fix table.close in batch_roi_export_to_table.py

[will-moore]

See https://github.com/ome/training-scripts/pull/89/files#r688333998

This fixes a bug in batch_roi_export_to_table.py since we must now do table.getOriginalFile() before closing the table.

To test,

• Upload script to a server
• Run from webclient, choosing a Project, Dasetet or a few Images - *Script is designed to work with idr0021 data
• Bulk-annotation table should be linked to the parent Project

[github-actions]

👈 Launch a binder notebook on branch fix_table_close_in_script

[pwalczysko]

Uploaded to outreach server. Ran the script on a dataset with 3 images.
I am getting a csv created and attached successfully to the Dataset, but no table because

Traceback (most recent call last):
  File "./script", line 439, in <module>
    run_script()
  File "./script", line 422, in run_script
    result = batch_roi_export(conn, script_params)
  File "./script", line 343, in batch_roi_export
    save_table(conn, images, image_data, script_params, project)
  File "./script", line 253, in save_table
    table.initialize(data)
  File "/opt/omero/server/venv3/lib64/python3.6/site-packages/omero_Tables_ice.py", line 1347, in initialize
    return _M_omero.grid.Table._op_initialize.invoke(self, ((cols, ), _ctx))
ValueError: invalid value for element 0 of sequence<double>

is reported in the exclamation mark icon

[will-moore]

Apologies, now that I read the script again, I realise this is designed to work with idr0021 data.
Probably for other data it's trying to choose channels that don't exist etc, and getting None values in the table.

I just uploaded the script to merge-ci (via webclient, logged-in as root) - it's under a test_pr_33 script menu.
And ran it against 1 Dataset from idr0021: https://merge-ci.openmicroscopy.org/web/webclient/?show=dataset-3184
This created a csv and OMERO.table on the parent Project.

[joshmoore]

This certainly matches the change that was needed elsewhere. 👍

[will-moore]

Try/finally in 2d4e600

[jburel]

Thanks @will-moore