README needs to mention about installation of scanreader, sbxreader, suite2p.

1. Currently, the full suite of pytests is very time consuming (taking over an hour), as we insert and delete between each test. By adding fixture scope, we could reduce this time significantly.

@pytest.fixture(scope="session")
def ingest_subjects(pipeline, subjects_csv):
    pass # keep as is

2. Currently, there are DuplicateErrors if _tear_down=False. By adding skip_duplicates to insert functions or pytest markers (e.g., skipif or xfail), we could make better use of non-teardown iterations.

• Data provider will need to provide a license. One option are the Creative Commons Licenses for data sample sharing (CC BY-SA 4.0).
• Provide guidelines for data usage.
• For citation purposes, consider creating a DOI for the dataset.

In testing workflow-session, we discovered that the children() method was inconsistent in the order it reported. This may impact the the parallel parent() method here. This can be swapped out for

assert schema.Table.full_table_name in schema.OtherTable.parents()

This issue was first encountered in the 03-process.ipynb

The import statement from workflow_calcium_imaging.pipeline import * does not import the Equipment table and it needs to be its own import statement.

test_find_valid_full_path and test_find_root_directory functions fail the integration test. They import the find_full_path and find_root_directory functions from incorrect places.

Instead of get_scan_image_files and get_scan_box_files in element_calcium_imaging. Similar to how it is implemented in element_array_ephys.

There are two requirements files: requirements.txt and requirements_test.txt. The integration test should be able to work with default requirements packages.

• Mirror the datajoint/pharus repository, once it is updated based off of today's discussion.

  /.github/ISSUE_TEMPLATE/bug_report.md
  /.github/ISSUE_TEMPLATE/feature_request.md
  /.github/ISSUE_TEMPLATE/config.yml
  /CONTRIBUTING.md
  /CODE_OF_CONDUCT.md
  

• Fetch and display summary images
• Fetch and display traces
• Example notebook

• FISSA

• Upload test data to DJArchive

tests/__init__.test_data points djarchive to a nonexistent dataset workflow-calcium-ephys-test-set

The README.md file refers to a notebooks/run_workflow.ipynb file but this file does not seem to exist.

The current pipeline imports element_lab.Location and then uses this as an upstream table for scan.ScanLocation which is described as follows:

Location: Reference table for ScanLocation, specifying the brain location where this scan is acquired

This does not match the description of element_lab.Location:

location of animal housing or experimental rigs

I propose modifying scan.ScanLocation to the following:

@schema
class ScanLocation(dj.Manual):
    definition = """
    -> Scan
    ---
    -> AnatomicalLocation
    """

And adding the following to workflow_calcium_imaging.pipeline.py or element_calcium_imaging.scan, changing the schema decorator:

@lab.schema
class AnatomicalLocation(dj.Manual):
    definition = """
    scan_location_id : varchar(16)
    ----
    anatomical_description: varchar(256)
    """

For users who want to try the workflow but don't have experience with databases.

This will work on a light djarchive data.
Add trigger (test_mode) to test file. Since the default ProcessingTask was load, triggering wasn't tested in the original pytest.

element_data_loader is required both in the requirements.txt file and the documentation. Otherwise trigger won't work.

Also update workflow-calcium-imaging to add support for multiple root data directories (e.g. paths.py).