Create python-package.yml with pytests. Fix code by different python …

…versions tests tests

.github/workflows/python-package.yml

@@ -0,0 +1,44 @@
+# This workflow will install Python dependencies, run tests and lint with a variety of Python versions
+# For more information see: https://docs.github.com/en/actions/automating-builds-and-tests/building-and-testing-python
+
+name: PyTests
+
+on:
+  push:
+    branches: [ "main" ]
+  pull_request:
+    branches: [ "main" ]
+
+jobs:
+  build:
+
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.7", "3.8", "3.9", "3.10", "3.11"]
+
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v3
+      with:
+        python-version: ${{ matrix.python-version }}
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        python -m pip install flake8 mypy pytest
+        if [ -f requirements-test.txt ]; then pip install -r requirements-test.txt; fi
+        pip install .[tests]
+    - name: flake8
+      run: |
+         flake8 --count --show-source --statistics mlup tests examples
+#    - name: mypy
+#      run: |
+#        - mypy .
+    - name: Unit tests
+      run: |
+        pytest tests/unit_tests
+    - name: Integration tests
+      run: |
+        pytest tests/integration_tests

mlup/config.py

@@ -120,7 +120,7 @@ def get_config_dict(self) -> Dict:
         return config
 
     def load_from_dict(self, conf: Dict):
-        logger.info(f'Load config from dict')
+        logger.info('Load config from dict')
         self.set_config_from_dict(conf)
 
     def load_from_json(self, file_path: Union[str, Path]):

mlup/console_scripts/command.py

@@ -17,14 +17,14 @@
 
 options:
   -h, --help            show this help message and exit
-  
+
 {AVAILABLE_COMMANDS}
 """
 
 
 def run_command(args: List[str]):
     try:
-        command, command_args = args[0], args[1:]
+        command, _ = args[0], args[1:]
     except IndexError:
         print(HELP)
         sys.exit(1)
@@ -37,7 +37,7 @@ def run_command(args: List[str]):
 
     try:
         module = importlib.import_module(f'mlup.console_scripts.{command}')
-    except ModuleNotFoundError as e:
+    except ModuleNotFoundError:
         print(f'Invalid command {command} - mlup.console_scripts.{command}.')
         print(AVAILABLE_COMMANDS)
         sys.exit(1)

mlup/ml/binarization/base.py

@@ -4,7 +4,7 @@
 from mlup.constants import LoadedFile
 
 
-@dataclass(kw_only=True)
+@dataclass
 class BaseBinarizer(metaclass=abc.ABCMeta):
     @classmethod
     @abc.abstractmethod

mlup/ml/binarization/joblib.py

@@ -19,7 +19,7 @@
 class JoblibBinarizer(PickleBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization joblib data.')
+        logger.info('Run deserialization joblib data.')
         if joblib is None:
             logger.error('For use joblib, please install it. pip install joblib.')
             raise ModelBinarizationError('For use joblib, please install it. pip install joblib.')

mlup/ml/binarization/lightgbm.py

@@ -14,7 +14,7 @@
 class LightGBMBinarizer(BaseBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization lightgbm data.')
+        logger.info('Run deserialization lightgbm data.')
         with TimeProfiler('Time to deserialization lightgbm data:'):
             try:
                 if data.path:

mlup/ml/binarization/memory.py

@@ -9,5 +9,5 @@
 class MemoryBinarizer(BaseBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization memory data.')
+        logger.info('Run deserialization memory data.')
         return data.raw_data

mlup/ml/binarization/onnx.py

@@ -27,7 +27,7 @@ def predict(self, input_data):
 class InferenceSessionBinarizer(BaseBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization onnxruntime data.')
+        logger.info('Run deserialization onnxruntime data.')
         with TimeProfiler('Time to deserialization onnxruntime data:'):
             try:
                 _data = data.raw_data

mlup/ml/binarization/pickle.py

@@ -15,7 +15,7 @@
 class PickleBinarizer(BaseBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization pickle data.')
+        logger.info('Run deserialization pickle data.')
         with TimeProfiler('Time to deserialization pickle data:'):
             try:
                 if isinstance(data.raw_data, BufferedIOBase):

mlup/ml/binarization/tensorflow.py

@@ -16,7 +16,7 @@
 class TensorFlowBinarizer(BaseBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization tensorflow data.')
+        logger.info('Run deserialization tensorflow data.')
         with TimeProfiler('Time to deserialization tensorflow data:'):
             try:
                 _data = data.raw_data
@@ -63,7 +63,7 @@ def is_this_type(cls, loaded_file: LoadedFile) -> float:
 class TensorFlowSavedBinarizer(BaseBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization tensorflow SavedModel data.')
+        logger.info('Run deserialization tensorflow SavedModel data.')
         with TimeProfiler('Time to deserialization tensorflow SavedModel data:'):
             try:
                 _data = data.raw_data

