Feature/upload

[egrace479]

Added upload option to Dashboard (csv or xls).

• Updated functions:
  • get_filenames: returns filenames and paths, also patches "Any" subspecies glitch (if user doesn't remove Any option before selecting specific subspecies and pressing "Display Images", it will just return images of the selected subspecies).
  • test_get_filenames and get_images updated to interact with updated get_filenames.
  • get_data: also takes boolean indicating if latitude & longitude values are in data to do lat-lon manipulations and sets locality column with lat-lon values if it's not already a column in the dataframe.
  • test_get_data: now checks two variations, one with lat & long, one without.
• Also removed additional taxa info from cat_list (in our test data it was just subspecies data repeated).
• Updated README to reflect upload requirements and added two datasets for testing/demo. Resolves Issue #23

[egrace479]

This is currently what it looks like before upload, then below the line gets replaced by the dashboard as it has been. Clicking on "Upload Data" again allows for a new upload.

Should I add the column requirements underneath as well? It currently will throw an error (missing column, wrong file type, etc.) and point the user to how-it-works:

[hlapp]

This is currently what it looks like before upload, then below the line gets replaced by the dashboard as it has been. [...] Should I add the column requirements underneath as well? It currently will throw an error (missing column, wrong file type, etc.) and point the user to how-it-works

It seems helpful to provide a link to "How it works" before someone runs into an error unintentionally because they didn't fully understand which columns they'd have to have.

[egrace479]

It seems this change set lumps together changes for multiple separate and not necessarily interrelated issues. Generally I'd suggest breaking changes up into meaningful and coherent pieces, this makes review and feedback much easier.

Overall this looks fine but I notice there's a fair amount of code duplication. In the conditional for whether geo-coordinates are available or not, it seems only a small part differs for the two cases, and the rest is duplicated. There are also strings or string prefixes (such as https://github.com/Imageomics/dashboard-prototype/raw/main/test_data/images/) that occur many times over, and could just be defined at the beginning and then reused.

For this PR, would it have been better to do the file retrieval generalization with a revised preloaded dataset and test as its own PR, then move the main div out separately and add the upload feature?

To your second point, I can adjust the geo-coordinate if statement to append the final div to hist_div dependent on the boolean. For the string duplication, are you referring to test_query? I can adjust that as well.

[egrace479]

Updated as per @hlapp's suggestions.

New landing page: Updated Error message page with embedded link that opens in new tab too:

[egrace479]

Is there a CSV I can test these changes with?

In test_data:

• HCGSD_base_filepath_noMap.csv: has no lat/lon columns
• HCGSD_base_filepath.csv: has base filepath in file_url (there's also one with the full filepath, including filename)
• Hoyal_Cuthill_GoldStandard_metadata_cleaned.csv: doesn't have file_url
• uploading any of the image files will trigger the CSV warning.

[johnbradley]

Thanks @egrace479. I was able to upload the the first two CSV files listed in the previous comment. Everything seemed to work fine for me.