In an unrelated PR #57, test coverage dropped to <90, so adding some tests to bump it up.

Also contains, nits (improper spacing) on exception messages for sync methods of async interfaces.

Why did the test pass? Is our codecoverage setup different than langchain that errors out?

@averikitsch the tests indeed passed now despite having coverage<90. Example, recent cloud build is green when the logs suggested:

FAIL Required test coverage of 90.0% not reached. Total coverage: 89.89%

My hypothesis is that this happens due to rounding issue with pytest, (89.89 ==> 90). I made a github PR to assert that our code coverage configs are indeed correct and tests fail if coverage drops below the 89% mark, and they do fail when coverage is 88.76%.

Ref: #63

Probably the default value (zero) of precision in pytest-cov plugin is to blame and a solution could be to set precision in addition to fail_under in the .coveragerc file.

But its an overkill for just 1% coverage.

Thank you!