Use pytest-backend-plugin to test against SaaS and on-prem and fix putfile SaaS bug

[tkilias]

This issue is more related to cloudstorage-extension setup in the ai lab, rather than the AI Lab itself.

Proposed fix:

• use pytest plugin in notebook-connector
• then the error will be observable in the notebook-connector

Background

In https://github.com/exasol/ai-lab/actions/runs/10525178182/job/29163449570 we had the following error

E           ------------------
E           from exasol.nb_connector import github, bfs_utils, cloud_storage
E           from exasol.nb_connector.connections import open_bucketfs_connection, open_pyexasol_connection
E           
E           jar_local_path = github.retrieve_jar(github.Project.CLOUD_STORAGE_EXTENSION, use_local_cache=True)
E           bfs_bucket = open_bucketfs_connection(ai_lab_config)
E           bfs_path = bfs_utils.put_file(bfs_bucket, jar_local_path)
E           
E           with open_pyexasol_connection(ai_lab_config) as conn:
E               cloud_storage.setup_scripts(conn, ai_lab_config.db_schema, bfs_path)
E               
E           print("Could Storage Extension was initialized")
E           ------------------
E           
E           
E           ---------------------------------------------------------------------------
E           TypeError                                 Traceback (most recent call last)
E           Cell In[2], line 6
E                 4 jar_local_path = github.retrieve_jar(github.Project.CLOUD_STORAGE_EXTENSION, use_local_cache=True)
E                 5 bfs_bucket = open_bucketfs_connection(ai_lab_config)
E           ----> 6 bfs_path = bfs_utils.put_file(bfs_bucket, jar_local_path)
E                 8 with open_pyexasol_connection(ai_lab_config) as conn:
E                 9     cloud_storage.setup_scripts(conn, ai_lab_config.db_schema, bfs_path)
E           
E           File ~/jupyterenv/lib/python3.10/site-packages/exasol/nb_connector/bfs_utils.py:24, in put_file(bucket, file_path, skip_if_exists)
E                22     raise ValueError(f"Local file doesn't exist: {file_path}")
E                23 local_name = file_path.name
E           ---> 24 if skip_if_exists and local_name in list(bucket):
E                25     _logger.info("File %s is already present in the bucketfs", local_name)
E                26 else:
E           
E           TypeError: 'SaaSBucket' object is not iterable

job-logs.txt

The reason that we recognized the error that late was first, that test was disabled in the ai-lab, but also because we don't test here against saas.

Acceptance Criteria

• use pytest-backend-plugin for testing agains onprem and SaaS
• fix bug