Integrate winner solutions, puffer 0.7, and syllabus

.github/workflows/workflow.yml

@@ -0,0 +1,24 @@
+name: tox
+on: [push, pull_request]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        py: ["3.8", "3.9", "3.10"]
+    steps:
+      - name: Setup python for test ${{ matrix.py }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.py }}
+      - uses: actions/checkout@v3
+      - name: Upgrade pip
+        run: python -m pip install -U pip setuptools wheel cython
+      - name: Install
+        run: python -m pip install -e '.[dev]'
+      - name: Check formatting
+        run: ruff format .
+      - name: Check lint
+        run: ruff check .

.pre-commit-config.yaml

@@ -0,0 +1,9 @@
+repos:
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    # Ruff version.
+    rev: v0.3.2
+    hooks:
+      # Run the linter.
+      - id: ruff
+      # Run the formatter.
+      - id: ruff-format

README.md

@@ -2,8 +2,69 @@
 
 # ![icon](https://neuralmmo.github.io/_build/html/_images/icon.png) Welcome to the Platform!
 
+[Documentation](https://neuralmmo.github.io "Neural MMO Documentation") is hosted by github.io.
 [![PyPI version](https://badge.fury.io/py/nmmo.svg)](https://badge.fury.io/py/nmmo)
 [![](https://dcbadge.vercel.app/api/server/BkMmFUC?style=plastic)](https://discord.gg/BkMmFUC)
 [![Twitter](https://img.shields.io/twitter/url/https/twitter.com/cloudposse.svg?style=social&label=Follow%20%40jsuarez5341)](https://twitter.com/jsuarez5341)
 
-Baselines repository for Neural MMO. New users should treat this as a starter project. [Documentation](https://neuralmmo.github.io "Neural MMO Documentation") is hosted by github.io.
+## Installation
+
+After cloning this repo, run:
+
+```
+pip install -e .[dev]
+```
+
+## Training
+
+To test if the installation was successful (with the `--debug` mode), run the following command:
+
+```
+python train.py --debug --no-track
+```
+
+To log the training process, edit the wandb section in `config.yaml` and remove `--no-track` from the command line. The `config.yaml` file contains various configuration settings for the project.
+
+### Agent zoo and your custom policy
+
+This baseline comes with four different models under the `agent_zoo` directory: `neurips23_start_kit`, `yaofeng`, `takeru`, and `hybrid`. You can use any of these models by specifying the `-a` argument.
+
+```
+python train.py -a hybrid
+```
+
+You can also create your own policy by creating a new module under the `agent_zoo` directory, which should contain `Policy`, `Recurrent`, and `RewardWrapper` classes.
+
+### Curriculum Learning using Syllabus
+
+The training script supports automatic curriculum learning using the [Syllabus](https://github.com/RyanNavillus/Syllabus) library. To use it, add `--syllabus` to the command line.
+
+```
+python train.py --syllabus
+```
+
+## Replay generation
+
+The `policies` directory contains a set of trained policies. For your models, create a directory and copy the checkpoint files to it. To generate a replay, run the following command:
+
+```
+python train.py -m replay -p policies
+```
+
+The replay file ends with `.replay.lzma`. You can view the replay using the [web viewer](https://kywch.github.io/nmmo-client/).
+
+## Evaluation
+
+The evaluation script supports the pvp and pve modes. The pve mode spawns all agents using only one policy. The pvp mode spawns groups of agents, each controlled by a different policy.
+
+To evaluate models in the `policies` directory, run the following command:
+
+```
+python evaluate.py policies pvp -r 10
+```
+
+This generates 10 results json files in the same directory (by using `-r 10`), each of which contains the results from 200 episodes. Then the task completion metrics can be viewed using:
+
+```
+python analysis/proc_eval_result.py policies
+```

agent_zoo/hybrid.py

@@ -0,0 +1,4 @@
+# ruff: noqa
+from .takeru import Policy
+from .takeru import Recurrent
+from .yaofeng import RewardWrapper

agent_zoo/neurips23_start_kit/__init__.py

@@ -0,0 +1,3 @@
+from .baseline_policy import Baseline as Policy
+from .baseline_policy import Recurrent
+from .reward_wrapper import RewardWrapper