[SYNSD-1233] Adding in a try/catch for HTTP 409 conflicts on entity creation

[BryanFauble]

Problem:

1. If an HTTP Read error occurs when receiving back the response from creating a Synapse entity an HTTP 409 error will be thrown when the request is retried.

Solution:

1. Try out removing the ReadError and ReadTimeout as exceptions that are retried. Instead let the SynapseHTTPError be raised and wrap the POST call in a try/except that will get the entity from Synapse if it was a conflict, then compare the results to what the expected state was.

Testing:

1. I did very light manual testing for this by running the syncToSynapse command and I manually triggered the exception to be raised.

[pep8speaks]

Hello @BryanFauble! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file synapseclient/api/entity_services.py:

Line 67:89: E501 line too long (225 > 88 characters)
Line 72:89: E501 line too long (113 > 88 characters)
Line 76:89: E501 line too long (187 > 88 characters)

• In the file tests/integration/synapseclient/models/async/test_file_async.py:

Line 11:1: F401 'httpcore._backends.anyio.AnyIOStream' imported but unused
Line 161:89: E501 line too long (221 > 88 characters)

Comment last updated at 2024-09-27 19:12:34 UTC

[thomasyu888]

👀 Does that mean if we add functions called in the async stream that we would have edit this call_count?

[BryanFauble]

Yeah, if more or less calls occur during the store_async function we would need to adjust this call count.

This specific way of replacing the read isn't working on GH actions. Though I did confirm the logic is working on my local machine. This needs to be fixed, but more digging into this is needed. I can get the hostname/IP addresses used, but being this "far" down in the libraries and working directly in httpcore, I lose a lot of context like request headers and the uri.