Plans for the near future mainly include, besides usual improvements and bug fixes, a UI redesign as well as additional types of charts and statistics (see [#101](https://github.com/muety/wakapi/issues/101), [#76](https://github.com/muety/wakapi/issues/76), [#12](https://github.com/muety/wakapi/issues/12)). If you have feature requests or any kind of improvement proposals feel free to open an issue or share them in our [user survey](https://github.com/muety/wakapi/issues/82).
There are different options for how to use Wakapi, ranging from our hosted cloud service to self-hosting it. Regardless of which option choose, you will always have to do the [client setup](#-client-setup) in addition.
**Note:** By default, SQLite is used as a database. To run Wakapi in Docker with MySQL or Postgres, see [Dockerfile](https://github.com/muety/wakapi/blob/master/Dockerfile) and [config.default.yml](https://github.com/muety/wakapi/blob/master/config.default.yml) for further options.
If you want to run Wakapi on **Kubernetes**, there is [wakapi-helm-chart](https://github.com/andreymaznyak/wakapi-helm-chart) for quick and easy deployment.
Wakapi relies on the open-source [WakaTime](https://github.com/wakatime/wakatime) client tools. In order to collect statistics for Wakapi, you need to set them up.
You can specify configuration options either via a config file (default: `config.yml`, customizable through the `-c` argument) or via environment variables. Here is an overview of all options.
| `app.aggregation_time` /<br>`WAKAPI_AGGREGATION_TIME` | `02:15` | Time of day at which to periodically run summary generation for all users |
| `app.report_time_weekly` /<br>`WAKAPI_REPORT_TIME_WEEKLY` | `fri,18:00` | Week day and time at which to send e-mail reports |
| `app.import_batch_size` /<br>`WAKAPI_IMPORT_BATCH_SIZE` | `50` | Size of batches of heartbeats to insert to the database during importing from external services |
| `app.inactive_days` /<br>`WAKAPI_INACTIVE_DAYS` | `7` | Number of days after which to consider a user inactive (only for metrics) |
| `app.heartbeat_max_age /`<br>`WAKAPI_HEARTBEAT_MAX_AGE` | `4320h` | Maximum acceptable age of a heartbeat (see [`ParseDuration`](https://pkg.go.dev/time#ParseDuration)) |
| `app.custom_languages` | - | Map from file endings to language names |
| `app.avatar_url_template` | (see [`config.default.yml`](config.default.yml)) | URL template for external user avatar images (e.g. from [Dicebear](https://dicebear.com) or [Gravatar](https://gravatar.com)) |
| `server.port` /<br>`WAKAPI_PORT` | `3000` | Port to listen on |
| `server.listen_ipv4` /<br>`WAKAPI_LISTEN_IPV4` | `127.0.0.1` | IPv4 network address to listen on (leave blank to disable IPv4) |
| `server.listen_ipv6` /<br>`WAKAPI_LISTEN_IPV6` | `::1` | IPv6 network address to listen on (leave blank to disable IPv6) |
| `server.listen_socket` /<br>`WAKAPI_LISTEN_SOCKET` | - | UNIX socket to listen on (leave blank to disable UNIX socket) |
| `security.password_salt` /<br>`WAKAPI_PASSWORD_SALT` | - | Pepper to use for password hashing |
| `security.insecure_cookies` /<br>`WAKAPI_INSECURE_COOKIES` | `false` | Whether or not to allow cookies over HTTP |
| `security.cookie_max_age` /<br>`WAKAPI_COOKIE_MAX_AGE` | `172800` | Lifetime of authentication cookies in seconds or `0` to use [Session](https://developer.mozilla.org/en-US/docs/Web/HTTP/Cookies#Define_the_lifetime_of_a_cookie) cookies |
| `security.allow_signup` /<br>`WAKAPI_ALLOW_SIGNUP` | `true` | Whether to enable user registration |
| `security.expose_metrics` /<br>`WAKAPI_EXPOSE_METRICS` | `false` | Whether to expose Prometheus metrics under `/api/metrics` |
| `sentry.dsn` /<br>`WAKAPI_SENTRY_DSN` | – | DSN for to integrate [Sentry](https://sentry.io) for error logging and tracing (leave empty to disable) |
There is also a [nice Grafana dashboard](https://grafana.com/grafana/dashboards/12790), provided by the author of [wakatime_exporter](https://github.com/MacroPower/wakatime_exporter).
Wakapi plays well together with [WakaTime](https://wakatime.com). For one thing, you can **forward heartbeats** from Wakapi to WakaTime to effectively use both services simultaneously. In addition, there is the option to **import historic data** from WakaTime for consistency between both services. Both features can be enabled in the _Integrations_ section of your Wakapi instance's settings page.
