ActivityWatch
Option to not announcing page data when browser is not active
Currently, the active tab is shown in the timeline even if I'm not looking at it.
Browser app is open 24/7 for me, with many open tabs. The data occupying most space in the timeline is the tab where I left the browser last time before switching to other apps. This is pretty misleading.
aw-watcher-obsidian seems smarter about this. Its timeline shows data blocks with more relevance to when I was using it.
aw-watcher-web-edge can contain short gaps (under a minute) here and there, but those don't change the overall picture.
Browsers do not have (reliable) APIs to detect that they are not active.
The solution to your UX issue is to let users filter away AFK data in the timeline by using the data from the AFK watcher (not modify the web watcher).
AFK won't work. I don't wan't to see browser data when I'm working with other apps.
I'm not familiar with browser APIs. ChatGPT provided something like this for Chromium:
chrome.windows.onFocusChanged.addListener(function(windowId) {
if (windowId === chrome.windows.WINDOW_ID_NONE) {
console.log('All browser windows are unfocused or minimized.');
} else {
console.log('Window is focused.');
}
});
MDN: https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/API/windows/onFocusChanged also mentions windows.WINDOW_ID_NONE
Instead of AFK, could the plugin subscribe for aw-watcher-window bucket to know when the browser is active?
The easiest solution would probably be quitting the web browser when you aren't using it :)
The easiest solution would probably be quitting the web browser when you aren't using it :)
Non-negotiable. Irresponsible to even suggest this.
Thinking how I could've achieved something similar by different means.
If there were a different set of categorization rules to split aw-watcher-window track into multiple tracks, I could've used it to split everything from msedge.exe and other browsers into a separate track.
That might give overall more visually readable timeline than color alone.
The downside is not having page URLs. If there were a mechanism to enrich (hydrate) data blocks instead of making a separate bucket - that would've also been helpful.
I have more complaints about current categorization, but that's for a different repo.