I tried to run the project locally.
I was facing issue to start the project and debugged that to start the project, main.py and requirements.txt files are required which are missing. Can somebody guide which is the correct branch to setup the project.

Many thanks in advance for the help.

Window titles, especially in browsers, tend to leak a lot of private data (e.g. Gmail's <title> exposing e-mails, all text editors exposing document names, etc.).

An easy-to-configure "kill switch" flag to discard any "window title" data from watchers would solve this issue and be arguably even more suitable if enabled by default for most users.

AW: 0.11.0.
OS: Windows 10 LTSC (version 1809) x64.
Browser: ungoogled-chromium 92.0.4515.107.

Today I've updated ungoogled-chromium from 89.0.4389.114-1 to 92.0.4515.107, that's theoretically built from the same source code by other developer + "Safe Browser" functionality isn't cut off.

I didn't expected that the settings will be lost 'cause I didn't change my browser, but they were lost.

Luckily I had exported settings but importing them doesn't show any activity or difference - there is even no activity in the Network tab of Chrome DevTools. Logs also doesn't contain any records about import.

Here's the attached and pretty-printed settings (some sensitive info has been cut off w/o loss in constistency)

As the title states, running make build on the source code works fine until it reaches

Processing /Users/user/Desktop/activitywatch/aw-server

Upon which it has been frozen for like 10 minutes.
I have tried rerunning the program yet nothing has changed.

We have the endpoint DELETE /0/buckets/{bucket_id}/events/{event_id} which can delete a single event, it would be useful with a DELETE /0/buckets/{bucket_id}/events/which simply just takes an array of event_ids.

I got this in the AW webui:

Error: Network Error. See dev console (F12) and/or server logs for more info.

When I ran aw-server or aw-qt I got:

[28038] Error loading Python lib '/usr/bin/libpython3.6m.so.1.0': dlopen: /usr/bin/libpython3.6m.so.1.0: cannot open shared object file: No such file or directory

It seems /opt/activitywatch/libpython3.6m.so.1.0 exists though

We should require explicit creation of buckets.

First I'd like to express again my thanks for such a wonderful and helpful project!

I've been using ActivityWatch for about two weeks now, and have also implemented a few very custom watchers for my own use cases. I'm currently running off of the latest tag, v0.8.0b7.

One thing I've noticed though that is a bit annoying to me is the Events and Query API behavior (and corresponding UI behavior) for long events that begin prior to the query period being cut off. For example, if I was AFK for a long amount of time until 30 minutes ago, and query for events in the last hour (or pull up the Timeline view in the UI for the last hour), I only see the "not-afk" event that started 30 minutes ago... and blank space leading up to it.

It seems to me like it would be really helpful if the events and query API endpoints had an option to return a "leading" event, if there's an event that started before the start of the query period but ended during the query period. I could also see this being represented in the timeline view with some sort of special icon or left border, indicating that the event began prior to the currently-viewed period.

Hi, firstly - I'm very impressed by the project and am taking notes from the documentation, code, design, project structure to benefit my own projects, props to the team! I already have some extension ideas and would love to contribute to the project.

A quick fix for the AUR install (would be applicable for any other linux packages) - the .desktop entry should be added to /usr/share/applications too, in order to appear in the desktop environment app search.

Bucket endpoint should return metadata about buckets.

• Storage and retrieval of bucket metadata need to be supported by aw_datastore

I need to use the aw-server in a Docker env.

Can't just use localhost for mongo db config.

How can I change the mongo connection config?

Hey, activitywatch server doesnt seem to have a consistently updated docker image on dockerhub. The most recent image on dockerhub (not sure if official) was last updated 2 years ago. I believe that by using CI/CD and automating the docker build process so that a docker image is built for a new release, this would allow the dockerhub images to be kept up to date with any new releases/commits.