Wakapi also integrates with [GitHub Readme Stats](https://github.com/anuraghazra/github-readme-stats#wakatime-week-stats) to generate fancy cards for you. Here is an example.
There is a [WakaTime plugin](https://github.com/lowlighter/metrics/tree/master/source/plugins/wakatime) for GitHub [Metrics](https://github.com/lowlighter/metrics/) that is also compatible with Wakapi.
It is recommended to use wakapi behind a **reverse proxy**, like [Caddy](https://caddyserver.com) or [nginx](https://www.nginx.com/), to enable **TLS encryption** (HTTPS).
However, if you want to expose your wakapi instance to the public anyway, you need to set `server.listen_ipv4` to `0.0.0.0` in `config.yml`.
Unit tests are supposed to test business logic on a fine-grained level. They are implemented as part of the application, using Go's [testing](https://pkg.go.dev/testing?utm_source=godoc) package alongside [stretchr/testify](https://pkg.go.dev/github.com/stretchr/testify).
API tests are implemented as black box tests, which interact with a fully-fledged, standalone Wakapi through HTTP requests. They are supposed to check Wakapi's web stack and endpoints, including response codes, headers and data on a syntactical level, rather than checking the actual content that is returned.
Our API (or end-to-end, in some way) tests are implemented as a [Postman](https://www.postman.com/) collection and can be run either from inside Postman, or using [newman](https://www.npmjs.com/package/newman) as a command-line runner.
To get a predictable environment, tests are run against a fresh and clean Wakapi instance with a SQLite database that is populated with nothing but some seed data (see [data.sql](testing/data.sql)). It is usually recommended for software tests to be [safe](https://www.restapitutorial.com/lessons/idempotency.html), stateless and without side effects. In contrary to that paradigm, our API tests strictly require a fixed execution order (which Postman assures) and their assertions may rely on specific previous tests having succeeded.
To keep things minimal, all JS and CSS assets are included as static files and checked in to Git. [TailwindCSS](https://tailwindcss.com/docs/installation#building-for-production) and [Iconify](https://iconify.design/docs/icon-bundles/) require an additional build step. To only require this at the time of development, the compiled assets are checked in to Git as well.
As explained in [#284](https://github.com/muety/wakapi/issues/284), precompressed (using Brotli) versions of some of the assets are delivered to save additional bandwidth. This was inspired by Caddy's [`precompressed`](https://caddyserver.com/docs/caddyfile/directives/file_server) directive. [`gzipped.FileServer`](https://github.com/muety/wakapi/blob/07a367ce0a97c7738ba8e255e9c72df273fd43a3/main.go#L249) checks for every static file's `.br` or `.gz` equivalents and, if present, delivers those instead of the actual file, alongside `Content-Encoding: br`. Currently, compressed assets are simply checked in to Git. Later we might want to have this be part of a new build step.
Since Wakapi heavily relies on the concepts provided by WakaTime, [their FAQs](https://wakatime.com/faq) largely apply to Wakapi as well. You might find answers there.
If you host Wakapi yourself, you have control over all your data. However, if you use our webservice and are concerned about privacy, you can also [exclude or obfuscate](https://wakatime.com/faq#exclude-paths) certain file- or project names.
Wakapi was started when I was a student, who wanted to track detailed statistics about my coding time. Although I'm a big fan of WakaTime I didn't want to pay <ahref="https://wakatime.com/pricing">$9 a month</a> back then. Luckily, most parts of WakaTime are open source!
Inferring a measure for your coding time from heartbeats works a bit differently than in WakaTime. While WakaTime has <ahref="https://wakatime.com/faq#timeout">timeout intervals</a>, Wakapi essentially just pads every heartbeat that occurs after a longer pause with 2 extra minutes.
It is unclear how to handle the three minutes in between. Did the developer do a 3-minute break, or were just no heartbeats being sent, e.g. because the developer was staring at the screen trying to find a solution, but not actually typing code?
This package is [Treeware](https://treeware.earth). If you use it in production, then we ask that you [**buy the world a tree**](https://plant.treeware.earth/muety/wakapi) to thank us for our work. By contributing to the Treeware forest you’ll be creating employment for local families and restoring wildlife habitats.
Coding in open source is my passion and I would love to do it on a full-time basis and make a living from it one day. So if you like this project, please consider supporting it 🙂. You can donate either through [buying me a coffee](https://buymeacoff.ee/n1try) or becoming a GitHub sponsor. Every little donation is highly appreciated and boosts my motivation to keep improving Wakapi!
I highly appreciate the efforts of **[@alanhamlett](https://github.com/alanhamlett)** and the WakaTime team and am thankful for their software being open source.
