chore: Bumping fastapi + starlette#3938
chore: Bumping fastapi + starlette#3938franciscojavierarceo merged 17 commits intofeast-dev:masterfrom
Conversation
Bumps [fastapi](https://github.com/tiangolo/fastapi) from 0.99.1 to 0.109.1. - [Release notes](https://github.com/tiangolo/fastapi/releases) - [Commits](fastapi/fastapi@0.99.1...0.109.1) --- updated-dependencies: - dependency-name: fastapi dependency-type: direct:production ... Signed-off-by: dependabot[bot] <[email protected]>
012dbd2 to
bcebc63
Compare
Signed-off-by: Chester Ong <[email protected]>
bcebc63 to
b9b8b26
Compare
Signed-off-by: Chester Ong <[email protected]>
|
Linting failed, could you check @bushwhackr ? |
|
Considering upgrading mypy too? 👀 |
|
I'm looking to pin the mypy version to 1.4.1 but like the #3942 there are quite a number of linting failures. Let me see if I can do one with minimum changes. |
Signed-off-by: Chester Ong <[email protected]>
Signed-off-by: Chester Ong <[email protected]>
…tedError Signed-off-by: Chester Ong <[email protected]>
Signed-off-by: Chester Ong <[email protected]>
0ef5334 to
d9ddf13
Compare
|
I'm of the opinion to skip tests mypy type errors till a later PR. Seems like alot of classes in sdk/python/tests/integration/feature_repos/universal/**/* have classes that do not confirm to the BaseClass abstract method definition. For example: subclass: |
Signed-off-by: Chester Ong <[email protected]>
Signed-off-by: Chester Ong <[email protected]>
Signed-off-by: Chester Ong <[email protected]>
Signed-off-by: Chester Ong <[email protected]>
Signed-off-by: Chester Ong <[email protected]>
Signed-off-by: Chester Ong <[email protected]>
|
@sudohainguyen could you rerun the tests. Linting should pass now |
|
great! re-running |
|
oh no 😢 @bushwhackr fail again |
Signed-off-by: Chester Ong <[email protected]>
e1ac385 to
94ced34
Compare
Signed-off-by: Chester Ong <[email protected]>
|
great, now unit tests 😓 |
…concrete Signed-off-by: Chester Ong <[email protected]>
|
@sudohainguyen my bad, please try again. Hopefully I can get this in before lunar new year |
|
Perfect! All checks passed |
sudohainguyen
left a comment
There was a problem hiding this comment.
great work! all problem solved
lgtm
is there any issue in terms of performance we should know? @bushwhackr
None at all, all changes are to fix linting issues. |
|
|
||
| lint-python: | ||
| cd ${ROOT_DIR}/sdk/python; python -m mypy | ||
| cd ${ROOT_DIR}/sdk/python; python -m mypy --exclude=/tests/ --follow-imports=skip feast |
There was a problem hiding this comment.
I would feel this is the biggest change, is that I only limit the linting checks to sdk/python/feast directory
There was a problem hiding this comment.
what if we still check tests? a lot of changes needed right?
|
to revert to the old mypy settings: |
|
alright, let's get rid of it in another PR |
| cursor = execute_snowflake_statement(conn, query) | ||
|
|
||
| if cursor.rowcount < 1 and not_found_exception: | ||
| if cursor.rowcount < 1 and not_found_exception: # type: ignore |
There was a problem hiding this comment.
I tried this and can pass the mypy check:
if cursor.rowcount and (cursor.rowcount < 1) and not_found_exception:
There was a problem hiding this comment.
You're right, this is due to another change I did to the mypy config: --follow-imports=skip.
run the below command to see the error:
python -m mypy --exclude=/tests/ sdk/python/feast/infra/registry/snowflake.py you will see an error.
`sdk/python/feast/infra/registry/snowflake.py:421: error: Unsupported operand types for > ("int" and "None") [operator]
There was a problem hiding this comment.
essentially cursor.rowcount could return none from snowflake's signature
@property
def rowcount(self) -> int | None:
return self._total_rowcount if self._total_rowcount >= 0 else None
|
any blockers? |
|
Need your merge actions here 🙏 @franciscojavierarceo |
I will fix (have fixed) this in my PR. #3942 |


What this PR does / why we need it:
Bumping vuln fastapi version. https://nvd.nist.gov/vuln/detail/CVE-2024-24762
Changes
sdk/python/feastdirectoryWhich issue(s) this PR fixes:
Fixes #3931 #3930