Skip to content

mattiasnordqvist/Golden-Fox

Repository files navigation

Golden-Fox

Foxy scheduling

I have always found scheduling messy. I mean, how do you even store this in a database? Imagine you could just say do this and that on the "2nd last day every month at 22:30"? Well, no need to imagine. Now you can!

Yes, we can!

Schedule.Fox("every day @ 10:00").From(DateTime.Now).First()  // => Would give you a datetime representing today or tomorrow at 10:00, depending on what the time is now.
Schedule.Fox("every day @ 10:00").From(DateTime.Now).Take(10)  // => An IEnumerable<DateTime> with 10 entries starting from today or tomorrow, depending on what the time is now.

With a little creativity you could design your own extension method so that this syntax is available for you

"sundays at 22:30".From("2015-08-14").First() // => 2015-08-16 22:30:00

See how simple this schedule is to store in a database? You need ONE column of one of the simplest types. No modeling required. You can now even store it in a configuration file wihtout hassle. The format is also easily recognized and understood by other programmers or users.

A typical situation when Golden Fox comes to use is when you are scheduling services to run at specific times. Given a schedule and a point in time, you can let Golden Fox find a future point in time according to the schedule.

Another is when you want to generate events at these intervals. It is easy to create an observable from a Golden Fox schedule.

var enumerator = Schedule.Fox("every second").From(DateTime.Now).GetEnumerator();
enumerator.MoveNext();

var observable = Observable.Generate(
    enumerator.Current,
    i => true,
    i => 
    {
        enumerator.MoveNext();
        return enumerator.Current;
    },
    i => i,
    i => new TimeSpan(0, 0, 0, (int)Math.Max(0, (i - DateTime.Now).TotalSeconds)));

observable.Subscribe(x => Console.WriteLine(x));

Does Golden Fox understand English?

This fox ain't golden for no reason. As long as the words you talk fit into this grammar, you're fine.

The Api

Schedule.Fox("every day @ 10:00").From(DateTime.Today);

or

new Fox("every day @ 10:00", DateTime.Today);

Returns an IEnumerable<DateTime starting with the next occurence from the given date (inclusive)

If you scratch the surface of the golden fox, you'll soon see that his fur is just a disguise. The bare bones of this creature is just a compiler which you can access directly like this

Fox.Compile("every day @ 10:00").Evaluate(DateTime.Today)

which will return just the first occurence from the given date. Not inclusively this time though. Consider

Fox.Compile("every day @ 10:00").Evaluate(DateTime.Today, true)

if you want inclusiveness. However, in this case it wouldn't matter, because DateTime.Today evaluates to Today at 00:00 which makes 10:00 the next occurence no matter if you include 00:00 or not as a viable option.

Intervals

Intervals are expressions like every minute or every day @ 10:00. Intervals can be combined by placing an and between them. Like ever minute and every day @ 10:00. However, combining these two wouldn't make any sense, because every day @ 10:00 is also a minute, but you get the point. An interval must specify a point in time that is recurring. Available intervals are

every hour

This means at 00:00, 01:00, 02:00, 03:00 and so forth. You can offset the minutes and seconds by using any of these syntaxes. every hour @ 15 minutes every hour @ hh:15. Offsets can also be combined with and, like every hour @ hh:15 and hh:45.

every minute

Like every hour, but for minutes. Offsets can be expressed like every minute @ 15 seconds, every minute @ mm:15 or every minute @ hh:mm:15

every second

Like every hour or every minute, except you can't offset it. Golden Fox don't care about milliseconds.

every day

every day can not stand alone. You must supply at least one time, like every day at 10:00. Multiple times can be used. every day at 10:00 and 18:00. A time can be expressed like hh:mm:ss or hh:mm. 13:45 equals 13:45:00.

every weekday

every thursday at 15:00 and thursdays at 15:00 means the same thing.

every day in week

3rd day every week @ 12:30 or 4th last day every week @ 12:30 means the same thing. Any nonsensical input is not guarded against. No one knows what would happen if you say 14th day every week @ 12:30.

every day in month

Works the same way as every day in week, except it works with months instead. last day every month @ 12:30 or 15th day every month @ 12:30 are both valid input. 31st day every month @ 12:30 is not valid input.

every day in year

Like the above but for years. last day every year at 12:00.

Constraints

On all of the intervals above, you can add constraints.

Between

Can be added to every hour, every minute and every second. It would look like this: every hour between 14:00 and 18:00. Constraints are inclusive so generating a sequence of dates from this interval would have 14:00 and 18:00 included.

From

Can be added to any interval. Defines a starting point for the interval. every day @ 10:00 from 2016-01-01 with input 2010-01-01 would give 2016-01-01 10:00:00. From can be a date, but you can also be more specific and supply a time, like from 2016-01-01 15:30

Until

This constraint is a little special. You use it just like you use From. every day until 2020-01-01. So what will happen the day we pass in a date higher than 2020-01-01? Well. There´s no sensible next occurence really, so we will throw an InvalidOperationException. Would we get an exception if we passed in 2010-01-01? Remeber, constraints are inclusive, so no, we wouldn't.

Fluent Api

In version 2.2.0, I added a fluent version of the api. It is available as a separate nuget package. You can explore it yourself by start typing any of these lines:

using GoldenFox.Fluent

Every. // use intellisense from here
1.St(). // use intellisense from here
First(). // use intellisense from here
2.Nd(). // use intellisense from here
Last(). // use intellisense from here

Contribute

Besides being used in production code in some of my projects, this is my pet project for learning how to create DSLs. All ideas on new features and how to improve the code are welcome.