PKHeX/PKHeX.Core/Editing/Showdown/ShowdownSet.cs

807 lines
25 KiB
C#
Raw Normal View History

using System;
using System.Collections.Generic;
using System.Text;
using static PKHeX.Core.Species;
namespace PKHeX.Core;
/// <summary>
/// Logic for exporting and importing <see cref="PKM"/> data in Pokémon Showdown's text format.
/// </summary>
public sealed class ShowdownSet : IBattleTemplate
{
private static readonly string[] StatNames = { "HP", "Atk", "Def", "Spe", "SpA", "SpD" };
private const string LineSplit = ": ";
private const string ItemSplit = " @ ";
private const string ParenJunk = "()[]";
private const int MAX_SPECIES = (int)MAX_COUNT - 1;
internal const string DefaultLanguage = GameLanguage.DefaultLanguage;
private static readonly GameStrings DefaultStrings = GameInfo.GetStrings(DefaultLanguage);
private static readonly ushort[] DashedSpecies =
{
(int)NidoranF, (int)NidoranM,
(int)HoOh,
(int)Jangmoo, (int)Hakamoo, (int)Kommoo,
(int)TingLu, (int)ChienPao, (int)WoChien, (int)ChiYu,
};
/// <inheritdoc/>
public ushort Species { get; private set; }
/// <inheritdoc/>
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 EntityContext Context { get; private set; } = RecentTrainerCache.Context;
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public string Nickname { get; private set; } = string.Empty;
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int Gender { get; private set; } = -1;
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int HeldItem { get; private set; }
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int Ability { get; private set; } = -1;
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int Level { get; private set; } = 100;
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public bool Shiny { get; private set; }
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int Friendship { get; private set; } = 255;
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int Nature { get; private set; } = -1;
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public string FormName { get; private set; } = string.Empty;
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public byte Form { get; private set; }
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int[] EVs { get; } = {00, 00, 00, 00, 00, 00};
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int[] IVs { get; } = {31, 31, 31, 31, 31, 31};
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public int HiddenPowerType { get; private set; } = -1;
public MoveType TeraType { get; private set; } = MoveType.Any;
2018-12-29 00:54:01 +00:00
/// <inheritdoc/>
public ushort[] Moves { get; } = {0, 0, 0, 0};
2018-07-14 16:55:22 +00:00
/// <inheritdoc/>
public bool CanGigantamax { get; private set; }
/// <inheritdoc/>
public byte DynamaxLevel { get; private set; } = 10;
/// <summary>
/// Any lines that failed to be parsed.
/// </summary>
public readonly List<string> InvalidLines = new(0);
private GameStrings Strings { get; set; } = DefaultStrings;
2018-07-14 16:55:22 +00:00
/// <summary>
/// Loads a new <see cref="ShowdownSet"/> from the input string.
/// </summary>
/// <param name="input">Single-line string which will be split before loading.</param>
public ShowdownSet(ReadOnlySpan<char> input) => LoadLines(input.EnumerateLines());
2018-07-14 16:55:22 +00:00
/// <summary>
/// Loads a new <see cref="ShowdownSet"/> from the input string.
/// </summary>
/// <param name="lines">Enumerable list of lines.</param>
public ShowdownSet(IEnumerable<string> lines) => LoadLines(lines);
2018-07-14 16:55:22 +00:00
private void LoadLines(SpanLineEnumerator lines)
{
ParseLines(lines);
SanitizeResult();
}
private void LoadLines(IEnumerable<string> lines)
{
ParseLines(lines);
SanitizeResult();
}
private void SanitizeResult()
{
FormName = ShowdownParsing.SetShowdownFormName(Species, FormName, Ability);
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
Form = ShowdownParsing.GetFormFromString(FormName, Strings, Species, Context);
// Handle edge case with fixed-gender forms.
if (Species is (int)Meowstic or (int)Indeedee or (int)Basculegion or (int)Oinkologne)
ReviseGenderedForms();
}
2021-12-05 01:56:56 +00:00
private void ReviseGenderedForms()
{
if (Gender == 1) // Recognized with (F)
2021-12-05 01:56:56 +00:00
{
FormName = "F";
Form = 1;
}
else
{
FormName = Form == 1 ? "F" : "M";
Gender = Form;
}
}
private const int MaxMoveCount = 4;
// Skip lines that are too short or too long.
