Also allow passing buffer instead of path for retrieve_file and store_file methods in SFTPHook #44247
+70
−22
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In the FTPSHook it's also possible to pass a buffer (like StringIO/BytesIO) instead of a path, I've refactored the retrieve_file and store_file methods in SFTPHook so those also support those types. This avoids the need to create a file to be able to download a file to FTP, you can directly download it into a buffer and vice versa if you need to upload content to an FTP server you don't need to create a file locally to be able to upload it. I just delegate this to the underlying paramiko getfo and putfo methods, nothing fancy had to be done to achieve this as the paramiko library already supported this. This shortcoming was described as a pitfall for the SFTPHook.
Also using our in-house StreamOperator which allows executing an operator in a multi-threaded way within the same worker (similar to expand but then distributing the work over multiple threads within the same worker instead of multiple workers), we discovered that when using the SFTPHook in a PythonOperator to retrieve a file, the connections don't get closed, which seems probable as each thread has to instantiate a new SFTPHook as the underlying library paramiko isn't thread-safe. When you look at the source code you'll notice that each operation within the SFTPHook get's a connection but actually never closes it as you have to call the close_conn method manually.
In the past we didn't use the SFTPHook but used directly the paramiko library like this and didn't have any issues:
So I think we should refactor the SFTP hook to allow it being used in a context manager to make sure the ftp connection is closed after each call so we can invoke it like this:
Otherwise we would need to call it like this, in the above you don't need to worry about the connection:
I've added a test case for both changes.
^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named
{pr_number}.significant.rst
or{issue_number}.significant.rst
, in newsfragments.