In my [previous public update](https://peakd.com/hive-139531/@gadrian/added-one-more-brick-to-the-hive-engine-snapshot-tool) for this tool, the most important addition was support for diesel pools.

I guess the tool started to shape up quite well with today's [commits](https://github.com/gadrian78/he-tokens-snapshot/tree/main). By the way, I'm anxious to see VSC starting to get used, so I can build something for it too.

In the meantime, this is what's new for today's update:
* added persistent caching with a 15 minutes expiration time for Coingecko prices, market and diesel pool data, to avoid hitting the APIs too often (when the script is run for different accounts)
* added save snapshots as JSON files automatically after processing. Here's a nice image:
![image.png](https://files.peakd.com/file/peakd-hive/gadrian/23tGXhPaDXTHwis8eXyAy2eYmyNYERE3bf52xGP8D6hCCC9tzcihxW8GMggPK5szkLyoZ.png)
* created `setup.sh` which:
  * takes two variables for the project directory and the snapshots base directory (which must be updated to work!)
  * takes a list of usernames (which again, needs to be updated for your needs)
  * installs Python dependencies
  * creates and activates a service and a timer to run our tool for the list of accounts at 8am or later, when the computer is turned on and has access to the internet (`FIXED:` I just realized that this shell script runs all accounts with the same set of tokens defined in `config.py` for all accounts; I suggest you to use the `multiaccount-manual-run.sh` for snapshots of a different set of tokens for each account, until I fix this)
  * tests this setup
* created `multiaccount-manual-run.sh` where someone can add all their accounts and associated lists of tokens and the script will run for each account (and create a snapshot json file for each account); persistent caching helps prevent asking for the same data from APIs multiple times. I added this shell script thinking about people who don't want the snapshots being taken automatically (maybe because they don't want to add a service and a timer on their system or because they want it to run it occasionally)
* refactored the code as it grew and I could improve its structure

The structure of snapshots is as follows. Under the `SNAPSHOTS_DIR` (defined in `config.py` or taken as an argument in command line), we have:
```
[SMAPSHOTS_DIR]/[USERNAME]/[SNAPSHOT_TYPE]/timestamped-filename.json
```
Something like this:
![image.png](https://files.peakd.com/file/peakd-hive/gadrian/23u5tS6u8hVEbj4qYpVNwVsagTE4jYvsk7as8GdWhBvUmzyGHMGMH6Uch55xgvYmDX6GV.png)

The exact filename depends on the snapshot type. For example, for a weekly snapshot, it would be something like `2025-W26.json`.

At this point, after I fix (FIXED!) the oversight about the automated script using the same list of tokens for every account, instead of distinct ones (which can easily be circumvented for the short term by running the `multiaccount-manual-run.sh`), this side of the tool is pretty much feature complete.

What could be done next would be to use the snapshot JSON files to generate some reports / charts, but that's another big project by itself.

What I can say is that, despite using AIs to help with this project, I learned (or remembered where I already knew somewhere in the back of my mind) a bunch through it.

### UPDATES:
* Fixed `setup.sh` to work with a list of accounts AND the associated list of tokens for each account.
  * the script also deactivates and removes old service and timer at the beginning, if they were previously set; helpful if you change the code and re-run `setup.sh`
  * I carved out the part about deactivation and removal of the the current service and timer and added it to the shell script `stop-automation.sh`, in case someone doesn't want automated snapshots taken anymore
* created an `uninstall.sh` script with the obvious functionality (read the updated `readme.md` on the repo for more info)

Any feedback?