Change polling for progress logging to exponential backoff

[FlorentinD]

No description provided.

[netlify]

✅ Deploy Preview for neo4j-graph-data-science-client canceled.

Name	Link
🔨 Latest commit	1a40613
🔍 Latest deploy log	https://app.netlify.com/sites/neo4j-graph-data-science-client/deploys/680ba47d952b0500092729bd

[Copilot]

Pull Request Overview

This PR changes the polling mechanism for progress logging to use exponential backoff, reducing the number of calls to check for progress updates while upgrading the Neo4j image version in the test environment.

• Update Neo4j image version in compose.yml
• Introduce exponential backoff using tenacity’s wait_exponential in several modules
• Add a test for validating the new logging interval behavior

Reviewed Changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
scripts/test_envs/gds_plugin_enterprise/compose.yml	Updated Neo4j image to neo4j:5-enterprise
graphdatascience/tests/unit/query_runner/progress/test_query_progress_logger.py	Added test_log_interval verifying progress logging interval
graphdatascience/retry_utils/retry_utils.py	Introduced new retry_until_future class for async retry based on a Future
graphdatascience/query_runner/session_query_runner.py	Updated progress logger to use wait_exponential for polling
graphdatascience/query_runner/progress/query_progress_logger.py	Switched polling interval to exponential backoff and adjusted progress bar parameters
graphdatascience/query_runner/neo4j_query_runner.py	Updated progress logger instantiation to use exponential backoff
changelog.md	Updated changelog reflecting reduced progress update calls

Comments suppressed due to low confidence (3)

graphdatascience/retry_utils/retry_utils.py:24

• [nitpick] The class name 'retry_until_future' does not follow the standard PascalCase naming convention; consider renaming it to 'RetryUntilFuture'.

class retry_until_future(retry_base):

graphdatascience/query_runner/progress/query_progress_logger.py:40

• [nitpick] The error message refers to 'polling interval' while the parameter is named 'log_interval'; consider updating the error message for clarity.

raise ValueError("polling interval must be a float or an instance of wait_base")

graphdatascience/tests/unit/query_runner/progress/test_query_progress_logger.py:43

• The use of time.sleep(0.5) requires importing the time module. Please add 'import time' at the top of the file.

time.sleep(0.5)

[FlorentinD]

I think we should run the progress logger also inside the threadpool.

Check every 0.5s if the main query still runs + cancel progress thread if its done.
right now we might delay the execution until 10s