Doesn't work with Lambda in Docker locally

[j-adamczyk]

I can't manage to get Mangum to work with AWS Lambda in Docker. Everything works if I just run it as a Python process, but not in Docker container.

Error:

{"errorMessage": "The adapter was unable to infer a handler to use for the event. This is likely related to how the Lambda function was invoked. (Are you testing locally? Make sure the request payload is valid for a supported handler.)", "errorType": "RuntimeError", "requestId": "8a98acd4-aab9-47bd-a21c-180f2c3fb483", "stackTrace": ["  File \"/var/lang/lib/python3.10/site-packages/mangum/adapter.py\", line 76, in __call__\n    handler = self.infer(event, context)\n", "  File \"/var/lang/lib/python3.10/site-packages/mangum/adapter.py\", line 68, in infer\n    raise RuntimeError(  # pragma: no cover\n"]}

I'm also getting the same error when I deploy on AWS Lambda.

My server:

import joblib
import uvicorn
from api_models import Properties
from fastapi import FastAPI
from mangum import Mangum
from starlette.middleware.cors import CORSMiddleware


app = FastAPI()
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

model = joblib.load(filename="./saved_models/current_model.pkl")


@app.get("/")
def read_root():
    return {"Hello": "World"}


@app.post("/make_prediction")
def make_prediction(properties: Properties):
    # some calculations
    return result


handler = Mangum(app)


if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=8000)

Dockerfile:

FROM public.ecr.aws/lambda/python:3.10

RUN yum install libgomp git -y && \
    yum clean all -y && \
    rm -rf /var/cache/yum

COPY requirements.txt ${LAMBDA_TASK_ROOT}
RUN pip install -r requirements.txt

COPY serving/backend ${LAMBDA_TASK_ROOT}
COPY saved_models ${LAMBDA_TASK_ROOT}/saved_models

CMD ["server.handler"]

I run it as docker run -p 9000:8080 prediction_lambda. I want to deploy this as Lambda Function URL, which uses API Gateway 2.0 style requests. My request:

import requests

r = requests.get(
    url="http://localhost:9000/2015-03-31/functions/function/invocations",
    json={
        "version": "2.0",
        "routeKey": "GET /",
        "rawPath": "/",
        "pathParameters": {}
    }
)
print(r.text)

So I'm tring to reach at least the / endpoint with GET to debug this, but eventually for production I want to use the POST method.

How can I fix this?

[DolevAlgam]

I get the same error.

[bencwallace]

You're missing the request context. I got this to work:

import requests

json = {
  "version": "2.0",
  "routeKey": "GET /",
  "rawPath": "/",
  "pathParameters": {},
  "requestContext": {
    "http": {
      "sourceIp": "192.0.0.1",
      "path": "/",
      "method": "GET"
    }
  }
}
r = requests.get(url="http://localhost:9000/2015-03-31/functions/function/invocations", json=json)
print(r.text)