Skip to main content

A gym environemnt for Malmo

Project description

# MalmoEnv #

MalmoEnv is an OpenAI “gym” Python Environment for Malmo/Minecraft, directly implemented Python to Java Minecraft.

A Python “gym env” can be created and used to run an agent in a Malmo mission. Each such env has a remote Minecraft instance associated to it (by DNS name or IP and Port). For multi-agent missions, the first agent’s (role 0) Minecraft client instance is used as a coordinator to allow all agents to rendezvous on mission starts (i.e. on env resets).

As it’s pure Python, you just need this one package, its direct dependencies and (Java) Minecraft! Example missions, including some from the 2018 MarLo competition can be found in the “missions” directory.

## Examples of use: ##

Install dependencies:

Java8 JDK ([AdoptOpenJDK](https://adoptopenjdk.net/)), python3, git

pip3 install gym lxml numpy pillow

To prepare Minecraft (after cloning this repository with git clone https://github.com/Microsoft/malmo.git):

cd malmo/Minecraft

(echo -n “malmomod.version=” && cat ../VERSION) > ./src/main/resources/version.properties

Running a single agent example mission (run each command in different cmd prompt/shells - use launchClient.bat on Windows):

./launchClient.sh -port 9000 -env or (On Windows) launchClient.bat -port 9000 -env

(In another shell) cd malmo/MalmoEnv optionally run python3 setup.py install

python3 run.py –mission missions/mobchase_single_agent.xml –port 9000 –episodes 10

A two agent example mission (run each command in different cmd prompt/shells):

./launchClient.sh -port 9000 -env

./launchClient.sh -port 9001 -env

In the two agent case, running each agent in it’s own shell, the run script (for agents other than the first) is given two ports - the first for the mission coordinator and a second (port2) for the other agent’s Minecraft:

python3 run.py –mission missions/mobchase_two_agents.xml –port 9000 –role 0 –experimentUniqueId “test1”

python3 run.py –mission missions/mobchase_two_agents.xml –port 9000 –port2 9001 –role 1 –experimentUniqueId “test1”

## Running multi-agent examples using multiple Python threads: ##

python3 runmultiagent.py –mission missions/mobchase_two_agents.xml

## Installing with pip ##

MalmoEnv is available as a pip wheel.

If you install with pip3 install malmoenv then you can download the Minecraft mod (assuming you have git available from the command line) with:

python3 -c “import malmoenv.bootstrap; malmoenv.bootstrap.download()”

The sample missions will be downloaded to ./MalmoPlatform/MalmoEnv/missions.

python3 -c “import malmoenv.bootstrap; malmoenv.bootstrap.launch_minecraft(9000)” can be used to start up the Malmo Minecraft Mod listening for MalmoEnv connections on port 9000 after downloading Malmo.

To test: cd MalmoPlatform/MalmoEnv; python3 runmultiagent.py –mission missions/mobchase_single_agent.xml

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

malmoenv-0.0.8.tar.gz (17.1 kB view details)

Uploaded Source

Built Distribution

malmoenv-0.0.8-py3-none-any.whl (30.7 kB view details)

Uploaded Python 3

File details

Details for the file malmoenv-0.0.8.tar.gz.

File metadata

  • Download URL: malmoenv-0.0.8.tar.gz
  • Upload date:
  • Size: 17.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.13.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/20.7.0 requests-toolbelt/0.9.1 tqdm/4.35.0 CPython/3.5.2

File hashes

Hashes for malmoenv-0.0.8.tar.gz
Algorithm Hash digest
SHA256 69fa0d4e6ce3a13a390c0c4d5b3e2257c1af46288d466971f1c64944ffd0d4a1
MD5 2742598ff927f86f560187e0fa7382f3
BLAKE2b-256 e07ff86d196b485ecb29befe1c39457fc1ebf572bf4cbea4254b12af175b52bb

See more details on using hashes here.

File details

Details for the file malmoenv-0.0.8-py3-none-any.whl.

File metadata

  • Download URL: malmoenv-0.0.8-py3-none-any.whl
  • Upload date:
  • Size: 30.7 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.13.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/20.7.0 requests-toolbelt/0.9.1 tqdm/4.35.0 CPython/3.5.2

File hashes

Hashes for malmoenv-0.0.8-py3-none-any.whl
Algorithm Hash digest
SHA256 4c65b8052274b480773b0d87e089a3d16945f67073d6987541fb65655e17cc2b
MD5 5e622a1f3fa33449159af309b047b5dd
BLAKE2b-256 cacc78423b811083ae56dd8d3a46f3a3bcacdf8b8f5c1ff36650ea9413092722

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page