You can use the custom runtime of Function Compute to write functions in Python. This topic describes how to deploy and invoke a Python function. You can modify the sample Python code as required.

Background information

Before you begin, make sure that the following operations are complete:
  • Install Node.js 8.6.0 or later. For more information, visit Downloads.
  • Install Docker. For more information, visit Docker Hub.

Step 1: Prepare the environment

The easiest way to install Funcraft is to download the executable binary files.

  1. Install Funcraft on an on-premises machine. For more information, see Install Funcraft.
  2. Run the fun --version command to check whether the installation is successful.
  3. Run the fun config command to configure Funcraft. Set the Account ID, Access Key ID, Access Key Secret, and Default region name parameters as prompted.
    fun config
    Aliyun Account ID 1234xxx
    Aliyun Access Key ID xxxx
    Aliyun Access Key Secret xxxx
    Default region name cn-xxxx
    The timeout in seconds for each SDK client invoking 300
    The maximum number of retries for each SDK client 5
    Allow to anonynously report usage statistics to improve the tool over time? (Y/n)

Step 2: Deploy and invoke a function

  1. Run the following command to clone the sample project to the on-premises machine:
    git clone          
    Note If you do not install Git, you can download the sample project from
  2. Optional:Modify the sample code in the python37-demo directory to implement your business logic.
  3. Run the following commands to go to the cloned sample project:
    1. Run the following command to go to the fc-custom-demo directory:
      cd fc-custom-demo
    2. Run the following command to go to the python37-demo directory:
      cd python37-demo    
  4. Run the following command to install dependencies:
    fun install -v   
    The following result is returned:
    Digest: sha256:790b073005953d1b41eef2b7799e3d24381d9fe3e1e2bd48101530ce996b****
    Status: Downloaded newer image for
     ---> 2479a0c4aab6
    Step 2/3 : RUN fun-install pip -i install flask
     ---> Running in 4aa275ae7e33
    Task => PipTask
         => PYTHONUSERBASE=/code/.fun/python pip install --user --upgrade flask -i
    Removing intermediate container 4aa275ae7e33
     ---> 7d100a3ab120
    Step 3/3 : RUN fun-install pip -i install gunicorn
     ---> Running in d6a2de957cf3
    Task => PipTask
         => PYTHONUSERBASE=/code/.fun/python pip install --user --upgrade gunicorn -i
    Removing intermediate container d6a2de957cf3
     ---> f4f101691d37
    Successfully built f4f101691d37
    Successfully tagged fun-cache-a5ed3048-bf7f-48db-bd5d-371592c01059:latest
    copying function artifact to D:\fc-custom-demo\python37-demo
    Install Success 
  5. Run the following command to deploy the project to Function Compute:
    fun deploy -y      
    The following result is returned:
    Waiting for service python37-demo to be deployed...
            Waiting for function fc-python37 to be deployed...
                    Waiting for packaging function fc-python37 code...
                    The function fc-python37 has been packaged. A total of 2 files were compressed and the final size was 641 B
            function fc-python37 deploy success
    service python37-demo deploy success
  6. Run the following command to invoke the deployed function:
    fun invoke -e "Hello World"          
    Note If you use the Windows operating system, you must convert the format of the bootstrap file to UNIX.
    The following result is returned:
    ========= FC invoke Logs begin =========
    FC Invoke Start RequestId: cd30369e-7dfa-439c-a68d-7fe16d5a7e05
    Hello World
    FC Invoke End RequestId: cd30369e-7dfa-439c-a68d-7fe16d5a7e05
    Duration: 54.13 ms, Billed Duration: 100 ms, Memory Size: 512 MB, Max Memory Used: 133.70 MB
    ========= FC invoke Logs end =========
    FC Invoke Result:
    Hello World