mlup/ml/binarization/torch.py

@@ -15,7 +15,7 @@
 class TorchBinarizer(BaseBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization torch data.')
+        logger.info('Run deserialization torch data.')
         with TimeProfiler('Time to deserialization torch data:'):
             try:
                 _data = data.raw_data
@@ -64,7 +64,7 @@ def is_this_type(cls, loaded_file: LoadedFile) -> float:
 class TorchJITBinarizer(BaseBinarizer):
     @classmethod
     def deserialize(cls, data: LoadedFile):
-        logger.info(f'Run deserialization torch JIT data.')
+        logger.info('Run deserialization torch JIT data.')
         with TimeProfiler('Time to deserialization torch JIT data:'):
             try:
                 _data = data.raw_data

mlup/ml/model.py

@@ -29,7 +29,7 @@
 logger = logging.getLogger('mlup')
 
 
-@dataclass(kw_only=True)
+@dataclass
 class ModelConfig:
     """
     MlupModel config class. This class have settings for model.
@@ -149,7 +149,7 @@ def ml_str(self, need_spaces: bool = False):
         return '\n'.join(res)
 
 
-@dataclass(kw_only=True, repr=True)
+@dataclass(repr=True)
 class MLupModel(ModelConfig):
     """This is main UP model class.
     Create object UP with your ML model, set your settings and run your app.
@@ -274,7 +274,7 @@ def _transform_predicted_data(self, predicted_data: Any):
             try:
                 processing_data = self._data_transformer_for_predicted.transform_to_json_format(predicted_data)
             except Exception as e:
-                logger.exception(f'Fail transform predicted data to response format.')
+                logger.exception('Fail transform predicted data to response format.')
                 raise PredictTransformDataError(str(e))
             finally:
                 logger.debug('Finish transform predicted data to response format.')
@@ -302,7 +302,7 @@ async def _predict(self, data_for_predict: Optional[Any] = None, **other_predict
                             *_predict_args
                         )
                 else:
-                    logger.debug(f'Running sync predict.')
+                    logger.debug('Running sync predict.')
                     with self._lock:
                         result = self._prepare_predict_method(**other_predict_args)(*_predict_args)
             except Exception as e:
@@ -423,4 +423,3 @@ async def predict_from(self, **predict_data):
             raise
         except Exception as e:
             raise PredictError(e)
-

mlup/up.py

@@ -24,12 +24,12 @@ def generate_default_config(path_to_file: Optional[str] = None) -> Optional[Dict
         up.to_yaml(path_to_file)
 
 
-@dataclass(kw_only=True)
+@dataclass
 class Config(ModelConfig, WebAppConfig):
     pass
 
 
-@dataclass(kw_only=True, repr=False)
+@dataclass(repr=False)
 class UP:
     """This is main UP class.
     Create object UP with your ML model, set your settings and run your web app.

mlup/utils/interspection.py

@@ -7,6 +7,7 @@
 from mlup.constants import IS_X, THERE_IS_ARGS, DEFAULT_X_ARG_NAME, BinarizationType, LoadedFile
 from mlup.utils.profiling import TimeProfiler
 
+
 logger = logging.getLogger('mlup')
 
 
@@ -107,7 +108,7 @@ def example(a, b = 100, *, c: float = 123):
                 param_data['type'] = types[type(param_obj.default)]
 
         if param_name.lower().strip() == 'x' and auto_detect_predict_params:
-            logger.info(f'Found X param in model params. Set List type')
+            logger.info('Found X param in model params. Set List type')
             param_data['type'] = 'List'
             param_data[IS_X] = True
             _found_X = True
@@ -145,7 +146,7 @@ def auto_search_binarization_type(loaded_file: LoadedFile) -> Optional[Type[Bina
     :rtype: Optional[Type[BinarizationType]]
 
     """
-    logger.info(f'Run auto search binarizer.')
+    logger.info('Run auto search binarizer.')
     probabilities = []
     with TimeProfiler('Time to auto search binarizer:', log_level='info'):
         for binarizer_path in BinarizationType:

mlup/utils/logging.py

@@ -21,7 +21,9 @@ def __init__(self, fmt: Optional[str] = None, *args, **kwargs):
     def set_fmt(self, fmt_name: str = 'default'):
         fmt = self._fmts[fmt_name]
         self._style = logging.PercentStyle(fmt)
-        self._style.validate()
+        # This "if" for python3.7-
+        if hasattr(self._style, 'validate'):
+            self._style.validate()
         self._fmt = self._style._fmt
 
 

mlup/utils/loop.py

@@ -1,4 +1,5 @@
 import asyncio
+import sys
 import threading
 
 
@@ -29,3 +30,10 @@ def run_async(func, *args, **kwargs):
         return thread.result
     else:
         return asyncio.run(func(*args, **kwargs))
+
+
+def create_async_task(coro, *, name=None):
+    task_kwargs = {}
+    if sys.version_info.minor >= 8:
+        task_kwargs['name'] = name
+    return asyncio.create_task(coro, **task_kwargs)

mlup/web/api_validators.py

@@ -1,7 +1,7 @@
 import logging
 from typing import List, Any, Dict, Tuple, Optional, Type
 
