PKHeX/PKHeX.Core/Legality/LearnSource/Group/LearnGroup1.cs

254 lines
9.1 KiB
C#
Raw Normal View History

Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
using System;
namespace PKHeX.Core;
/// <summary>
/// Group that checks the source of a move in <see cref="GameVersion.Gen1"/>.
/// </summary>
public sealed class LearnGroup1 : ILearnGroup
{
public static readonly LearnGroup1 Instance = new();
private const int Generation = 1;
public ILearnGroup? GetPrevious(PKM pk, EvolutionHistory history, IEncounterTemplate enc, LearnOption option) => pk.Context switch
{
EntityContext.Gen1 when enc.Generation == 1 && pk is PK1 pk1 && HasDefinitelyVisitedGen2(pk1) => LearnGroup2.Instance,
EntityContext.Gen1 when enc.Generation == 2 => LearnGroup2.Instance,
EntityContext.Gen2 => null,
_ => enc.Generation == 1 ? LearnGroup2.Instance : null,
};
public bool HasVisited(PKM pk, EvolutionHistory history) => history.HasVisitedGen1;
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
public bool Check(Span<MoveResult> result, ReadOnlySpan<ushort> current, PKM pk, EvolutionHistory history, IEncounterTemplate enc,
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
MoveSourceType types = MoveSourceType.All, LearnOption option = LearnOption.Current)
{
if (enc.Generation == Generation && types.HasFlagFast(MoveSourceType.Encounter))
CheckEncounterMoves(result, current, enc, pk);
var evos = history.Gen1;
for (var i = 0; i < evos.Length; i++)
Check(result, current, pk, evos[i], i, option, types);
if (GetPrevious(pk, history, enc, LearnOption.Current) is null)
FlagEvolutionSlots(result, current, history);
return MoveResult.AllParsed(result);
}
private static void FlagEvolutionSlots(Span<MoveResult> result, ReadOnlySpan<ushort> current, EvolutionHistory history)
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
{
for (int i = 0; i < result.Length; i++)
{
if (current[i] == 0)
continue;
var detail = result[i];
if (!detail.Valid || detail.Generation is not (1 or 2))
continue;
var info = detail.Info;
if (info.Method is LearnMethod.LevelUp)
{
var level = info.Argument;
var stage = detail.EvoStage;
var chain = detail.Generation is 1 ? history.Gen1 : history.Gen2;
var species = chain[stage].Species;
if (IsAnyOtherResultALowerEvolutionStageAndHigherLevel(result, i, history, level, species))
result[i] = MoveResult.Unobtainable();
}
}
}
private static bool IsAnyOtherResultALowerEvolutionStageAndHigherLevel(Span<MoveResult> result, int index, EvolutionHistory history, byte level, ushort species)
{
// Check if any other result is a lower evolution stage and higher level.
for (int i = 0; i < result.Length; i++)
{
if (i == index)
continue;
var detail = result[i];
if (!detail.Valid || detail.Generation is not (1 or 2))
continue;
(var method, _, byte level2) = detail.Info;
if (method is not LearnMethod.LevelUp)
continue;
var stage = detail.EvoStage;
var chain = detail.Generation is 1 ? history.Gen1 : history.Gen2;
var species2 = chain[stage].Species;
if (level2 > level && species2 <= species)
return true;
}
return false;
}
private static void FlagFishyMoveSlots(Span<MoveResult> result, ReadOnlySpan<ushort> current, IEncounterTemplate enc)
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
{
var occupied = current.Length - current.Count((ushort)0);
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
if (occupied == 4)
return;
Span<ushort> moves = stackalloc ushort[4];
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
GetEncounterMoves(enc, moves);
// Count the amount of initial moves not present in the current list.
int count = CountMissing(current, moves);
if (count == 0)
return;
// There are ways to skip level up moves by leveling up more than once.
// Evolving at targeted levels can evade learning moves too.
