feature: dataset from paths

Author: samedii

.github/workflows/publish.yml

@@ -19,35 +19,30 @@ jobs:
 
       - name: Install dependencies
         run: |
-          curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python
-          source $HOME/.poetry/env
-          poetry install
+          curl -sSL https://install.python-poetry.org | python - --version 1.2.1
+          $HOME/.local/bin/poetry install --no-root
 
       - name: Run tests
         run: |
-          source $HOME/.poetry/env
-          poetry run pytest
+          $HOME/.local/bin/poetry run pytest
 
       - name: Build wheels
         run: |
-          source $HOME/.poetry/env
-          poetry version $(git tag --points-at HEAD)
-          poetry build
+          $HOME/.local/bin/poetry version $(git tag --points-at HEAD)
+          $HOME/.local/bin/poetry build
 
       - name: Test install package
         run: |
-          source $HOME/.poetry/env
           mkdir test_install
           cd test_install
-          poetry init
-          poetry add ../dist/$(ls dist/*.whl)
+          $HOME/.local/bin/poetry init
+          $HOME/.local/bin/poetry add ../dist/$(ls dist/*.whl)
 
-          poetry run python -c "import datastream"
+          $HOME/.local/bin/poetry run python -c "import datastream"
 
       - name: Upload
         env:
           USERNAME: __token__
           PASSWORD: ${{ secrets.PYPI_TOKEN }}
         run: |
-          source $HOME/.poetry/env
-          poetry publish --username=$USERNAME --password=$PASSWORD
+          $HOME/.local/bin/poetry publish --username=$USERNAME --password=$PASSWORD

.github/workflows/test.yml

@@ -26,19 +26,16 @@ jobs:
 
       - name: Install dependencies
         run: |
-          curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python
-          source $HOME/.poetry/env
-          poetry install
+          curl -sSL https://install.python-poetry.org | python - --version 1.2.1
+          $HOME/.local/bin/poetry install install
 
       - name: Run tests
         run: |
-          source $HOME/.poetry/env
-          poetry run pytest
+          $HOME/.local/bin/poetry install run pytest
 
       - name: Build wheels
         run: |
-          source $HOME/.poetry/env
-          poetry build
+          $HOME/.local/bin/poetry install build
 
   build-docs:
     runs-on: ubuntu-latest

datastream/dataset.py

@@ -89,6 +89,30 @@ def from_dataframe(dataframe: pd.DataFrame) -> Dataset[pd.Series]:
             get_item=lambda df, index: df.iloc[index],
         )
 
+    @staticmethod
+    def from_paths(paths: Iterable[str, Path], pattern: str) -> Dataset[pd.Series]:
+        '''
+        Create ``Dataset`` from paths using regex pattern that extracts information
+        from the path itself.
+        :func:`Dataset.__getitem__` will return a row from the dataframe and
+        :func:`Dataset.map` should be given a function that takes a row from
+        the dataframe as input.
+
+        >>> image_paths = ["dataset/damage/1.png"]
+        >>> (
+        ...     Dataset.from_paths(image_paths, pattern=r".*/(?P<class_name>\w+)/(?P<index>\d+).png")
+        ...     .map(lambda row: row["class_name"])
+        ... )[-1]
+        'damage'
+        '''
+        paths = list(paths)
+        return Dataset.from_dataframe(
+            pd.Series(paths)
+            .astype(str)
+            .str.extract(pattern)
+            .assign(path=paths)
+        )
+
     def __getitem__(
             self: Dataset[T],
             select: Union[int, slice, Iterable, Callable[[pd.DataFrame], Iterable[int]]]