dabla opened a new pull request, #72336:
URL: https://github.com/apache/airflow/pull/72336
<!-- SPDX-License-Identifier: Apache-2.0
https://www.apache.org/licenses/LICENSE-2.0 -->
<!--
Thank you for contributing!
Please provide above a brief description of the changes made in this pull
request.
Write a good git commit message following this guide:
https://chris.beams.io/posts/git-commit/
Please make sure that your code changes are covered with tests.
And in case of new features or big changes remember to adjust the
documentation.
For user-facing UI changes, please attach before/after screenshots (or a
short
screen recording) so reviewers can assess the visual impact.
Feel free to ping (in general) for the review if you do not see reaction for
a few days
(72 Hours is the minimum reaction time you can expect from volunteers) - we
sometimes miss notifications.
In case of an existing issue, reference it using one of the following:
* closes: #ISSUE
* related: #ISSUE
-->
Add deferrable mode to `SFTPOperator`
closes: #65475
## Credits
This PR is a revival of the original work by **SunilKumarReddy**
([@sunildataengineer](https://github.com/sunildataengineer)) in
[apache/airflow#68298](https://github.com/apache/airflow/pull/68298), which was
closed without being merged. All credit for the design and initial
implementation belongs to the original author.
## Summary
`SFTPOperator` now supports a `deferrable=True` parameter that frees the
Airflow worker slot during the file transfer and reclaims it only when the
transfer finishes. This is implemented via a new `SFTPOperationTrigger` that
drives the existing `SFTPHookAsync` to perform the actual I/O in the
Triggerer process.
As a by-product, all synchronous transfer logic (GET / PUT / DELETE,
including
directory handling and concurrency) was refactored into a shared
`SFTPHook.transfer()` method so that the synchronous `execute()` path and the
async `SFTPOperationTrigger` share a single authoritative implementation,
eliminating the duplicate code that existed before.
## Changes
### `providers/sftp/hooks/sftp.py`
- Added `SFTPOperation` enum (`GET`, `PUT`, `DELETE`), moved from
`operators/sftp.py` so both the hook and the trigger can import it without
a circular dependency.
- Added `SFTPHook.transfer()` — a synchronous dispatcher for GET / PUT /
DELETE that handles both single-file and directory transfers (including the
concurrent path). The operator's `execute()` now delegates here entirely.
- Added `SFTPHookAsync.transfer()` — the async counterpart used by
`SFTPOperationTrigger`, built on `asyncssh` and `asyncio.gather` with a
semaphore to honour the `concurrency` parameter.
- Refreshed `_validate_within_directory()` to use `os.path.expanduser` and a
clean `startswith(base + os.sep)` containment check.
### `providers/sftp/operators/sftp.py`
- Removed the now-redundant local `SFTPOperation` class (re-exported from the
hook for backward compat).
- Added `deferrable: bool = False` parameter to `SFTPOperator.__init__`.
- `execute()` is slimmed down: validation is unchanged; the actual file
transfer is a single call to `sftp_hook.transfer()`.
- When `deferrable=True`, `execute()` raises `TaskDeferred` to hand control
to `SFTPOperationTrigger`.
- Added `execute_complete()` callback to resume after the trigger fires,
returning `local_filepath` or raising `AirflowException` on error.
### `providers/sftp/triggers/sftp.py`
- Added `SFTPOperationTrigger` — a new `BaseTrigger` that calls
`SFTPHookAsync.transfer()` inside its `run()` coroutine and yields a
`TriggerEvent` with `{"status": "success"|"error", "message": "..."}`.
- Minor clean-up: replaced `timezone.convert_to_utc` import (which was using
the `common-compat` shim) with a direct import from
`airflow.utils.timezone`.
### `providers/sftp/tests/`
- Extended `test_sftp.py` (operator tests) with coverage for
`deferrable=True`
— verifies that `execute()` raises `TaskDeferred` and that
`execute_complete()` handles both success and error events.
- Added `providers/sftp/newsfragments/68298.feature.rst`.
## Usage
```python
from airflow.providers.sftp.operators.sftp import SFTPOperator, SFTPOperation
upload = SFTPOperator(
task_id="upload_report",
ssh_conn_id="sftp_default",
local_filepath="/tmp/report.csv",
remote_filepath="/uploads/report.csv",
operation=SFTPOperation.PUT,
deferrable=True, # worker slot freed while the transfer runs
)
```
> **Note on directory transfers:** deferrable mode is best suited to single
> large files. For bulk transfers of many files, consider using an async
> `PythonOperator` with `SFTPHookAsync` directly, which provides true async
> multiplexing over a single event loop.
---
##### Was generative AI tooling used to co-author this PR?
<!--
If generative AI tooling has been used in the process of authoring this PR,
please
change below checkbox to `[X]` followed by the name of the tool, uncomment
the "Generated-by".
-->
- [X] Yes (please specify the tool below)
GitHub Copilot (Claude Sonnet 4.6)
<!--
Generated-by: [Tool Name] following [the
guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions)
-->
---
* Read the **[Pull Request
Guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#pull-request-guidelines)**
for more information. Note: commit author/co-author name and email in commits
become permanently public when merged.
* For fundamental code changes, an Airflow Improvement Proposal
([AIP](https://cwiki.apache.org/confluence/display/AIRFLOW/Airflow+Improvement+Proposals))
is needed.
* When adding dependency, check compliance with the [ASF 3rd Party License
Policy](https://www.apache.org/legal/resolved.html#category-x).
* For significant user-facing changes create newsfragment:
`{pr_number}.significant.rst`, in
[airflow-core/newsfragments](https://github.com/apache/airflow/tree/main/airflow-core/newsfragments).
You can add this file in a follow-up commit after the PR is created so you
know the PR number.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