Tab urls (collected through aw-watcher-web) do not show up in search results for Tools > Search or Tools > Query (only the app name and app title show up).
Possible cause:

1. The bucket for aw-watcher-web is not included in the search
2. Somewhere in aw-core the url is filtered out
3. Somewhere in aw-server the url is filtered out

I've tried this query with raw json output but the urls are missing. I thought if I include browser_events, the urls would show up.

afk_events = query_bucket(find_bucket("aw-watcher-afk_"));
window_events = query_bucket(find_bucket("aw-watcher-window_"));
browser_events = query_bucket(find_bucket("aw-watcher-web-firefox"));
window_events = filter_period_intersect(window_events, afk_events);
window_events = filter_period_intersect(window_events, browser_events);
merged_events = merge_events_by_keys(window_events, ["app", "title"]);
RETURN = sort_by_duration(merged_events);

Workaround: I can do the queries I need directly on peewee-sqlite.v2.db in the terminal but I thought if the search/query functionality is already present in the UI then I could use it:

awuser@box:/aw/data/activitywatch/aw-server$ sqlite3 peewee-sqlite.v2.db 'select timestamp, datastr from eventmodel;' | sed -e 's/|{/",/ ; s/^/{"time": "/' | grep -i "github" | tail -3
{"time": "2021-10-21 03:16:11.200000+00:00","url": "https://github.com/ActivityWatch/aw-server/issues/85", "title": "Search results for `Tools > Search` or `Tools > Query` don't include urls \u00b7 Issue #85 \u00b7 ActivityWatch/aw-server", "audible": false, "incognito": false, "tabCount": 16}
{"time": "2021-10-21 03:16:12.827000+00:00","url": "https://github.com/ActivityWatch/aw-core/", "title": "ActivityWatch/aw-core: Core library for ActivityWatch", "audible": false, "incognito": false, "tabCount": 16}
{"time": "2021-10-21 03:16:15.641000+00:00","url": "https://github.com/ActivityWatch/aw-server/issues/85", "title": "Search results for `Tools > Search` or `Tools > Query` don't include urls \u00b7 Issue #85 \u00b7 ActivityWatch/aw-server", "audible": false, "incognito": false, "tabCount": 15}

After running the latest beta version (v0.12.3b5) of aw-server, the dashboard is not accessible.

$ /Applications/ActivityWatch.app/Contents/MacOS/aw-server
2023-07-21 12:21:04 [INFO ]: Using storage method: peewee  (aw_server.main:33)
2023-07-21 12:21:04 [INFO ]: Starting up...  (aw_server.main:41)
2023-07-21 12:21:04 [INFO ]: Using database file: /Users/BelKed/Library/Application Support/activitywatch/aw-server/peewee-sqlite.v2.db  (aw_datastore.storages.peewee:150)
 * Serving Flask app 'aw-server'
 * Debug mode: off
2023-07-21 12:21:04 [INFO ]: WARNING: This is a development server. Do not use it in a production deployment. Use a production WSGI server instead.
 * Running on http://localhost:5600  (werkzeug:187)
2023-07-21 12:21:04 [INFO ]: Press CTRL+C to quit  (werkzeug:187)
2023-07-21 12:21:06 [INFO ]: 404 (127.0.0.1): GET / HTTP/1.1  (flask:25)

This is useful as a method for heartbeats. Such as if a user just became non-afk and now the duration of the non-afk period wants to be updated every x minutes until the user becomes afk again.

It's not the full heartbeats API we are planning for, but it's a lot easier do and is expected to work well enough for now.

Arbitrary websites can request localhost urls unless they have the appropriate headers, and this only protects against some request types. Specifically, POST is not blocked by a lack of an Access-Control-Allow-Origin header. ActivityWatch does not appear to have one, so the browser should block requests from webpages loaded from remote urls.

$ curl -vv -H "Origin: http://attacker.com" http://localhost:5600
*   Trying 127.0.0.1:5600...
* TCP_NODELAY set
* Connected to localhost (127.0.0.1) port 5600 (#0)
> GET / HTTP/1.1
> Host: localhost:5600
> User-Agent: curl/7.68.0
> Accept: */*
> Origin: http://attacker.com
> 
* Mark bundle as not supporting multiuse
* HTTP 1.0, assume close after body
< HTTP/1.0 200 OK
< Server: Werkzeug/2.1.2 Python/3.9.14
< Date: Mon, 26 Jun 2023 15:13:41 GMT
< Content-Disposition: inline; filename=index.html
< Content-Type: text/html; charset=utf-8
< Content-Length: 1683
< Last-Modified: Mon, 26 Sep 2022 19:14:27 GMT
< Cache-Control: no-cache
< Date: Mon, 26 Jun 2023 15:13:41 GMT
< Connection: close

https://about.gitlab.com/blog/2021/09/07/why-are-developers-vulnerable-to-driveby-attacks/

