remove upper bound numpy requirement (#562)

* remove upper bound numpy requirement

* use numpy larger than 1.22

* use python >=3.8

* remove 3.7 and add 3.10 to tests

* disable pylint false alarm

* remove 3.7 from workflows

* revert pylint disable and change inv for reciprocal

* fix test

.github/workflows/publish-docs.yml

@@ -22,7 +22,7 @@ jobs:
       - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v2
         with:
-          python-version: 3.7
+          python-version: 3.8
 
       - name: Install Dependencies
         run: |

.github/workflows/publish-pypi.yml

@@ -14,7 +14,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v1
         with:
-          python-version: "3.7"
+          python-version: "3.8"
 
       - name: Install dependencies
         run: |

.github/workflows/test.yml

@@ -11,7 +11,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: ["3.7", "3.8", "3.9"]
+        python-version: ["3.8", "3.9", "3.10"]
 
     name: Set up Python ${{ matrix.python-version }}
     steps:

README.md

@@ -13,7 +13,7 @@ Bambi is a high-level Bayesian model-building interface written in Python. It's
 
 ## Installation
 
-Bambi requires a working Python interpreter (3.7+). We recommend installing Python and key numerical libraries using the [Anaconda Distribution](https://www.anaconda.com/products/individual#Downloads), which has one-click installers available on all major platforms.
+Bambi requires a working Python interpreter (3.8+). We recommend installing Python and key numerical libraries using the [Anaconda Distribution](https://www.anaconda.com/products/individual#Downloads), which has one-click installers available on all major platforms.
 
 Assuming a standard Python environment is installed on your machine (including pip), Bambi itself can be installed in one line using pip:
 

bambi/tests/test_predict.py

@@ -346,7 +346,7 @@ def test_predict_include_group_specific():
     assert not np.isclose(
         idata_1.posterior["y_mean"].values,
         idata_2.posterior["y_mean"].values,
-    ).any()
+    ).all()
 
     # Since it's an intercept-only model, predictions are the same for all observations if
     # we drop group-specific terms.

docs/index.rst

@@ -25,7 +25,7 @@ Bambi is a high-level Bayesian model-building interface written in Python. It wo
 
 Dependencies
 ============
-Bambi is tested on Python 3.7+ and depends on ArviZ, formulae, NumPy, pandas and PyMC (see `requirements.txt <https://github.com/bambinos/bambi/blob/main/requirements.txt>`_ for version information).
+Bambi is tested on Python 3.8+ and depends on ArviZ, formulae, NumPy, pandas and PyMC (see `requirements.txt <https://github.com/bambinos/bambi/blob/main/requirements.txt>`_ for version information).
 
 Installation
 ============

pyproject.toml

@@ -1,3 +1,3 @@
 [tool.black]
 line-length = 100
-target-version = ['py37', 'py38']
+target-version = ['py38', 'py39']

requirements.txt

@@ -1,6 +1,6 @@
 arviz>=0.11.2
 formulae==0.3.4
-numpy>=1.16.1, <1.22.0
+numpy>1.22
 pandas>=1.0.0
 pymc>=4.0.0
 scipy>=1.7.0