"Process" files interaction

[SeanLeRoy]

This might be too unrelated, but in a lot of beta user interviews they ask about where and how we do some of our images modifications alluding to tools like bioio and segmentation algs. I’ve been wondering if it would be powerful to provide a UI in BFF that would provide users the exact Python script to run to use on their files or perhaps create a backend that just runs the script for them and streams back the image data.

An example workflow of a user wanting to convert to .zarr would be:

• Upload .csv containing metadata/references to files
• Select files to convert
• Right click
• Select a new menu option: “Process” (might not be the best word here)
• Select which type of file processing, in this case: “File type conversion”
• A modal appears with some prompts about the files to convert (like optional options about physical pixel size for example)
• A Python script appears that is automatically filtered to just the selected files and will perform the .zarr conversion as selected

[SeanLeRoy]

From Graham:

Its sort of providing a practical answer to how I was begging our institute to ponder DataFlow years ago... where from my high-level perspective I could luxuriously ignore details of "how" and allow myself to make blanket statements like "at a high level everyone in biology and therefor at AICS is doing the exact same thing = gathering data files; processing those files in various ways; storing the results in a way (usually new files) that they can look at/ or continue processing/ or share/ or download/ or permanently store/ or publish/ or whatever them" ... so if we had some interface that treats that 3 step process at a high-level like that (with that middle step being black-boxy for now), I think it would have universal appeal. Seems like provenance+ could be a helpful addition to connect any new files created- similar to what you and Chris were proposing on the fly at the Altos meeting yesterday if I recall.

From Nathalie:

People want the tools where the data is stored so they don’t need to move their files. This is the biggest incentive to use an image data repository. Through our interviews with experts from the scientific community (major stakeholders), we hear that people will use cloud repos to access tools for visualization, processing, and analysis more so than for sharing their data.