AW: 0.11.0.2
OS: Archlinux x64.
Browser: Firefox

In my ~/.cache/activitywatch/log/aw-server/aw-server_2021-xxxx.log with following log for aw-server:

2021-08-02 04:50:39 [INFO ]: Using storage method: peewee  (aw_server.main:32)
2021-08-02 04:50:39 [INFO ]: Starting up...  (aw_server.main:37)
2021-08-02 04:50:40 [INFO ]: Using database file: /home/mizho/.local/share/activitywatch/aw-server/peewee-sqlite.v2.db  (aw_datastore.storages.peewee:116)
2021-08-02 04:50:40 [INFO ]:  * Running on http://localhost:5600/ (Press CTRL+C to quit)  (werkzeug:113)
2021-08-02 04:50:41 [INFO ]: Creating bucket 'aw-watcher-window_ZANM'  (aw_datastore.datastore.Datastore:52)
2021-08-02 04:50:41 [INFO ]: Creating bucket 'aw-watcher-afk_ZANM'  (aw_datastore.datastore.Datastore:52)
2021-08-02 04:50:51 [INFO ]: Received heartbeat, but bucket was previously empty, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:274)
2021-08-02 04:50:51 [INFO ]: Received heartbeat, but bucket was previously empty, inserting as new event. (bucket: aw-watcher-afk_ZANM)  (aw_server.api:274)
2021-08-02 04:51:08 [INFO ]: 404 (127.0.0.1): �[33mGET /img/icons/apple-touch-icon-152x152.png HTTP/1.1�[0m  (flask:25)
2021-08-02 04:54:10 [INFO ]: Creating bucket 'aw-watcher-web-firefox'  (aw_datastore.datastore.Datastore:52)
2021-08-02 04:54:14 [INFO ]: Received heartbeat, but bucket was previously empty, inserting as new event. (bucket: aw-watcher-web-firefox)  (aw_server.api:274)
2021-08-02 04:56:22 [INFO ]: Creating bucket 'aw-watcher-vim_ZANM'  (aw_datastore.datastore.Datastore:52)
2021-08-02 04:56:22 [INFO ]: Received heartbeat, but bucket was previously empty, inserting as new event. (bucket: aw-watcher-vim_ZANM)  (aw_server.api:274)
2021-08-02 05:00:23 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 05:03:20 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 05:05:07 [INFO ]: Creating bucket 'aw-watcher-web-chrome'  (aw_datastore.datastore.Datastore:52)
2021-08-02 05:05:12 [INFO ]: Received heartbeat, but bucket was previously empty, inserting as new event. (bucket: aw-watcher-web-chrome)  (aw_server.api:274)
2021-08-02 05:05:33 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 05:05:34 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-vim_ZANM)  (aw_server.api:262)
2021-08-02 05:11:03 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 05:18:02 [INFO ]: Creating bucket 'aw-stopwatch'  (aw_datastore.datastore.Datastore:52)
2021-08-02 05:30:50 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 05:41:37 [INFO ]: 404 (127.0.0.1): �[33mGET /img/icons/android-chrome-192x192.png HTTP/1.1�[0m  (flask:25)
2021-08-02 05:46:57 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-vim_ZANM)  (aw_server.api:262)
2021-08-02 05:55:37 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-vim_ZANM)  (aw_server.api:262)
2021-08-02 06:01:14 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 06:10:34 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-vim_ZANM)  (aw_server.api:262)
2021-08-02 06:14:19 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 06:17:14 [INFO ]: 404 (127.0.0.1): �[33mGET /img/icons/apple-touch-icon-152x152.png HTTP/1.1�[0m  (flask:25)
2021-08-02 06:22:40 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 06:23:09 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-vim_ZANM)  (aw_server.api:262)
2021-08-02 06:33:59 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-window_ZANM)  (aw_server.api:262)
2021-08-02 06:34:40 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-vim_ZANM)  (aw_server.api:262)
2021-08-02 06:35:36 [INFO ]: Received heartbeat after pulse window, inserting as new event. (bucket: aw-watcher-vim_ZANM)  (aw_server.api:262)

I've errors in the localhost:5600 UI such as:
TypeError: t is undefined. See dev console (F12) and/or server logs for more info.

