This library provides helper API to write and run Advent of Code puzzles solving code for the .NET platform.
Here is the sample how one can write puzzle solver runner which will
- Automatically load input from the adventofcode.com site.
- Run solver on the parsed input.
- Automatically verify at the site the answer provided by solver.
- Run assert method to make test green or red.
[Fact(DisplayName = "Verifies that the puzzle answer is correct")]
public async Task Verifies_that_the_puzzle_answer_is_correct()
{
var httpClient = new HttpClient();
var sessionToken = "<your adventofcode.com session token>";
var runner = new SiteRunner(httpClient, new Uri("https://adventofcode.com"), sessionToken);
await runner
.Puzzle(2021, day: 1, level: 1)
.SolveUsing<Day01Level1Solver>()
.HandlingResultCorrectness(Assert.True)
.ParsingInputWith(int.Parse)
.Run();
}
Add NuGet package AdventOfCode.Runner to your project project:
dotnet add package AdventOfCode.Runner
Solver is the type implementing AdventOfCode.IAsyncSolver<TEntry, TResult>
interface.
public interface IAsyncSolver<in TEntry, TResult>
{
ValueTask<TResult> Solve(IAsyncEnumerable<TEntry> entries);
}
The type arguments are described in the next two sections.
TEntry
is the type of the input entry which is parsed from the single line of the puzzle text input provided by site at the https://adventofcode.com/<year>/day/<day>/input
page.
E.g. for the year 2021 day 1 puzzles the site provides the input similar to this:
176
184
188
142
...
It makes sense to parse each line into the int
value and provide solver with int
values instead of strings. So the TEntry
here will be int
.
TEntry
is not necessary has to be the simple type like int
. It can be any type that you can parse from the single line of the input and that is more appropriate for your solver code.
TResult
is the type of the answer your solver is calculated. In the vast majority of the cases it will be int
.
The only method declared by the IAsyncSolver<in TEntry, TResult>
interface is the Solve
method. It takes the typed input passed through the argument and calculates and returns answer.
The method is designed as asynchronous (returns ValueTask<TResult>
) to do not restrict you in methods of puzzle solving.
It also takes the IAsyncEnumerable<TEntry>
of input entries allowing you to deal with any kind of input. Either with one buffered in the memory or with one streamed from the any source.
Sometimes it is inconvenient to work with streamed input or with asynchronous result. In this case you can add the new base class implementing IAsyncSolver
and declaring more convenient Solve-like method to implement. There is an abstract class SolverWithArrayInput derived from the IAsyncSolver
which you can use as a base type for your solvers when you don't want deal with asynchronous input and result.
Here is the sample of the solver implementation for the fake puzzle which requires to calculate sum of the input numbers:
public class SumSolver : IAsyncSolver<int, int>
{
public async ValueTask<int> Solve(IAsyncEnumerable<int> entries)
{
return await entries.SumAsync();
}
}
The API is based on the concept of a solvers runner represented by AdventOfCode.Runner<TEntry, TResult>
type.
When created, the runner takes as input and encapsulates a number of services that are called as the runner is running. In this way, the runner can be considered as an orchestrator of the services passed to it.
The runner is started by the Run method:
public class Runner<TEntry, TResult>
{
...
public async ValueTask Run() {...}
...
}
Services are passed to the runner constructor:
public class Runner<TEntry, TResult>
{
...
public Runner(
IAsyncEnumerable<string> inputLines,
Func<string, int, TEntry> lineParser,
IAsyncSolver<TEntry, TResult> solver,
IEnumerable<IRawInputObserver> rawInputObservers,
IEnumerable<IInputObserver<TEntry>> inputObservers,
IEnumerable<IResultObserver<TResult>> resultObservers,
IEnumerable<Func<ValueTask>> afterRunHandlers
) {...}
public Runner(
IAsyncEnumerable<TEntry> inputEntries,
IAsyncSolver<TEntry, TResult> solver,
IEnumerable<IRawInputObserver> rawInputObservers,
IEnumerable<IInputObserver<TEntry>> inputObservers,
IEnumerable<IResultObserver<TResult>> resultObservers,
IEnumerable<Func<ValueTask>> afterRunHandlers
) {...}
...
}
The following describes the services that are passed to the runner when it is created, and that are invoked while it is running.
The puzzle input is represented in the runner as an async enumerable of typed entries
IAsyncEnumerable<TEntry> inputEntries
or as an async enumerable of raw lines
IAsyncEnumerable<string> inputLines
Only one type of the input (typed or raw) can be passed to the runner. When the raw input is passed, then the line parser is required to pass also so that the typed input could be created:
Func<string, int, TEntry> lineParser
- The first argument of the
lineParser
is raw input line to be parsed. - The second argument is zero-based index of the line in the input lines enumerable.
- Parser returns typed input entry.
