Ⅰ. Introduction


    "Taisite-Platform" is an interface automation test platform developed by "Software Testing & Machine Learning Enthusiasts".


    The platform is dedicated to effectively combining artificial intelligence technology with software testing, enabling the platform to be more intelligent and generalized while ensuring test accuracy requirements and optimize the user experience at the same time. The goal is to become the most intimate, best-used, highest value open source test platform.

    Technology stack

    The platform follows the idea of "separate development frontend and backend". The technology stack is: "Python + Vue + Mongodb".

    (Thanks for the inspiration that this open source project brought to me.)

    User environment

    Chrome is a good choice

    Open source protocol


    Ⅱ. Taisite-platform features (os: What is the difference between other test platforms?)

    0. It has been put into production environment for more than 1 year, it is stable~

    1. The platform follows a "small but fine" strategy to maximize the development of all features and cost-effectiveness, helping test teams quickly build easy-to-follow/maintain interfaces   automated test system.

    2. The platform follows the "zero-encoding" principle, and users can complete more complex business process interface tests without programming.

    3. The platform follows the principle of “good-looking is justice” and the operation interface is shown as follows:


    1. The platform has an excellent timed task experience. After starting the scheduled task, you can disable/arbitrarily edit the task content and take effect immediately. It also has a rich alarm strategy.   The page is shown below:


    1. The platform has import/export capabilities that support testers "favorite" Excel format for easy batch generation/modification of use cases.


    1. The platform has a rich test result verification system and supports text similarity verification. (get more detail)

    2. The platform provides a test task scheduling interface to facilitate integration with development projects.

    3. ......

    (There are so many amazing little features waiting for you to explore & tap)

    Ⅲ .Taisite-platform function diagram ( em.. you need to translate by yourself :))



    IV . Deploy

    Deploy under windows

    0. Clone

    git clone

    1. Install python 3 env

    2. deploy NLP model

    Download model

    2.2 Extract the compression package

    2.3 Install python dependent-packages

    pip install tensorflow==1.14.0  -i
    pip install bert-serving-server==1.9.1 -i

    2.4 Start the model

    // Execute after the current directory is switched to the model folder directory

    bert-serving-start -model_dir ./chinese_L-12_H-768_A-12/ -num_worker=1

    After the startup is successful, the output is as follows:


    3. Deploy Mongodb database

    4. Set system environment variables


    Where AUTOTEST_PLATFORM_ENV defaults to production (required)

    AUTOTEST_PLATFORM_MONGO_HOST and AUTOTEST_PLATFORM_MONGO_PORT indicate the address and port of the database (required)

    AUTOTEST_PLATFORM_MONGO_USERNAME and AUTOTEST_PLATFORM_MONGO_PASSWORD represent the account password of the database (if not required)

    AUTOTEST_PLATFORM_NLP_SERVER_HOST (Natural Language Model Service) defaults to native boot (not required)

    AUTOTEST_PLATFORM_MONGO_DEFAULT_DBNAME is the default data table name (required)

    After the setting is completed, you can test it with the following commands (CMD switches to the project root directory)

    python ./backend/

    If the configuration is successful, you can see the input configuration data.

    5. Package the front-end dist file (I have done this for you, skip it if you don't need secondary development)

    5.1 Install the Vue environment, download node.js and configure the environment, download the npm package manager

    5.2 Cmd into the frontend directory, configure cnpm:

    npm install -g cnpm --registry=   

    5.3 Execute the install dependency package command:

    cnpm install

    5.4 Execute the package command:

    cnpm run build

    If successfully packaged, the dist folder will be generated in the project root directory.

    6. Start backend

    // Switch to the project root directory to execute

    pip install -r ./backend/requirements.txt -i

    // Start backend (default 5050 port)

    python ./backend/

    // Create a platform administrator account password

    python ./backend/

    7. Access project

    You can now log in using using the created administrator account password.


    Docker containerized deployment in Linux environment

    0. Clone

    git clone

    1. Natural language model deployment

    sudo -i
    docker pull shaoyuyishiwo/bertserver
    docker run --name autotest-platform-bertserver -d shaoyuyishiwo/bertserver 

    2. Mongo database deployment (skip this step if an existing database is available)

    2.1 Start database & data mount to host

    sudo -i
    docker pull mongo 
    docker run  --name autotest-platform-mongo -p 27017:27017 -v /data/db:/data/db -v /data/configdb:/data/configdb ``-d mongo

    2.2 Create a database account

    docker exec -it autotest-platform-mongo /bin/bash
    > use admin
    switched to db admin
    > db.createUser({user:"${USERNAME}",pwd:"${PASSWORD}",roles:["root"]})
    Successfully added user: { "user" : "admin", "roles" : [ "root" ] }

    2.3 Database memory expansion (recommended)

    > db.adminCommand({setParameter:1, internalQueryExecMaxBlockingSortBytes:335544320})
    { "was" : 33554432, "ok" : 1 }

    3. Environment variable configuration

    // Edit /etc/profile file

    sudo -i
    vi /etc/profile

    If there is a warning, select (E)dit anyway (enter E)

    3.1 Insert the following data at the end of the text (enter i to get into insert status)

    export AUTOTEST_PLATFORM_ENV=production

    The variable is a dynamic value. The deployer can input it according to the actual situation. The DBNAME value can be arbitrarily customized (database table name). The BERT_IPADRESS and MONGO_HOST values can be queried by the following commands:

    docker inspect autotest-platform-bertserver
    docker inspect autotest-platform-mongo // If you used the above steps to deploy the database

    The output is shown below:


    3.2 After inserting, click the ESC button, type :wq and click Enter to save.

    3.3 Environment variables take effect immediately after executing the following command

    source /etc/profile

    4. Start the project

    Before you start the project, you need to change the timezone info by modifying the RUN script in Dockerfile.backend which stay in first-level directory of the project. The default timezone is Asia/Shanghai.

    // Execute the deployment file in the project root directory

    sh deploy ${PORT} 

    The ${PORT} variable fills in the project access port, and the administrator account password is also created when the project starts, as shown in the following figure:


    5. Access project

    The browser can access the ${PORT} port of the deployment server address.



    The following output represents the NLP model startup failure


    Solution steps:

    1. Remove the code from ./backend/app/


    1. Modify the following code in ./backend/testframe/interfaceTest/ to pass:


    When you start the project after you finish, you will not depend on the natural language model~

    V . Contact me

    if you have any questions , feel free to email me ,





    贡献者 4

    泰斯特Test @weixin_41908648
    D dependabot[bot] @dependabot[bot]


    • Vue 52.6 %
    • Python 27.9 %
    • HTML 7.3 %
    • JavaScript 7.3 %
    • CSS 4.9 %