TypeError: t is undefined
    datasets TimelineBarChart.vue:99
    Lodash 10
    value TimelineBarChart.vue:99
    value TimelineBarChart.vue:45
    VueJS 12
vue.runtime.esm.js:1888:12

The only probably relevant thing that it may contribute to this is I have custom (very simple) categorize which I have to each time export the latest and then import again.

First, is the 404 error in aw-server log related here, and how can I fix the seemingly vu.js icon issue?
Second, is there a better way than export/import the categories (if relevant)

I saw the documentation for CORS here: https://docs.activitywatch.net/en/latest/security.html#cors-configuration

I am trying to create my own webinterface and I want to configure the server so that my webinterface can access the local API running on localhost:5600.

Would it be possible to add a configuration option to set allowed origins in the settings? I suppose it might be good to add this configuration option to aw-qt.

I am currently writing a few test cases for aw-watcher-terminal and therefore want to check if the events were successfully sent to the server.

In my implementation I need to start aw-server (--testing) and aw-watcher-terminal (--testing). After these are started the python test file then adds a unique id to the event, then fetches the latest 100 events and filters out all events with this id. (Then I can check if all properties of this event are correct)

Do you have a better way to do this?

Trying to run a query with the following line:

events = categorize(events, [[["AW GitHub"], {"type": "regex", "regex": "github.com/ActivityWatch;AW"}]]);

Gives:

2023-08-10 12:06:24 [WARNING] 400 Client Error: BAD REQUEST for url: http://127.0.0.1:5600/api/0/query/
2023-08-10 12:06:24 [WARNING] Error message received: {'type': 'QueryParseException', 'message': 'Invalid syntax for value to assign'}

This is probably due to the ; in one of the rules.

We need a model to filter out sensitive data by default.

For example if a window title contains: "[title] - Firefox (Private Browsing)" we should redact [title] to some magic string such as "REDACTED".

For some cases we might want to filter the window out entirely, giving 0 information about which window is running, better catch too much than too little.

It should be the goal that every user has a set of "clean" data. The filtering should also be able to be run on an existing database of data, so that cleaner data can be output. Preferably, the data should be so clean that there is little (or even no) reason not to share it (which would be great since easy access to a large dataset could make research in some areas a lot easier!).

The question left is where this processing step should take place. We want the filtering/redacting to happen before data is sent anywhere but it should also be able to be enforceable on a server (if the server owner doesn't trust the servers security, if in the cloud for example) and have clients notified of this so that they can do the filtering on their side, removing the need to send sensitive data at all. It might therefore be prudent to write a module in aw-core that implements this functionality since it should be useable from the server and all clients (which transmit sensitive data).

This feature should be on by default, we don't need anything advanced yet, first priority is to redact titles from Incognito/Private Browsing, that's a good step in the right direction.

This should have a far higher priority than Zero-Knowledge storage right now, because it's a lot easier and is more user friendly (In ZK storage: if you lose your keys you lose your data).

https://github.com/noirbizarre/flask-restplus

Hello, I've been using AW and I love it.
However, just today, I checked my I/O accumulated stats using iotop -ao
It seems like aw-server used about 1GB/5mins. That, for a consumer SSD, is way above the the average daily usage, which is probably about 20-40GB/day.

It could also be that I'm running many applications, with the browser plugin enabled which in turn is the cause for so much I/O usage. But, still, I think the team should try and do something about it...

One, we could increase the default hearbeat/data-send intervals, or we could store the data to a temporary place and update it on disk every 10mins or so. I think the 2nd could be provided as a configuration option, if not given already.

It's possible to do the 2nd one manually, but I still hope it's something to consider. Either for this, or for the upcoming aw-server-rust rewrite!

That is the data usage for about 2mins of monitoring( time manually measured ). For firefox, there's an config option which can hugely reduce its usage( which I'll go do after this ), but I'm not sure if anything similar exists for aw-server or its' plugins

Hi all - can't build aw-server.
I get many errors of this nature:

/////////////////////////////////////////////////////////////
cargo build --release --bin aw-server
Compiling plex v0.2.5
Compiling url v2.3.1
error[E0658]: use of unstable library feature 'proc_macro_diagnostic'
--> /home/ec2-user/.cargo/registry/src/github.com-1ecc6299db9ec823/plex-0.2.5/src/lexer.rs:176:26
|
176 | .error(format!("invalid regular expression: {}", e))
| ^^^^^
|
= note: see issue #54140 rust-lang/rust#54140 for more information
... [ + many similar errors]
/////////////////////////////////////////////////////////////
Debugging info:

