development-log

Reservoir Development Log

development-log

fits-stream

2021-12-24 update

updated the figma for 100 fits nft project and collected a bunch of different metadata ideas for adding to the NFT.

• NFT CARD DESIGN
  • make the card background change depending on the season attribute
    • spring: green
    • summer: yellow
    • fall: orange
    • winter: blue
  • 
• metadata
  • 
  • includes season, labels (which I have a filtered list of relevant clothing ones which could be used to “guess” # of layers), activities, venue types, and the location

2021-11-01 update

changed the default view to just show favorites and allow for filtering. This was harder than it should’ve been because I had to dive into vanilla javascript for handling filtering based on a checkbox input (instead of using hugo’s syntax, which just affects the generated html) 2021-11-01

2021-10-08 update

• okay i exported my fits album photos successfully and now have a folder with the photos. I need to:
  • figure out how to add needed metadata easily (caption).
    • with osxphoto can actually grab the description via {desc} property of the template engine, which might be the best method for now…
    • or just manually add the description on website yourself? don’t like that this doesn’t support doing it in the moment.
    • could host a simple web server and have api upload description for a specific photo
  • upload photos regularly via script (local cron job vs. github action script which would have to use icloud rest API) and set the proper s3 url
    • https://s3-<REGION-NAME>.amazonaws.com/<BUCKET-NAME>/<FILE/KEY-NAME>
    • https://s3-us-west-2.amazonaws.com/personal-apple-photos/fits-stream/2021-10-07\ fit.jpeg
    • generate the JSON file from the list of photos and then upload all the photos, skipping the ones that are already there. need some way of keeping lightweight state? just a json file that has a list of the uploaded files?
• stop using dogsheep-photos and write my own script that uses osx-photos to do what i want like grab all the metadata and generate json file while also uploading to s3 and stores the state too
• need to finalize design too
  • inspiration:
    • stream.thesephist.com
    • https://www.sz-mag.com/news/2020/08/sacai-s-s21-mens-lookbook/
      • two column layout with expand image on click
      • 

2021-10-05 update

• started working on this with the use case to enable a fashion fits stream on my personal website
  • need all the photos available in a big JSON
  • either have script that downloads from S3 using signed key via github actions and then hosts it all locally in the website or just render the src from a public bucket?
    • probably the former to avoid any public access of my infra and avoid costs
• set up an s3 bucket personal-apple-photos to hold my photo data remotely.
  • Image data is so difficult to store and serve… Whole challenge in itself compared to text.. What other formats are like this and what can they enable?
• ran the script to upload but only got like 90 photos? presume issue is that they are not downloaded from icloud since i use the “optimize storage option.” The easy solution is to just turn that off for now or turn on the “store originals on mac” for now.. but want to support generically
  • How to even support this for someone when you don’t have actual access to their photos DB? This tension of local vs cloud keeps coming up
    • coda is the like the epitome of cloud-first
• trying to manually use osx-photos to download the missing data. It seems to use AppleKit to communicate with photos to try to force it to download, although it does mention this part is buggy, and I’m not sure the script is working because it’s taking forever.
  • ok this script i think is ded
• maybe swap to using https://pypi.org/project/pyicloud/ and modify the dogsheep-photos script to use that instead of using the local DB.
  • would need to figure out how to get all the metadata though. Although in this case, I guess I only care about the actual photos for the fits thing.