Drop Python 3.6 support (#685)

* Drop python 3.6 support

* Update doc

* Update gitlab CI

* Update doc env

* Fix gitlab CI

.github/workflows/ci.yml

@@ -16,7 +16,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.6, 3.7, 3.8, 3.9]
+        python-version: [3.7, 3.8, 3.9]
 
     steps:
     - uses: actions/checkout@v2

.gitlab-ci.yml

@@ -1,4 +1,4 @@
-image: stablebaselines/stable-baselines3-cpu:1.1.0a5
+image: stablebaselines/stable-baselines3-cpu:1.3.1a3
 
 type-check:
   script:
@@ -6,6 +6,9 @@ type-check:
 
 pytest:
   script:
+  - python --version
+  # Fix to get atari ROMs
+  - pip install atari-py==0.2.5
   # MKL_THREADING_LAYER=GNU to avoid MKL_THREADING_LAYER=INTEL incompatibility error
   - MKL_THREADING_LAYER=GNU make pytest
 

Dockerfile

@@ -1,7 +1,7 @@
 ARG PARENT_IMAGE
 FROM $PARENT_IMAGE
 ARG PYTORCH_DEPS=cpuonly
-ARG PYTHON_VERSION=3.6
+ARG PYTHON_VERSION=3.7
 
 RUN apt-get update && apt-get install -y --no-install-recommends \
          build-essential \

README.md

@@ -82,7 +82,7 @@ Documentation is available online: [https://sb3-contrib.readthedocs.io/](https:/
 **Note:** Stable-Baselines3 supports PyTorch >= 1.8.1.
 
 ### Prerequisites
-Stable Baselines3 requires python 3.6+.
+Stable Baselines3 requires python 3.7+.
 
 #### Windows 10
 

docs/conda_env.yml

@@ -4,9 +4,9 @@ channels:
   - defaults
 dependencies:
   - cpuonly=1.0=0
-  - pip=20.2
-  - python=3.6
-  - pytorch=1.8.1=py3.6_cpu_0
+  - pip=21.1
+  - python=3.7
+  - pytorch=1.8.1=py3.7_cpu_0
   - pip:
     - gym>=0.17.2
     - cloudpickle

docs/guide/install.rst

@@ -6,7 +6,7 @@ Installation
 Prerequisites
 -------------
 
-Stable-Baselines3 requires python 3.6+ and PyTorch >= 1.8.1.
+Stable-Baselines3 requires python 3.7+ and PyTorch >= 1.8.1.
 
 Windows 10
 ~~~~~~~~~~

docs/misc/changelog.rst

@@ -4,11 +4,12 @@ Changelog
 ==========
 
 
-Release 1.3.1a4 (WIP)
+Release 1.3.1a5 (WIP)
 ---------------------------
 
 Breaking Changes:
 ^^^^^^^^^^^^^^^^^
+- Dropped python 3.6 support (as announced in previous release)
 - Renamed ``mask`` argument of the ``predict()`` method to ``episode_start`` (used with RNN policies only)
 - local variables ``action``, ``done`` and ``reward`` were renamed to their plural form for offpolicy algorithms (``actions``, ``dones``, ``rewards``),
   this may affect custom callbacks.

setup.py

@@ -134,11 +134,10 @@ setup(
     long_description=long_description,
     long_description_content_type="text/markdown",
     version=__version__,
-    python_requires=">=3.6",
+    python_requires=">=3.7",
     # PyPI package information.
     classifiers=[
         "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.6",
         "Programming Language :: Python :: 3.7",
         "Programming Language :: Python :: 3.8",
         "Programming Language :: Python :: 3.9",

stable_baselines3/version.txt

@@ -1 +1 @@
-1.3.1a4
+1.3.1a5