(venv) [ec2-user@ip-172-31-62-148 activitywatch]$ ./checkversions.sh
git version 2.39.1
GNU Make 3.82
Built for x86_64-koji-linux-gnu
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later http://gnu.org/licenses/gpl.html
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Python 3.8.16
Poetry (version 1.3.2)
v16.19.1 (node)
8.19.3 (npm)
rustc 1.67.1 (d5a82bbd2 2023-02-07)
cargo 1.67.1 (8ecd4f20a 2023-01-10)
/////////////////////////////////////////////////////////////
(venv) [ec2-user@ip-172-31-62-148 activitywatch]$ uname -a
Linux ip-172-31-62-148.us-west-2.compute.internal 4.14.304-226.531.amzn2.x86_64 #1 SMP Wed Feb 1 21:34:38 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
/////////////////////////////////////////////////////////////

I believe I have all the right pre-requisites and versions and have followed your build instructions carefully.

Best regards,

Chris.

Would be a simpler way to extend functionality depending on incoming data. Don't integrate them with the server but just enable which data that the server receives should be broadcast.

For example:
A Python program could listen to the websocket and when a AFK or no-longer-AFK event is received then do a notify-send, informing the user that the afkwatcher now thinks that the user is AFK (or not). Would be handy to detect errors in logging without having to actively check logs.

Same issue as ActivityWatch/aw-qt#19, only this process is a Flask server, and run is inherited from Flask, so I can't put a if os.getppid() == 1 inside it.

Flask has ways to shut itself down, but typically by setting up a /shutdown page that sends the shutdown call when requested, here we don't want to rely on requests, more like setup a watchdog. Here gives similar options. There my questions is asked, and most of the answer is "app.run() is only meant for development", not too helpful.

Now, does the server really have to be restarted ? Should aw-qt test whether the server is already up, and maybe connect to it rather than start it, if it's already alive ?

I've now, on more than one occasion, gotten a strange freeze where aw-server does not respond to any requests.

The first time was a couple days ago, the second time I noticed this morning as the web UI didn't load any data. I restarted the server through aw-qt, still frozen (but it seemed to have processed some requests before freezing). Restarted again through the terminal, now it worked and started processing the 2 day watcher backlog.

I have no idea what is causing this, but it's a critical bug.

Might be nice, if the version would be visible in aw-webui it would also make it easier for users to know which version of aw they are running.

Depends on the JSON datastore issue here: ActivityWatch/aw-core#3

Absolutely no security is in place and it's not currently meant for interhost communication. If one really wants one can set up a port forwarding with SSH.

WakaTime has a settings file that enables setting a api_url and api_key. This can be used to effectively route all WakaTime requests directly to another entity (probably for debugging purposes). There is even an example config file called paranoid.cfg used in testing.

It would be nice to have a way to receive the heartbeats that WakaTime sends out so we can collect the data.

Now, to the matter on how that would be done in a nice way. Bundling it into the actwa-server core seems nasty to me, but creating plugins that are initialized with a reference to the flask app object and then set up the routes on it's own shouldn't be too involved. For prototyping it might be best to just bundle it into actwa-server directly (but keep it in it's own file so it will be easy to refactor into a seperate package and repo later).

The API is documented here: https://wakatime.com/developers/

Since we still don't have a way to collect data from all devices into a single location, has anyone thought about storing data into Prometheus or Influx and visualizing with Grafana?

IPv6 localhost [::1] server does not respond

Server will not respond on IPv6 localhost Windows 10

Should include a guide for how to set it up with Lets Encrypt.

We do NOT have the security to possibly expose a unsecured piece of software on computers. The sync/security issue is postponed but is critical for many future features (such as sync/networked datastore).

So it happened, I have on that data from the end of 2019 and I was sad to delete but it is possible to recover the data on it.

sqlite3 /home/[your-user]/.local/share/activitywatch/aw-server/peewee-sqlite.v2.db ".dump" > dump.sql
rm -f /home/[your-user]/.local/share/activitywatch/aw-server/peewee-sqlite.v2.db
sqlite3 -init dump.sql /home/[your-user]/.local/share/activitywatch/aw-server/peewee-sqlite.v2.db
.quit

With this commands I was able to recreate a database and now activity watch starts. In case of this issue you have aw-qt reporting an error, if you press restart server aw-qt crashes...
With the more details button you see the error on peewee about the malformed database, so you need to do those commands and restart aw-qt.

