Add special case to time_str_to_utc to handle invalid tz names

[sam-w]

Description

See #3921

tl;dr: malformed date strings are being returned from the Gmail API and being passed to the time_str_to_utc function here:

onyx/backend/onyx/connectors/gmail/connector.py

Line 182 in 5c7487e

updated_at_datetime = time_str_to_utc(updated_at)

This is the least obtrusive solution I could come up with:

• if a datestring doesn't parse, see if it contains an offset followed by something in parens (+1234 (SOMETHING))
• if the thing in parens is not a valid tz name, remove it parse the date with just the offset

Another option would be to use fuzzy parsing (i.e. dt = parse(datetime_str, fuzzy=True)) but that risks causing surprising behaviour elsewhere, given how widely this util is used.

How Has This Been Tested?

Added unit tests.

Backporting (check the box to trigger backport action)

Note: You have to check that the action passes, otherwise resolve the conflicts manually and tag the patches.

• This PR should be backported (make sure to check that the backport attempt succeeds)
• [Optional] Override Linear Check

[vercel]

@sam-w is attempting to deploy a commit to the Danswer Team on Vercel.

A member of the Team first needs to authorize it.