IAsyncSolver<TEntry, TResult> solver
Runner calls the solver passing the typed input entries to it and gets the calculated result from it.
Input observers accept each input entry and handle (one of the task of the input observing is a logging).
There are raw input observers:
IEnumerable<IRawInputObserver> rawInputObservers
public interface IRawInputObserver
{
ValueTask Observe(string line);
}
or typed input observers:
IEnumerable<IInputObserver<TEntry>> inputObservers
public interface IInputObserver<in TEntry>
{
ValueTask Observe(string line, TEntry entry);
}
The observer will be called in an order they follow in the observers enumerable before the solver starts calculating. The raw input observers are called before each line parsing.
Result observer is called once the solver completes calculation. Observers will be called in an order they follow in the observers enumerable and will accept the result.
IEnumerable<IResultObserver<TResult>> resultObservers
public interface IResultObserver<in TResult>
{
ValueTask Observe(TResult result);
}
To run something at the end of the processing one can use after-run handlers:
IEnumerable<Func<ValueTask>> afterRunHandlers
Method Run
starts asynchronous process of the orchestrated services executing:
public class Runner<TEntry, TResult>
{
public async ValueTask Run() {...}
}
Providing the input services and observers one can create specific process of obtaining puzzle input and handling the calculated result.
Here is the lifecycle of the the Run
execution:
OPTION 1 OPTION 2
______________________________ ______________________________
------------------------------ ------------------------------
| Asynchronously enumerating | | Asynchronously enumerating |
| raw input | | typed input |
------------------------------ ------------------------------
| |
------------------------------- |
| Calling raw input observers | |
------------------------------- |
| |
---------------------------- |
| Parsing each line during | |
| input enumerating | |
---------------------------- |
| |
-------------------------------- |
| Calling typed input observer | |
-------------------------------- |
|____________________________________|
|
-----------------------------------------
| Passing typed input enumerable to the |
| solver.Solve() |
-----------------------------------------
|
-----------------------------
| Solver calculating result |
-----------------------------
|
----------------------------
| Calling result observers |
----------------------------
|
------------------------------
| Calling after-run handlers |
------------------------------
To simplify solver runner creation and passing services to it there is a fluent API which guides you to build runner.
Look at the following sample:
static IAsyncEnumerable<string> LoadInputLinesAsync()
{
// loads input from somewhere...
}
...
private class Year2021Day1Level1Solver : SolverWithArrayInput<int, long>
{
protected override long Solve(int[] entries)
{
return entries.Sum();
}
}
...
// Build runner and start it
await AdventOfCode
.SolveUsing(new Year2021Day1Level1Solver())
.WithInput(LoadInputLinesAsync())
.ParsingInputWith(int.Parse)
.ObservingResultWith(result => Console.WriteLine($"Result: {result}"))
.Run();
The fluent API of the builder which is used here is self-explained. Methods of the builder have overrides to be convenient in different scenarios.
Follow the auto-completion in your IDE to explore more building options.
Normally to check your solution you have to deal with adventofcode.com:
- first, to load the input to provide it to you solving code (most of the time solving code is copying into file and solving code reads from it),
- second, to submit your answer.
The library provides extensions of the runner and it's builder to do all these communications with the site automatically.
To use it, you create the instance of the SiteRunner
type and write the code similar to the following:
var siteRunner = new SiteRunner(...);
await siteRunner
.Puzzle(2021, 1, level: 1)
.SolveUsing<int, int>(typeof(Day1Level1Resolver))
.HandlingResultCorrectness((result, isResultCorrect) =>
Console.WriteLine($"Your result {result} is{(isResultCorrect ? "" : " not")} correct"))
.ParsingInputWith(int.Parse)
.Run();
As you can expect there are overrides and extensions over the SiteRunner
API.
To create the instance of the SiteRunner
type you need to call it's constructor providing the following objects to it:
- Instance of the
System.Net.Http.HttpClient
; - Site url (adventofcode.com);
- Your personal session token to authorized runner calls to the site.
var siteRunner = new SiteRunner(
new HttpClient(),
new Uri("https://adventofcode.com"),
"<your session token>"
);
Session token is used to authorize you on adventofcode.com site.
To get it:
- login to the adventofcode.com,
- find it in the cookie under the key
session
session=111222...999
Do not share your token!
One of the way to keep it secret is to use .net secret manager (see, Safe storage of app secrets in development in ASP.NET Core).
Your can configure and create http client in any way that works for you.
You can pre-configure your http client instance to set it's base address to the site url (https://adventofcode.com
) and to set cookie with session token. In this case you don't need to pass url and session token to the SiteRunner
constructor.
You can use single instance of the SiteRunner
over your application.
SiteRunner
API can be easily extended to meet your needs.
For example, have a look at the vladsm/adventofcode-2021-dotnet repository where the solution is verifying with the site runner and makes unit test green or red (PuzzlesTestsExtensions.cs).
Happy solving!