Not retrieving pictures or correct type of picture for flood event tracking spreadsheet

[acgold]

Problem

• Updated flood event tracking code tries to find the closest picture to a flood event measurement, but sometimes does not return a picture link
• If it does return a picture link, the picture "download" link is pasted in the flood event spreadsheet, rather than the "web view" link

Solution

• Investigate if measurements that do not have a picture link actually have pics that exist for that time period
• Test out a larger time window for picture retrieval, but make sure that the time interval doesn't return multiple pictures
• Investigate the python package used to see if we can get the web view link rather than download link

[acgold]

Anyone want to give this a shot? Maybe @lgrimley or @ththelen

[ththelen]

@acgold I can take a stab at it. It will be good to brush up on my python skills. Can you point me towards where I can find the the flood event tracking code, and perhaps give a quick tutorial on how I can run/test my solutions if it won't be immediately obvious once I find the code

[lgrimley]

@acgold Is this the same problem that Jasmine identified with the timestamp on the image as different than the timestamp reported in the spreadsheet? or is it a completely different problem? @ththelen I can help if you need someone to bounce ideas off of

[acgold]

Thanks, @ththelen and @lgrimley! Within drift_correction.py, there is the following function for updating the spreadsheet https://github.com/sunny-day-flooding-project/sdfp-drift/blob/6c29cad7f7d8488dbd8d51c5618d1ff779402fa4/drift_correction.py#L367 which calls this function for fetching images https://github.com/sunny-day-flooding-project/sdfp-drift/blob/6c29cad7f7d8488dbd8d51c5618d1ff779402fa4/drift_correction.py#L481

Setting up local testing is kind of involved, so I will ping you in Teams with a self-contained test dataset. For future reference or if you want to give it a shot, the way I typically test things like this is to:

• log into open shift from command line on my computer
• port forward to the right container for the database
• set all of the environmental variables with a short script I have locally
• run all of the drift correction code by running most of the main function within drift_correction.py
• then run the function I'm interested in

Maybe in the next couple weeks a group of us can meet to go quickly go through/set up this testing workflow?

[acgold]

This is probably related to the issue of differing timestamps, although I dropped the ball on that bug and forgot to record that issue at the time. The timestamps of the images will very likely not be the exact time of the measurement, but it should be within a few minutes.

Let's open a new issue if we find large time mismatches because that would be a much larger bug

[ththelen]

@acgold I would be very interested to learn how to set up your testing workflow. My experience with open shift and port forwarding is minimal, so I think I would need some hand holding to get to the point where I could run that operation on my own. That said, I would be happy to learn if you have time to give some guidance.