mengchuan87/MachineWolf

AI Platform performance runner

MachineWolf is transferred to the new project MachineDevil . Please goto and view

---

MachineWolf is a Test Studio for AI 、Deep Learning or Machine Learning framwork、platform.

Quickly Start

• Runing script at local PC

  sudo chmod +x init_dev.sh
  bash ./init_dev.sh
  locust -f ./example/locust/test_http.py --conf ./example/locust/host.conf

• Execute testsuites in docker container

  1. Pull the images from docker-hub

  docker pull banrieen/machinewolf

  2. Start container

  docker run -d     -p 8088:8080  -p 8090:8090     --name "ml-workspace"  -v "${PWD}:/workspace"  --env NOTEBOOK_ARGS="--NotebookApp.notebook_dir=/home"  --shm-size 2048m  --restart always     banrieen/machinewolf:latest
  # Open web IDE
  # http://<xxx.xxx.xxx.xxx>:8088 

• Running locust scripts by taurus

  bzt example/taurus/quick_test.yml

• Running jmeter scripts by taurus

  bzt example/jmeter/trace_user_footprint.jmx

• Running yaml scripts by taurus

  bzt example/taurus/quick_test.yml

• Runing pytest testsuites, such as non-api， HA， throughput test scripts

  pytest example/pytest/test_ha.py

Example of testreport

CLI dashboard

Export testreport

• testreport/result.csv_stats.csv
• testreport/result.csv_stats_history.csv
• testreport/result.csv_failures.csv
• testreport/result.csv_exceptions.csv

Schema of test studio

The test suite is an independent and flexible organization test suite based on the purpose of being inclusive and accommodating extraction. Support a variety of cutting-edge and excellent tools and concepts; currently test schemes (testscheme), data (datas.yaml), scripts (.py, .jmx), and execution plans (host.yml, taurus.yml) are organized flexibly Together. There are still some examples that need to be improved and supplemented.

|-- testhub/
    `-- testscheme
        |-- 5g_manufacturing
        |-- annotations_cvat
    `-- testsuites
        |-- annotations_cvat
            |-- host.conf
            |-- test_cvat_suites.py
            |-- datas.yaml
        |-- dlws
        |-- e2e_aiarts
        |-- ha_aiarts
        |-- jobmanager
        |-- songshanhu
    `-- testlib
        |-- fake_users
        |-- postgres_client
        |-- csv_client

Security

In order to avoid information leaks, invalid information floods.

• All test scripts, explanatory text and configuration files remove all ID, ACCOUNT, HOST information

• Does not retain any test environment information, and any test data

• Replace sensitive information with canonical logos：

  • account： <HOSTNAME>:<PASSWORLD>
  • host： <HOST>:<PORT>
  • link： <LINKTYPE>:<LINKADDRESS>
  • cert： <KEYGEN> 或 <TOKEN>
  • email： <EMAIL-NAME@EMAIL-SERVICE.COM>

Please refer to the release notes for detailsRELEASE。

License

MIT

Comunity