First off, it simply didn't work, so I got hold of the request URL from the aw-server log and found the response to be:

URL: http://127.0.0.1:5666/api/0/views/aw-webui_windowactivity_summary_erb-main2-arch%3Flimit=-1&start=2017-03-12T00%253A00%253A00%252B01%253A00&end=2017-03-13T00%253A00%253A00%252B01%253A00

{
    "msg": "There's no view with the name 'aw-webui_windowactivity_summary_erb-main2-arch?limit=-1&start=2017-03-12T00%3A00%3A00%2B01%3A00&end=2017-03-13T00%3A00%3A00%2B01%3A00'"
}

So, after handcrafting the URL into the following I get another error:

URL: http://127.0.0.1:5666/api/0/views/aw-webui_windowactivity_timeline_erb-main2-arch?limit=-1&start=2017-03-12T00:00:00%2B01:00&end=2017-03-13T00:00:00%2B01:00

{
    "msg": "Cannot query bucket that doesn't exist in transform: 'aw-watcher-window_erb-main2-arch'"
}

And indeed, aw-watcher-window does not exist since I'm running in testing mode, surely it should work in testing mode as well.

And also, "that doesn't exist in transform" confused me, I thought the bucket was not in the transform, which made no sense to me. Now I realize that what the message tried to say was that the bucket specified in the transform does not exist in the database, which makes sense.

Related to ActivityWatch/aw-webui#392.
Seems like the database gets locked when requesting another request before the current one is finished.

2023-01-24 22:28:31 [ERROR]: Exception on /api/0/buckets/aw-watcher-window_BelKeds-MBP/events [GET]  (aw-server:1449)
Traceback (most recent call last):
  File "peewee.py", line 3237, in execute_sql
sqlite3.OperationalError: database is locked

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "flask/app.py", line 1517, in full_dispatch_request
  File "flask/app.py", line 1503, in dispatch_request
  File "aw_server/rest.py", line 46, in decorator
  File "flask_restx/api.py", line 405, in wrapper
  File "flask/views.py", line 84, in view
  File "flask_restx/resource.py", line 46, in dispatch_request
  File "aw_server/rest.py", line 208, in get
  File "aw_server/api.py", line 43, in g
  File "aw_server/api.py", line 191, in get_events
  File "aw_datastore/datastore.py", line 101, in get
  File "aw_datastore/storages/peewee.py", line 304, in get_events
  File "peewee.py", line 1962, in inner
  File "peewee.py", line 2033, in execute
  File "peewee.py", line 2206, in _execute
  File "peewee.py", line 3250, in execute
  File "peewee.py", line 3244, in execute_sql
  File "peewee.py", line 3010, in __exit__
  File "peewee.py", line 192, in reraise
  File "peewee.py", line 3237, in execute_sql
peewee.OperationalError: database is locked

From my understanding of the code, and documentation, every time the localhost page reloads, the top activities in all categories are recalculated from the beginning. If you have a lot of events, that can lead to a timeout( ActivityWatch/activitywatch#217). I think performance could be improved by creating a new table in the database where you store the summarized events. The columns would be:

Key(URLs, domains, app_events, title_events), Value(i.e github.com, localhost:5600), Duration, and Day.

Thus every time the page reloads, it only summarizes the new events, then adds the new domains/urls/etc and updates the existing ones. Then you just retrieve the results. This has two advantages:

1. Decreases the time to calculate the summary for the day, as you perform a lot fewer calculations.
2. Makes analyzing statistics for a week, month, year a lot faster since the table will have a summary of the events for every day. Therefore, you just retrieve the right amount of days and perform analysis on them.

@ErikBjare , @johan-bjareholt, what do you guys think? Are there any current initiatives like this already in the works? If not, I can start working on it

Matrix.org has a tutorial to make a free small server for their service: https://matrix.org/docs/guides/free-small-matrix-server/

Is there a simple setup like this that could be used for ActivityWatch? If so, could someone write one such tutorial, or do you just need to run the python program?

Even if not, the pricing comparison they show could be useful to have somewhere:

http://docs.peewee-orm.com/en/latest/peewee/playhouse.html#flask-utils

http://werkzeug.pocoo.org/docs/0.12/contrib/profiler/

--profiling would do it.