// https://bulbapedia.bulbagarden.net/wiki/List_of_glitches_(Generation_I)#Level-up_learnset_skipping
// Just flag missing initial move slots.
int ctr = count;
for (int i = 0; i < result.Length; i++)
{
if (current[i] != 0)
continue;
result[i] = MoveResult.EmptyInvalid;
if (--ctr == 0)
break;
}
}
private static int CountMissing(ReadOnlySpan<ushort> current, ReadOnlySpan<ushort> moves)
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
{
int count = 0;
foreach (var expect in moves)
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
{
if (expect == 0)
break;
if (!current.Contains(expect))
count++;
}
return count;
}
private static void CheckEncounterMoves(Span<MoveResult> result, ReadOnlySpan<ushort> current, IEncounterTemplate enc, PKM pk)
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
{
Span<ushort> moves = stackalloc ushort[4];
if (enc is IMoveset {Moves: {HasMoves: true} x})
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
x.CopyTo(moves);
else
GetEncounterMoves(enc, moves);
LearnVerifierHistory.MarkInitialMoves(result, current, moves);
// Flag empty slots if never visited Gen2 move deleter.
if (pk is not PK1 pk1)
return;
if (HasDefinitelyVisitedGen2(pk1))
return;
FlagFishyMoveSlots(result, current, enc);
}
private static bool HasDefinitelyVisitedGen2(PK1 pk1)
{
if (!ParseSettings.AllowGen1Tradeback)
return false;
var rate = pk1.Catch_Rate;
return rate is 0 || GBRestrictions.IsTradebackCatchRate(rate);
}
private static void GetEncounterMoves(IEncounterTemplate enc, Span<ushort> moves)
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
{
if (enc.Version is GameVersion.YW or GameVersion.RBY)
LearnSource1YW.Instance.GetEncounterMoves(enc, moves);
else
LearnSource1RB.Instance.GetEncounterMoves(enc, moves);
}
private static void Check(Span<MoveResult> result, ReadOnlySpan<ushort> current, PKM pk, EvoCriteria evo, int stage, LearnOption option = LearnOption.Current, MoveSourceType types = MoveSourceType.All)
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
{
var rb = LearnSource1RB.Instance;
if (!rb.TryGetPersonal(evo.Species, evo.Form, out var rp))
return; // should never happen.
var yw = LearnSource1YW.Instance;
if (!yw.TryGetPersonal(evo.Species, evo.Form, out var yp))
return; // should never happen.
if (ParseSettings.AllowGen1Tradeback && ParseSettings.AllowGen2MoveReminder(pk))
evo = evo with { LevelMin = 1 };
for (int i = result.Length - 1; i >= 0; i--)
{
ref var entry = ref result[i];
if (entry.Valid && entry.Generation > 2)
continue;
var move = current[i];
var chk = yw.GetCanLearn(pk, yp, evo, move, types);
if (chk != default && GetIsPreferable(entry, chk, stage))
{
entry = new(chk, (byte)stage, Generation);
continue;
}
chk = rb.GetCanLearn(pk, rp, evo, move, types);
if (chk != default && GetIsPreferable(entry, chk, stage))
entry = new(chk, (byte)stage, Generation);
}
}
private static bool GetIsPreferable(in MoveResult entry, in MoveLearnInfo chk, int stage)
{
if (entry == default)
return true;
if (entry.Info.Method is LearnMethod.LevelUp)
{
if (chk.Method is not LearnMethod.LevelUp)
return true;
if (entry.EvoStage == stage)
return entry.Info.Argument < chk.Argument;
}
else if (entry.Info.Method.IsEggSource())
{
return true;
}
else if (chk.Method is LearnMethod.LevelUp)
{
return false;
}
return entry.EvoStage < stage;
}
public void GetAllMoves(Span<bool> result, PKM pk, EvolutionHistory history, IEncounterTemplate enc, MoveSourceType types = MoveSourceType.All, LearnOption option = LearnOption.Current)
{
if (types.HasFlagFast(MoveSourceType.Encounter) && enc.Generation == Generation)
FlagEncounterMoves(enc, result);
foreach (var evo in history.Gen1)
GetAllMoves(result, pk, evo, types);
}
private static void GetAllMoves(Span<bool> result, PKM pk, EvoCriteria evo, MoveSourceType types)
{
if (ParseSettings.AllowGen1Tradeback && ParseSettings.AllowGen2MoveReminder(pk))
evo = evo with { LevelMin = 1 };
LearnSource1YW.Instance.GetAllMoves(result, pk, evo, types);
LearnSource1RB.Instance.GetAllMoves(result, pk, evo, types);
}
private static void FlagEncounterMoves(IEncounterTemplate enc, Span<bool> result)
{
if (enc is IMoveset { Moves: { HasMoves: true } x })
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
{
result[x.Move4] = true;
result[x.Move3] = true;
result[x.Move2] = true;
result[x.Move1] = true;
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
}
else
{
Span<ushort> moves = stackalloc ushort[4];
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
GetEncounterMoves(enc, moves);
foreach (var move in moves)
result[move] = true;
}
}
}