-from pydantic import BaseModel, validator, create_model, Field
+from pydantic import BaseModel, create_model, Field
 
 from mlup.constants import IS_X, DEFAULT_X_ARG_NAME
 from mlup.ml.model import MLupModel

mlup/web/app.py

@@ -71,7 +71,7 @@ async def wrap(*args, response: FastAPIResponse, **kwargs):
     return wrap
 
 
-@dataclass(kw_only=True)
+@dataclass
 class WebAppConfig:
     """
     WebAppConfig config class. This class have settings for web app.
@@ -227,7 +227,7 @@ def wb_str(self, need_spaces: bool = False):
         return '\n'.join(res)
 
 
-@dataclass(kw_only=True, repr=True)
+@dataclass(repr=True)
 class MLupWebApp:
     """This is main UP web app class.
     Create object UP with your ML model, set your settings and run your web app.

mlup/web/architecture/batching.py

@@ -99,7 +99,7 @@ def _save_predicted_data_from_batch(self, predicted_data: List, error: Optional[
                 local_result = []
 
     async def _predict(self):
-        logger.debug(f'Run predict for batch')
+        logger.debug('Run predict for batch')
         predicted_data = self.batch_queue
         error = None
         try:
@@ -112,7 +112,7 @@ async def _predict(self):
         self._save_predicted_data_from_batch(predicted_data, error)
         self.batch_queue.clear()
         self.batch_predict_ids.clear()
-        logger.debug(f'End predict for batch')
+        logger.debug('End predict for batch')
 
     async def _start_worker(self, sleep_time: float = 0.1):
         self._running = True

mlup/web/architecture/worker_and_queue.py

@@ -9,6 +9,7 @@
 from mlup.errors import WebAppLoadError, PredictError, PredictWaitResultError
 from mlup.ml.model import MLupModel
 from mlup.utils.collections import TTLOrderedDict
+from mlup.utils.loop import create_async_task
 from mlup.web.architecture.base import BaseWebAppArchitecture
 from mlup.constants import WebAppArchitecture
 
@@ -102,7 +103,7 @@ async def run(self):
         logger.info('Run model in worker')
         if self.results_storage is None or self.queries_queue is None:
             raise WebAppLoadError('Not called .load() in web.load()')
-        self.worker_process = asyncio.create_task(self._start_worker())
+        self.worker_process = create_async_task(self._start_worker())
 
     @property
     def is_running(self) -> bool:

pyproject.toml

@@ -55,13 +55,19 @@ Documentation = "https://github.com/nxexox/pymlup/docs"
 Repository = "https://github.com/nxexox/pymlup"
 
 [project.optional-dependencies]
-scikit-learn = ["scikit-learn>=1.2.0,<1.3.0"]
+scikit-learn = [
+    "scikit-learn>=1.2.0,<1.3.0;python_version>='3.8'",
+    "scikit-learn;python_version<'3.8'",
+]
 lightgbm = ["lightgbm>=4.0.0,<5.0.0"]
 tensorflow = ["tensorflow>=2.0.0,<3.0.0"]
-torch = ["torch>=2.0.0,<3.0.0"]
+torch = [
+    "torch>=2.0.0,<3.0.0;python_version>='3.8'",
+    "torch;python_version<'3.8'",
+]
 onnx = [
     "onnx>=1.0.0,<2.0.0",
-    "onnxruntime>=1.0.0,<2.0.0",
+    "onnxruntime<1.16",
 ]
 tests = [
     # Tests
@@ -70,13 +76,17 @@ tests = [
 
     # Tests models
     "joblib>=1.2.0,<1.3.0",
-    "pandas>=2.0.0,<3.0.0",
-    "scikit-learn>=1.2.0,<1.3.0",
-    "tensorflow>=2.0.0,<3.0.0",
+    "pandas>=2.0.0,<3.0.0;python_version>='3.8'",
+    "pandas;python_version<'3.8'",
+    "scikit-learn>=1.2.0,<1.3.0;python_version>='3.8'",
+    "scikit-learn;python_version<'3.8'",
+    "tensorflow>=2.0.0,<3.0.0;python_version>='3.8'",
+    "tensorflow;python_version<'3.8'",
     "lightgbm>=4.0.0,<5.0.0",
-    "torch>=2.0.0,<3.0.0",
+    "torch>=2.0.0,<3.0.0;python_version>='3.8'",
+    "torch;python_version<'3.8'",
     "onnx>=1.0.0,<2.0.0",
-    "onnxruntime>=1.0.0,<2.0.0",
+    "onnxruntime<1.16",
     "tf2onnx>=1.0.0,<2.0.0",
     "skl2onnx>=1.0.0,<2.0.0",