// Longest line is ~74 (Gen2 EVs)
// Length permitted: 3-80
// The shortest Pokémon name in Japanese is "ニ" (Ni) which is the name for the Pokémon, Nidoran♂ (male Nidoran). It has only one letter.
// We will handle this 1-2 letter edge case only if the line is the first line of the set, in the rare chance we are importing for a non-English language?
private const int MinLength = 3;
private const int MaxLength = 80;
private static bool IsLengthOutOfRange(ReadOnlySpan<char> trim) => (uint)(trim.Length - MinLength) > MaxLength + MinLength;
private void ParseLines(SpanLineEnumerator lines)
{
int movectr = 0;
bool first = true;
foreach (var line in lines)
{
ReadOnlySpan<char> trim = line.Trim();
if (IsLengthOutOfRange(trim))
{
// Try for other languages just in case.
if (first && trim.Length != 0)
{
ParseFirstLine(trim);
first = false;
continue;
}
InvalidLines.Add(line.ToString());
continue;
}
if (first)
{
ParseFirstLine(trim);
first = false;
continue;
}
if (ParseLine(trim, ref movectr))
return; // End of moves, end of set data
}
}
private void ParseLines(IEnumerable<string> lines)
{
int movectr = 0;
bool first = true;
foreach (var line in lines)
{
ReadOnlySpan<char> trim = line.Trim();
if (IsLengthOutOfRange(trim))
{
// Try for other languages just in case.
if (first && trim.Length != 0)
{
ParseFirstLine(trim);
first = false;
continue;
}
InvalidLines.Add(line);
continue;
}
if (first)
{
ParseFirstLine(trim);
first = false;
continue;
}
if (ParseLine(trim, ref movectr))
return; // End of moves, end of set data
}
}
private bool ParseLine(ReadOnlySpan<char> line, ref int movectr)
{
if (line[0] is '-' or '')
{
var moveString = ParseLineMove(line);
int move = StringUtil.FindIndexIgnoreCase(Strings.movelist, moveString);
if (move < 0)
InvalidLines.Add($"Unknown Move: {moveString}");
else if (Array.IndexOf(Moves, (ushort)move) != -1)
InvalidLines.Add($"Duplicate Move: {moveString}");
else
Moves[movectr++] = (ushort)move;
return movectr == MaxMoveCount;
}
if (movectr != 0)
return true;
bool valid;
var split = line.IndexOf(LineSplit, StringComparison.Ordinal);
if (split == -1)
{
valid = ParseSingle(line); // Nature
}
else
{
var left = line[..split].Trim();
var right = line[(split + LineSplit.Length)..].Trim();
valid = ParseEntry(left, right);
}
if (!valid)
InvalidLines.Add(line.ToString());
return false;
}
private bool ParseSingle(ReadOnlySpan<char> identifier)
{
if (!identifier.EndsWith("Nature", StringComparison.OrdinalIgnoreCase))
return false;
var firstSpace = identifier.IndexOf(' ');
if (firstSpace == -1)
return false;
var naturestr = identifier[..firstSpace];
return (Nature = StringUtil.FindIndexIgnoreCase(Strings.natures, naturestr)) >= 0;
}
private bool ParseEntry(ReadOnlySpan<char> identifier, ReadOnlySpan<char> value) => identifier switch
{
"Ability" => (Ability = StringUtil.FindIndexIgnoreCase(Strings.abilitylist, value)) >= 0,
"Nature" => (Nature = StringUtil.FindIndexIgnoreCase(Strings.natures , value)) >= 0,
"Shiny" => Shiny = StringUtil.IsMatchIgnoreCase("Yes", value),
"Gigantamax" => CanGigantamax = StringUtil.IsMatchIgnoreCase("Yes", value),
"Friendship" => ParseFriendship(value),
"EVs" => ParseLineEVs(value),
"IVs" => ParseLineIVs(value),
"Level" => ParseLevel(value),
"Dynamax Level" => ParseDynamax(value),
"Tera Type" => ParseTeraType(value),
_ => false,
};
private bool ParseLevel(ReadOnlySpan<char> value)
{
