'venv')}"]

.azure/ipynb-publish.yml

@@ -61,6 +61,7 @@ jobs:
 
     - bash: |
         set -e
+        sudo apt install -y tree
         pip --version
         pip install --requirement requirements.txt
         pip install --requirement requirements/data.txt

.azure/ipynb-tests.yml

@@ -41,6 +41,7 @@ jobs:
 
     - bash: |
         set -e
+        sudo apt install -y tree
         pip --version
         pip install --requirement requirements.txt
         pip install --requirement requirements/data.txt

.github/workflows/ci_testing.yml

@@ -19,7 +19,6 @@ jobs:
     # Timeout: https://stackoverflow.com/a/59076067/4521646
     timeout-minutes: 55
     env:
-      ACCELERATOR: CPU
       PATH_DATASETS: ${{ github.workspace }}/.datasets
 
     steps:
@@ -75,6 +74,7 @@ jobs:
           KAGGLE_KEY: ${{ secrets.KAGGLE_KEY }}
         run: |
           set -e
+          sudo apt install -y tree
           while IFS= read -r line; do
               python .actions/assistant.py bash-test $line
               cat .actions/_ipynb-test.sh

.pre-commit-config.yaml

@@ -14,9 +14,9 @@ repos:
       - id: end-of-file-fixer
       - id: trailing-whitespace
       - id: check-case-conflict
+      - id: check-json
       - id: check-yaml
       - id: check-toml
-      - id: pretty-format-json
       - id: check-added-large-files
         args: ['--maxkb=250', '--enforce-all']
       - id: check-docstring-first
@@ -26,7 +26,7 @@ repos:
     rev: v2.31.1
     hooks:
       - id: pyupgrade
-        args: [--py36-plus]
+        args: [--py37-plus]
         name: Upgrade code
 
   - repo: https://github.com/myint/docformatter

requirements/default.txt

@@ -1,4 +1,4 @@
 ipython[notebook]
-pytorch-lightning>=1.3
-torchmetrics>=0.3
+pytorch-lightning>=1.4
+torchmetrics>=0.6
 torch>=1.6, <1.9

requirements/devel.txt

@@ -1,8 +1,5 @@
-virtualenv
-jupytext  # converting
+virtualenv>=20.10
+jupytext>=1.10  # converting
 pytest>=6.0
-nbval  # testing
+nbval>=0.9.6  # testing
 papermill>=2.3.4  # render
-black
-flake8
-isort

templates/img-classify/.meta.yml

@@ -0,0 +1,20 @@
+title: Simple image classification with Lightning Flash
+author: PL team
+created: 2022-04-14
+updated: 2021-06-16
+license: CC BY-SA
+build: 0
+tags:
+  - Image
+description: |
+  This is a template to show simple image classification case if for some reason accelerator is required.
+requirements:
+  - lightning-flash[image]>=0.7
+  - pandas>=1.0
+  - matplotlib>=3.0
+  - seaborn
+accelerator:
+  - GPU
+datasets:
+  web:
+    - https://pl-flash-data.s3.amazonaws.com/hymenoptera_data.zip

templates/img-classify/classify.py

@@ -0,0 +1,69 @@
+# %%
+import os
+
+import flash
+import matplotlib.pyplot as plt
+import pandas as pd
+import seaborn as sn
+from flash.image import ImageClassificationData, ImageClassifier
+from pytorch_lightning.loggers import CSVLogger
+
+PATH_DATASETS = os.environ.get("PATH_DATASETS", ".")
+# this dataset is automatically downloaded and extracted based on meta link
+# this archive includes the one more level - folder with the same name
+DATA_HYMENOPLERA = os.path.join(PATH_DATASETS, "hymenoptera_data", "hymenoptera_data")
+
+# %% [markdown]
+# ## 1. Create the DataModule
+
+# %%
+datamodule = ImageClassificationData.from_folders(
+    train_folder=f"{DATA_HYMENOPLERA}/train/",
+    val_folder=f"{DATA_HYMENOPLERA}/val/",
+    batch_size=1024,
+)
+
+# %% [markdown]
+# ## 2. Build the task
+
+# %%
+model = ImageClassifier(backbone="resnet18", labels=datamodule.labels)
+
+# %% [markdown]
+# ## 3. Create the trainer and finetune the model
+
+# %%
+logger = CSVLogger(save_dir="logs/")
+trainer = flash.Trainer(logger=logger, max_epochs=3, gpus=1)
+trainer.finetune(model, datamodule=datamodule, strategy="freeze")
+
+# %%
+metrics = pd.read_csv(f"{trainer.logger.log_dir}/metrics.csv")
+del metrics["step"]
+metrics.set_index("epoch", inplace=True)
+print(metrics.dropna(axis=1, how="all").head())
+
+g = sn.relplot(data=metrics, kind="line")
+plt.gcf().set_size_inches(12, 4)
+plt.grid()
+
+# %% [markdown]
+# ## 4. Predict what's on a few images! ants or bees?
+
+# %%
+datamodule = ImageClassificationData.from_files(
+    predict_files=[
+        f"{DATA_HYMENOPLERA}/val/bees/65038344_52a45d090d.jpg",
+        f"{DATA_HYMENOPLERA}/val/bees/590318879_68cf112861.jpg",
+        f"{DATA_HYMENOPLERA}/val/ants/540543309_ddbb193ee5.jpg",
+    ],
+    batch_size=3,
+)
+predictions = trainer.predict(model, datamodule=datamodule, output="labels")
+print(predictions)
+
+# %% [markdown]
+# ## 5. Save the model!
+
+# %%
+trainer.save_checkpoint("image_classification_model.pt")

templates/simple/.meta.yml

@@ -3,7 +3,7 @@ author: PL team
 created: 2021-06-15
 updated: 2021-06-17
 license: CC
-build: 8
+build: 9
 description: |
   This is a template to show how to contribute a tutorial.
 requirements: