Description
Use case
I updated the schema of my application, changing the amount field from int to float to comply with my company's requirements. However, I noticed that the default JSON parser was losing floating-point precision during requests.
def test_api_gateway_resolver_numeric_value():
app = ApiGatewayResolver()
@app.post("/my/path")
def test_handler():
return app.current_event.json_body
# WHEN calling the event handler
event = {}
event.update(LOAD_GW_EVENT)
event["body"] = '{"amount": 2.2999999999999998}'
event["httpMethod"] = "POST"
result = app(event, {})
# THEN process event correctly
assert result["statusCode"] == 200
assert result["multiValueHeaders"]["Content-Type"] == [content_types.APPLICATION_JSON]
assert result["body"] == '{"amount": 2.2999999999999998}' # assert '{"amount":2.3}' == '{"amount": 2.2999999999999998}'Solution/User Experience
I've already implemented a solution and will be opening a PR soon.
The fix involves passing a custom deserializer instance to the ApiGatewayResolver constructor to handle floating-point precision correctly.
Alternative solutions
Acknowledgment
- This feature request meets Powertools for AWS Lambda (Python) Tenets
- Should this be considered in other Powertools for AWS Lambda languages? i.e. Java, TypeScript, and .NET
Metadata
Metadata
Assignees
Type
Projects
Status
Working on it