Allow top-level await in Jupyter notebooks

[charliermarsh]

Summary

Top-level await is allowed in Jupyter notebooks (see: autoawait).

Closes #6584.

Test Plan

Had to test this manually. Created a notebook, verified that the yield was flagged but the await was not.

[github-actions]

PR Check Results

Ecosystem

✅ ecosystem check detected no changes.

Benchmark

Linux

group                                      main                                    pr
-----                                      ----                                    --
formatter/large/dataset.py                 1.00      4.6±0.28ms     8.8 MB/sec     1.02      4.7±0.40ms     8.6 MB/sec
formatter/numpy/ctypeslib.py               1.00   931.4±41.45µs    17.9 MB/sec     1.01   941.0±45.45µs    17.7 MB/sec
formatter/numpy/globals.py                 1.00     92.0±4.79µs    32.1 MB/sec     1.01     92.7±6.41µs    31.8 MB/sec
formatter/pydantic/types.py                1.00  1822.9±113.63µs    14.0 MB/sec    1.01  1845.5±101.59µs    13.8 MB/sec
linter/all-rules/large/dataset.py          1.00     14.2±0.36ms     2.9 MB/sec     1.08     15.4±0.55ms     2.6 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      3.7±0.19ms     4.5 MB/sec     1.05      3.8±0.15ms     4.3 MB/sec
linter/all-rules/numpy/globals.py          1.00   517.8±16.02µs     5.7 MB/sec     1.01   520.9±24.41µs     5.7 MB/sec
linter/all-rules/pydantic/types.py         1.00      7.3±0.20ms     3.5 MB/sec     1.08      7.9±0.29ms     3.2 MB/sec
linter/default-rules/large/dataset.py      1.00      7.3±0.16ms     5.6 MB/sec     1.07      7.8±0.21ms     5.2 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00  1591.8±53.96µs    10.5 MB/sec     1.01  1602.9±80.78µs    10.4 MB/sec
linter/default-rules/numpy/globals.py      1.02   197.6±14.95µs    14.9 MB/sec     1.00    193.4±7.80µs    15.3 MB/sec
linter/default-rules/pydantic/types.py     1.00      3.2±0.09ms     7.9 MB/sec     1.05      3.4±0.13ms     7.5 MB/sec

Windows

group                                      main                                   pr
-----                                      ----                                   --
formatter/large/dataset.py                 1.00      4.1±0.02ms     9.9 MB/sec    1.00      4.1±0.06ms     9.9 MB/sec
formatter/numpy/ctypeslib.py               1.00   781.1±13.61µs    21.3 MB/sec    1.01   785.9±16.21µs    21.2 MB/sec
formatter/numpy/globals.py                 1.00     78.8±1.07µs    37.4 MB/sec    1.01     79.7±1.58µs    37.0 MB/sec
formatter/pydantic/types.py                1.00  1599.5±18.37µs    15.9 MB/sec    1.01  1617.0±42.65µs    15.8 MB/sec
linter/all-rules/large/dataset.py          1.00     12.7±0.10ms     3.2 MB/sec    1.00     12.8±0.19ms     3.2 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      3.6±0.02ms     4.7 MB/sec    1.01      3.6±0.02ms     4.7 MB/sec
linter/all-rules/numpy/globals.py          1.00    373.2±6.70µs     7.9 MB/sec    1.02    381.2±8.47µs     7.7 MB/sec
linter/all-rules/pydantic/types.py         1.00      6.6±0.04ms     3.9 MB/sec    1.01      6.7±0.07ms     3.8 MB/sec
linter/default-rules/large/dataset.py      1.00      7.0±0.04ms     5.8 MB/sec    1.02      7.1±0.05ms     5.7 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00  1459.9±16.50µs    11.4 MB/sec    1.01  1477.0±28.46µs    11.3 MB/sec
linter/default-rules/numpy/globals.py      1.00    150.9±2.53µs    19.6 MB/sec    1.00    150.6±3.05µs    19.6 MB/sec
linter/default-rules/pydantic/types.py     1.00      3.1±0.04ms     8.1 MB/sec    1.01      3.2±0.02ms     8.1 MB/sec

[dhruvmanila]

Looks good, thanks!

For the long term we might want to check if autoawait is disabled or not otherwise we might give false negative.

[charliermarsh]

Makes sense. I think the false positive is more annoying than the false negative is worrisome, so let's proceed.