0.6.0

.github/workflows/python-package.yml

@@ -24,21 +24,33 @@ jobs:
       uses: actions/setup-python@v3
       with:
         python-version: ${{ matrix.python-version }}
+
     - name: Install dependencies
       run: |
+        pwd
+        ls
         python -m pip install --upgrade pip
-        python -m pip install flake8 pytest
-        if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
-        if [ -f utils_requirements.txt ]; then pip install -r utils_requirements.txt; fi
+        python -m pip install -r requirements.txt
+        python -m pip install -r cascade/tests/requirements.txt
+        python -m pip install -r utils_requirements.txt
+
     - name: Lint with flake8
       run: |
         # stop the build if there are Python syntax errors or undefined names
         flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
         # exit-zero treats all errors as warnings. The GitHub editor is 127 chars wide
         flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
+
     - name: Test
       run: |
         pwd
         ls
         python --version
-        python -m unittest discover ./cascade/tests
+        cd ./cascade/tests
+        pytest --cov=cascade .
+    - name: Test utils
+      run: |
+        pwd
+        ls
+        cd ./cascade/utils/tests
+        pytest

cascade/__init__.py

@@ -14,7 +14,7 @@
 limitations under the License.
 """
 
-__version__ = '0.5.2'
+__version__ = '0.6.0'
 __author__ = 'Ilia Moiseev'
 __author_email__ = 'ilia.moiseev.5@yandex.ru'
 
@@ -25,7 +25,7 @@
 from . import tests
 
 # cascade does not have
-# `from . import utils`
+# from . import utils
 # because it will bring additional dependencies that may not be needed by the user
 # if you need to use cascade.utils, you can install utils_requirements.txt and then 
 # import as any other cascade module

cascade/base/__init__.py

@@ -1,2 +1,3 @@
 from .meta_handler import MetaHandler
 from .traceable import Traceable
+from .meta_handler import CustomEncoder as JSONEncoder

cascade/base/meta_handler.py

@@ -16,23 +16,28 @@
 
 import os
 import json
+from typing import Union, Dict, List
 import datetime
+from typing import List, Dict
 from json import JSONEncoder
 
+import yaml
 import numpy as np
 
 
 class CustomEncoder(JSONEncoder):
     def default(self, obj):
+        if isinstance(obj, type):
+            return str(obj)
         if isinstance(obj, (datetime.datetime, datetime.date, datetime.time)):
             return obj.isoformat()
+
         elif isinstance(obj, datetime.timedelta):
             return (datetime.datetime.min + obj).time().isoformat()
 
         elif isinstance(obj, (np.int_, np.intc, np.intp, np.int8,
                               np.int16, np.int32, np.int64, np.uint8,
                               np.uint16, np.uint32, np.uint64)):
-
             return int(obj)
 
         elif isinstance(obj, (np.float_, np.float16, np.float32, np.float64)):
@@ -44,29 +49,39 @@ def default(self, obj):
         elif isinstance(obj, (np.ndarray,)):
             return obj.tolist()
 
-        elif isinstance(obj, (np.bool_)):
+        elif isinstance(obj, np.bool_):
             return bool(obj)
 
-        elif isinstance(obj, (np.void)):
+        elif isinstance(obj, np.void):
             return None
 
         return super(CustomEncoder, self).default(obj)
 
+    def obj_to_dict(self, obj):
+        return json.loads(self.encode(obj))
 
-class MetaHandler:
+
+class BaseHandler:
+    def read(self, path) -> List[Dict]:
+        raise NotImplementedError()
+
+    def write(self, path, obj, overwrite=True) -> None:
+        raise NotImplementedError()
+
+
+class JSONHandler(BaseHandler):
     """
     Handles the logic of dumping and loading json files
     """
-    def read(self, path) -> dict:
+    def read(self, path) -> Union[Dict, List[Dict]]:
         """
         Reads json from path
 