if (!int.TryParse(value.Trim(), out var val))
return false;
if ((uint)val is 0 or > 100)
return false;
Level = val;
return true;
}
private bool ParseFriendship(ReadOnlySpan<char> value)
{
if (!int.TryParse(value.Trim(), out var val))
return false;
if ((uint)val > byte.MaxValue)
return false;
Friendship = val;
return true;
}
private bool ParseDynamax(ReadOnlySpan<char> value)
{
Context = EntityContext.Gen8;
var val = Util.ToInt32(value);
if ((uint)val > 10)
return false;
return (DynamaxLevel = (byte)val) is (>= 0 and <= 10);
}
private bool ParseTeraType(ReadOnlySpan<char> value)
{
Context = EntityContext.Gen9;
var types = Strings.types;
var val = StringUtil.FindIndexIgnoreCase(types, value);
if (val < 0)
return false;
TeraType = (MoveType)val;
return true;
}
/// <summary>
/// Gets the standard Text representation of the set details.
/// </summary>
public string Text => GetText();
2018-07-14 16:55:22 +00:00
/// <summary>
/// Gets the localized Text representation of the set details.
/// </summary>
/// <param name="lang">2 character language code</param>
public string LocalizedText(string lang = DefaultLanguage) => LocalizedText(GameLanguage.GetLanguageIndex(lang));
/// <summary>
/// Gets the localized Text representation of the set details.
/// </summary>
/// <param name="lang">Language ID</param>
private string LocalizedText(int lang)
{
var strings = GameInfo.GetStrings(lang);
return GetText(strings);
}
private string GetText(GameStrings? strings = null)
{
2022-08-30 22:00:45 +00:00
if (Species is 0 or > MAX_SPECIES)
return string.Empty;
if (strings != null)
Strings = strings;
var result = GetSetLines();
return string.Join(Environment.NewLine, result);
}
public List<string> GetSetLines()
{
var result = new List<string>();
// First Line: Name, Nickname, Gender, Item
var form = ShowdownParsing.GetShowdownFormName(Species, FormName);
result.Add(GetStringFirstLine(form));
// IVs
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 ivs = GetStringStats(IVs, Context.Generation() < 3 ? 15 : 31);
if (ivs.Length != 0)
result.Add($"IVs: {string.Join(" / ", ivs)}");
// EVs
var evs = GetStringStats(EVs, 0);
if (evs.Length != 0)
result.Add($"EVs: {string.Join(" / ", evs)}");
// Secondary Stats
if ((uint)Ability < Strings.Ability.Count)
result.Add($"Ability: {Strings.Ability[Ability]}");
if (TeraType != MoveType.Any)
result.Add($"Tera Type: {Strings.Types[(int)TeraType]}");
if (Level != 100)
result.Add($"Level: {Level}");
if (Shiny)
result.Add("Shiny: Yes");
if (DynamaxLevel != 10 && Context == EntityContext.Gen8)
result.Add($"Dynamax Level: {DynamaxLevel}");
if (CanGigantamax)
result.Add("Gigantamax: Yes");
if ((uint)Nature < Strings.Natures.Count)
result.Add($"{Strings.Natures[Nature]} Nature");
// Moves
result.AddRange(GetStringMoves());
return result;
}
private string GetStringFirstLine(string form)
{
string specForm = Strings.Species[Species];
if (form.Length != 0)
specForm += $"-{form.Replace("Mega ", "Mega-")}";
else if (Species == (int)NidoranM)
specForm = specForm.Replace("♂", "-M");
else if (Species == (int)NidoranF)
specForm = specForm.Replace("♀", "-F");
string result = GetSpeciesNickname(specForm);
// omit genderless or nonspecific
if (Gender is 1)
result += " (F)";
else if (Gender is 0)
result += " (M)";
if (HeldItem > 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
var items = Strings.GetItemStrings(Context);
