使用SAM Local Invoke时无法导入app模块的问题求助

项目信息 #

项目结构（执行tree -L 2输出） #

.
├── Dockerfile
├── envs
│   ├── dev.json
│   └── prd.json
├── event.json
├── requirements.txt
├── src
│   ├── app.py
│   └── utils.py
└── template.yml

Dockerfile内容 #

# Specify the base image
FROM public.ecr.aws/lambda/python:3.11 AS build-stage

# Install build dependencies for psutil (pyproject.toml)
RUN yum install -y gcc python3-devel gcc-c++

# Set the working directory inside the container
WORKDIR /app

# Copy requirements.txt from the root of your project
COPY requirements.txt .

# Install Python dependencies without plotly to a temporary directory
RUN pip3 install --no-cache-dir --upgrade pip && \
    pip3 install --no-cache-dir --upgrade -r requirements.txt --target /tmp/python-deps

# -------------------- Stage 2 --------------------
# Build the final stage
FROM public.ecr.aws/lambda/python:3.11 AS final-stage

# Remove unnecessary dependencies (optional, but recommended)
RUN yum remove -y gcc python3-devel gcc-c++ && \
    yum clean all

# Copy the Python dependencies from the build-stage
COPY --from=build-stage /tmp/python-deps /app

# Copy source code
COPY src /app/src

# Argument to specify environment
ARG ENVIRONMENT=dev
ENV ENVIRONMENT=${ENVIRONMENT}

# Copy the appropriate environment file based on the argument
COPY envs/${ENVIRONMENT}.json /app/.env

# Set the entry point for the Lambda function
CMD [ "app.lambda_handler" ]

template.yml内容 #

AWSTemplateFormatVersion: '2010-09-09'
Transform: AWS::Serverless-2016-10-31
Description: AWS SAM template for testing Docker-based Lambda function

Parameters:
  Environment:
    Type: String
    Description: The environment (dev or prd)
    Default: dev
    AllowedValues: [dev, prd]

Globals:
  Function:
    Timeout: 180
    MemorySize: 256

Resources:
  DockerDemo:
    Type: AWS::Serverless::Function
    Properties:
      CodeUri: src/
      PackageType: Image
       ImageConfig:
         Command: [app.lambda_handler]
      Environment:
        Variables:
          ENVIRONMENT: !Ref Environment
      ImageUri: !Sub docker-demo:${Environment}
    Metadata:
      DockerTag: !Ref Environment
      DockerContext: .
      Dockerfile: Dockerfile
      DockerBuildArgs:
        ENVIRONMENT: !Ref Environment
      DockerEnvFiles:
        - envs/${Environment}.json

app.py内容 #

# Importing our utils
import json
import utils

def lambda_handler(event : dict, context : dict) -> dict:
    try:
        
        env_vars = utils.load_env()
        utils.load_mysql(env_vars = env_vars)
        
        return {
            'status' : 200,
            'message' : 'Everything worked nice!'
        }
        
    except Exception as e:
        return {
            'status' : 400,
            'message' : f'Something went wrong: \'{e}\''
        }

执行命令与错误 #

1. 执行sam build --use-container，构建成功。
2. 执行sam local invoke DockerDemo --event event.json，出现以下错误：

Invoking Container created from dockerdemo:dev
Local image was not found.
Removing rapid images for repo dockerdemo
Building image..................
Using local image: dockerdemo:rapid-x86_64.

START RequestId: b110d6b1-81bd-4c13-9cfe-27692c1a37b5 Version: $LATEST
Traceback (most recent call last): Unable to import module 'app': No module named 'app'
END RequestId: b110d6b1-81bd-4c13-9cfe-27692c1a37b5
REPORT RequestId: b110d6b1-81bd-4c13-9cfe-27692c1a37b5  Init Duration: 1.02 ms  Duration: 435.02 ms     Billed Duration: 436 ms Memory Size: 256 MB     Max Memory Used: 256 MB
{"errorMessage": "Unable to import module 'app': No module named 'app'", "errorType": "Runtime.ImportModuleError", "requestId": "b110d6b1-81bd-4c13-9cfe-27692c1a37b5", "stackTrace": []}% 

环境：MacBook Pro M2（Apple Silicon芯片），无架构适配问题。

问题根源 #

Lambda Python运行时默认从/var/task目录加载模块，但当前Dockerfile将代码放在/app/src目录，Python无法识别该路径下的app模块；同时template.yml中PackageType: Image与CodeUri配置存在冲突。

修复步骤 #

1. 调整Dockerfile，适配Lambda运行时规范 #

修改Dockerfile，将代码和依赖包放到Lambda默认的/var/task目录：

# Specify the base image
FROM public.ecr.aws/lambda/python:3.11 AS build-stage

# Install build dependencies for psutil (pyproject.toml)
RUN yum install -y gcc python3-devel gcc-c++

# Set the working directory inside the container
WORKDIR /app

# Copy requirements.txt from the root of your project
COPY requirements.txt .

# Install Python dependencies without plotly to a temporary directory
RUN pip3 install --no-cache-dir --upgrade pip && \
    pip3 install --no-cache-dir --upgrade -r requirements.txt --target /tmp/python-deps

# -------------------- Stage 2 --------------------
# Build the final stage
FROM public.ecr.aws/lambda/python:3.11 AS final-stage

# Remove unnecessary dependencies (optional, but recommended)
RUN yum remove -y gcc python3-devel gcc-c++ && \
    yum clean all

# Copy Python dependencies to Lambda's default task directory
COPY --from=build-stage /tmp/python-deps /var/task

# Copy source code to Lambda's default task directory
COPY src /var/task

# Argument to specify environment
ARG ENVIRONMENT=dev
ENV ENVIRONMENT=${ENVIRONMENT}

# Copy the appropriate environment file to task directory
COPY envs/${ENVIRONMENT}.json /var/task/.env

# Set the entry point for the Lambda function
CMD [ "app.lambda_handler" ]

2. 修正template.yml的冲突配置 #

使用镜像打包时，无需指定CodeUri，简化配置：

AWSTemplateFormatVersion: '2010-09-09'
Transform: AWS::Serverless-2016-10-31
Description: AWS SAM template for testing Docker-based Lambda function

Parameters:
  Environment:
    Type: String
    Description: The environment (dev or prd)
    Default: dev
    AllowedValues: [dev, prd]

Globals:
  Function:
    Timeout: 180
    MemorySize: 256

Resources:
  DockerDemo:
    Type: AWS::Serverless::Function
    Properties:
      PackageType: Image
      Environment:
        Variables:
          ENVIRONMENT: !Ref Environment
      ImageUri: !Sub docker-demo:${Environment}
    Metadata:
      DockerTag: !Ref Environment
      DockerContext: .
      Dockerfile: Dockerfile
      DockerBuildArgs:
        ENVIRONMENT: !Ref Environment

3. 重新构建并测试 #

执行以下命令验证修复效果：

sam build --use-container
sam local invoke DockerDemo --event event.json

额外说明 #

• Lambda Python运行时会自动将/var/task加入Python路径，遵循此规范可避免手动配置路径的麻烦。
• 若需保留/app目录，也可通过添加ENV PYTHONPATH=/app/src让Python找到模块，但推荐使用Lambda默认目录以减少额外配置。

内容的提问来源于stack exchange，提问作者AbdoCherry