         Parameters
         ----------
         path:
-            Path to the file. If no extension provided, then .json assumed
+            Path to the file. If no extension provided, then .json will be added
         """
-        assert os.path.exists(path)
         _, ext = os.path.splitext(path)
         if ext == '':
             path += '.json'
@@ -77,16 +92,77 @@ def read(self, path) -> dict:
                 meta = json.loads(meta)
             return meta
 
-    def write(self, name, obj, overwrite=True) -> None:
+    def write(self, name, obj:List[Dict], overwrite=True) -> None:
         """
         Writes json to path using custom encoder
         """
-
         if not overwrite and os.path.exists(name):
             return
 
-        with open(name, 'w') as json_meta:
-            json.dump(obj, json_meta, cls=CustomEncoder, indent=4)
+        with open(name, 'w') as f:
+            json.dump(obj, f, cls=CustomEncoder, indent=4)
+
+
+class YAMLHandler(BaseHandler):
+    def read(self, path) -> Union[Dict, List[Dict]]:
+        """
+        Reads yaml from path
+
+        Parameters
+        ----------
+        path:
+            Path to the file. If no extension provided, then .yml will be added
+        """
+        _, ext = os.path.splitext(path)
+        if ext == '':
+            path += '.yml'
+
+        with open(path, 'r') as meta_file:
+            meta = yaml.safe_load(meta_file)
+            return meta
+
+    def write(self, path, obj, overwrite=True) -> None:
+        if not overwrite and os.path.exists(path):
+            return
+
+        obj = CustomEncoder().obj_to_dict(obj)
+        with open(path, 'w') as f:
+            yaml.safe_dump(obj, f)
 
-    def encode(self, obj):
-        return CustomEncoder().encode(obj)
+
+class TextHandler(BaseHandler):
+    def read(self, path) -> Dict:
+        """
+        Reads text file from path and returns dict in the form {path: 'text from file'}
+
+        Parameters
+        ----------
+        path:
+            Path to the file
+        """
+
+        with open(path, 'r') as meta_file:
+            meta = {path: ''.join(meta_file.readlines())}
+            return meta
+
+    def write(self, path, obj, overwrite=True) -> None:
+        raise NotImplementedError('MetaHandler does not write text files, only reads')       
+
+
+class MetaHandler:
+    def read(self, path) -> List[Dict]:
+        handler = self._get_handler(path)
+        return handler.read(path)
+
+    def write(self, path, obj, overwrite=True) -> None:
+        handler = self._get_handler(path)
+        return handler.write(path, obj, overwrite=overwrite)
+
+    def _get_handler(self, path) -> BaseHandler:
+        ext = os.path.splitext(path)[-1]
+        if ext == '.json':
+            return JSONHandler()
+        elif ext == '.yml':
+            return YAMLHandler()
+        else:
+            return TextHandler()

cascade/base/traceable.py

@@ -1,17 +1,19 @@
 import warnings
-from typing import List, Dict
+from typing import List, Dict, Union
 
 
 class Traceable:
     def __init__(self, *args, meta_prefix=None, **kwargs) -> None:
         if meta_prefix is None:
             meta_prefix = {}
-        if isinstance(meta_prefix, str):
-            from . import MetaHandler
-
-            meta_prefix = MetaHandler().read(meta_prefix)
+        elif isinstance(meta_prefix, str):
+            meta_prefix = self._read_meta_from_file(meta_prefix)
         self.meta_prefix = meta_prefix
 
+    def _read_meta_from_file(self, path: str) -> Union[List[Dict], Dict]:
+            from . import MetaHandler
+            return MetaHandler().read(path)
+
     def get_meta(self) -> List[Dict]:
         """
         Returns
@@ -30,10 +32,13 @@ def get_meta(self) -> List[Dict]:
             self._warn_no_prefix()
         return [meta]
 
