Merge branch 'releaseprep' into cookBux

environment.yml

-name: saqc
-channels:
-  - defaults
-dependencies:
-  - numpy
-  - pandas
-  - scipy
-  - matplotlib
-  - numba
-  - click
-  - scikit-learn
-  - pyarrow
-  - pip
-  - pip:
-    - python-intervals
-    - dtw
-    - mlxtend
-    - outlier-utils

requirements.txt

@@ -2,7 +2,6 @@ Click==8.0.1
 dtw==1.4.0
 hypothesis==6.23.1
 matplotlib==3.4.3
-mlxtend==0.19.0
 numba==0.54.0
 numpy==1.20.3
 outlier-utils==0.0.3

saqc/core/history.py

@@ -402,7 +402,7 @@ class History:
 
         if not hist.empty and (
             not hist.columns.equals(pd.Index(range(len(hist.columns))))
-            or hist.columns.dtype != int
+            or not np.issubdtype(hist.columns.dtype, np.integer)
         ):
             raise ValueError(
                 "column names must be continuous increasing int's, starting with 0."

saqc/funcs/tools.py

@@ -3,7 +3,6 @@
 
 from typing import Optional, Tuple
 
-from py import process
 from typing_extensions import Literal
 import numpy as np
 from dios import DictOfSeries

setup.py

@@ -5,7 +5,7 @@ with open("README.md", "r") as fh:
 
 setup(
     name="saqc",
-    version="1.4",
+    version="2.0",
     author="Bert Palm, David Schaefer, Peter Luenenschloss, Lennard Schmidt",
     author_email="david.schaefer@ufz.de",
     description="Data quality checking and processing tool/framework",
@@ -15,19 +15,18 @@ setup(
     packages=find_packages(exclude=("tests",)),
     python_requires=">=3.7, <3.10",
     install_requires=[
-        "numpy",
-        "pandas",
-        "scipy",
-        "scikit-learn",
-        "numba",
-        "matplotlib",
-        "click",
-        "pyarrow",
-        "typing_extensions",
-        "outlier-utils",
-        "dtw",
-        "PyWavelets",
-        "mlxtend",
+        "numpy==1.20.*",
+        "pandas==1.3.*",
+        "scipy==1.7.*",
+        "scikit-learn==1.0.*",
+        "numba==0.54.*",
+        "matplotlib==3.4.*",
+        "Click==8.0.*",
+        "pyarrow==4.0.*",
+        "typing_extensions==3.10.*",
+        "outlier-utils==0.0.3",
+        "dtw==1.4.*",
+        "seaborn==0.11.*",
     ],
     license_files=("LICENSE.md",),
     entry_points={

tests/integration/test_integration.py

 #!/usr/bin/env python
 from click.testing import CliRunner
-import os
+from pathlib import Path
 
 
-def test__main__py():
+def test__main__py(tmp_path):
     import saqc.__main__
 
     # if not run from project root
-    projpath = os.path.dirname(saqc.__file__) + "/../"
+    projpath = Path(saqc.__file__).parents[1]
     args = [
         "--config",
-        projpath + "ressources/data/config.csv",
+        Path(projpath, "ressources/data/config.csv"),
         "--data",
-        projpath + "ressources/data/data.csv",
+        Path(projpath, "ressources/data/data.csv"),
         "--outfile",
-        "/tmp/test.csv",  # the filesystem temp dir
+        Path(tmp_path, "test.csv"),  # the filesystem temp dir
     ]
     runner = CliRunner()
-
     for scheme in ["float", "positional", "dmp", "simple"]:
         result = runner.invoke(saqc.__main__.main, args + ["--scheme", scheme])
         assert result.exit_code == 0, result.output