-    def update_meta(self, obj: Dict) -> None:
+    def update_meta(self, obj: Union[Dict, str]) -> None:
         """
         Updates meta_prefix, which is then updates dataset's meta when get_meta() is called
         """
+        if isinstance(obj, str):
+            obj = self._read_meta_from_file(obj)
+
         if hasattr(self, 'meta_prefix'):
             self.meta_prefix.update(obj)
         else:

cascade/data/concatenator.py

@@ -64,10 +64,10 @@ def __repr__(self) -> str:
 
     def get_meta(self) -> List[Dict]:
         """
-        Concatenator calls `get_meta()` of all its datasets and appends to its own meta
+        Concatenator calls `get_meta()` of all its datasets
         """
         meta = super().get_meta()
-        meta[0]['data'] = []
+        meta[0]['data'] = {}
         for ds in self._datasets:
-            meta[0]['data'] += ds.get_meta()
+            meta[0]['data'][repr(ds)] = ds.get_meta()
         return meta

cascade/data/dataset.py

@@ -76,6 +76,11 @@ def __getitem__(self, item):
     def __iter__(self):
         for item in self._data:
             yield item
+
+    def get_meta(self):
+        meta = super().get_meta()
+        meta[0]['obj_type'] = str(type(self._data))
+        return meta
 
 
 class Wrapper(Dataset):
@@ -95,7 +100,7 @@ def __len__(self) -> int:
     def get_meta(self):
         meta = super().get_meta()
         meta[0]['len'] = len(self)
-        meta[0]['obj_type'] = type(self._data)
+        meta[0]['obj_type'] = str(type(self._data))
         return meta
 
 

cascade/data/pickler.py

@@ -25,10 +25,10 @@ class Pickler(Modifier):
     """
     def __init__(self, path, dataset=None, *args, **kwargs) -> None:
         """
-        Loads pickled dataset or dumps one depending on parameters passed
+        Loads pickled dataset or dumps one depending on parameters passed:
 
-        If only path is passed - loads dataset from path provided if path exists
-        if path provided with a dataset dumps dataset to the path
+        1. If only path is passed - loads dataset from path provided if path exists
+        2. if path provided with a dataset dumps dataset to the path
 
         Parameters
         ----------

cascade/data/random_sampler.py

@@ -19,7 +19,20 @@
 
 
 class RandomSampler(Sampler):
+    """
+    Shuffles dataset. Can randomly sample from dataset 
+    if num_samples is not None and less than length of dataset.
+    """
     def __init__(self, dataset: Dataset, num_samples=None, **kwargs) -> None:
+        """
+        Parameters
+        ----------
+        dataset: Dataset
+            Input dataset to sample from
+        num_samples: int, optional
+            Should be less than len(dataset), but oversampling can be added in the future.
+            If None, then just shuffles the dataset.
+        """
         if num_samples is None:
             num_samples = len(dataset)
         super().__init__(dataset, num_samples, **kwargs)

cascade/docs/requirements.txt

@@ -1,3 +1,4 @@
 sphinx
 furo
 sphinx-copybutton
+nbsphinx

cascade/docs/source/cascade.base.rst

@@ -0,0 +1,19 @@
+cascade.base
+============ 
+.. autoclass:: cascade.base.Traceable
+   :members:
+
+|
+
+|
+
+|
+
+.. autoclass:: cascade.base.MetaHandler
+   :members:
+
+|
+
+|
+
+|

cascade/docs/source/cascade.data.rst

@@ -82,15 +82,14 @@ cascade.data
 
 | 
 
-
 .. autoclass:: cascade.data.FolderDataset
    :members:
 
 |
 
 |
 
-| 
+|
 
 .. autoclass:: cascade.data.Pickler
    :members:
@@ -99,6 +98,15 @@ cascade.data
 
 |
 
+|
+
+.. autoclass:: cascade.data.RandomSampler
+   :members:
+
+|
+
+|
+
 | 
 
 .. autoclass:: cascade.data.SequentialCacher