if ((uint)HeldItem < items.Length)
result += $" @ {items[HeldItem]}";
}
return result;
}
private string GetSpeciesNickname(string specForm)
{
if (Nickname.Length == 0)
return specForm;
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
bool isNicknamed = SpeciesName.IsNicknamedAnyLanguage(Species, Nickname, Context.Generation());
if (!isNicknamed)
return specForm;
return $"{Nickname} ({specForm})";
}
private static string[] GetStringStats(ReadOnlySpan<int> stats, int ignoreValue)
{
var count = stats.Length - stats.Count(ignoreValue);
if (count == 0)
return Array.Empty<string>();
var result = new string[count];
int ctr = 0;
for (int i = 0; i < stats.Length; i++)
{
var statIndex = GetStatIndexStored(i);
var statValue = stats[statIndex];
if (statValue == ignoreValue)
continue; // ignore unused stats
var statName = StatNames[statIndex];
result[ctr++] = $"{statValue} {statName}";
}
return result;
}
private IEnumerable<string> GetStringMoves()
{
var moves = Strings.Move;
foreach (var move in Moves)
{
if (move == 0 || move >= moves.Count)
continue;
if (move != (int)Move.HiddenPower)
{
yield return $"- {moves[move]}";
continue;
}
var type = 1 + HiddenPowerType; // skip Normal
var typeName = Strings.Types[type];
yield return $"- {moves[move]} [{typeName}]";
}
}
private static int GetStatIndexStored(int displayIndex) => displayIndex switch
{
3 => 4,
4 => 5,
5 => 3,
_ => displayIndex,
};
/// <summary>
/// Forces some properties to indicate the set for future display values.
/// </summary>
/// <param name="pk">PKM to convert to string</param>
public void InterpretAsPreview(PKM pk)
{
if (pk.Format <= 2) // Nature preview from IVs
Nature = Experience.GetNatureVC(pk.EXP);
}
/// <summary>
/// Converts the <see cref="PKM"/> data into an importable set format for Pokémon Showdown.
/// </summary>
/// <param name="pk">PKM to convert to string</param>
/// <returns>New ShowdownSet object representing the input <see cref="pk"/></returns>
public ShowdownSet(PKM pk)
{
2022-08-30 22:00:45 +00:00
if (pk.Species == 0)
return;
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
Context = pk.Context;
Nickname = pk.Nickname;
Species = pk.Species;
HeldItem = pk.HeldItem;
Ability = pk.Ability;
pk.GetEVs(EVs);
pk.GetIVs(IVs);
pk.GetMoves(Moves);
Nature = pk.StatNature;
Gender = (uint)pk.Gender < 2 ? pk.Gender : 2;
Friendship = pk.CurrentFriendship;
Level = pk.CurrentLevel;
Shiny = pk.IsShiny;
if (pk is IGigantamax g)
CanGigantamax = g.CanGigantamax;
if (pk is IDynamaxLevel d)
DynamaxLevel = d.DynamaxLevel;
if (Array.IndexOf(Moves, (ushort)Move.HiddenPower) != -1)
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
HiddenPowerType = HiddenPower.GetType(IVs, Context);
if (pk is ITeraType t)
TeraType = t.TeraType;
if (pk is IHyperTrain h)
{
for (int i = 0; i < 6; i++)
{
if (h.IsHyperTrained(i))
IVs[i] = pk.MaxIV;
}
}
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
FormName = ShowdownParsing.GetStringFromForm(Form = pk.Form, Strings, Species, Context);
}
private void ParseFirstLine(ReadOnlySpan<char> first)
{
int itemSplit = first.IndexOf(ItemSplit, StringComparison.Ordinal);
if (itemSplit != -1)
{
var itemName = first[(itemSplit + ItemSplit.Length)..];
var speciesName = first[..itemSplit];
ParseItemName(itemName);
ParseFirstLineNoItem(speciesName);
}
else
{
ParseFirstLineNoItem(first);
}
}
private void ParseItemName(ReadOnlySpan<char> itemName)
{
if (TrySetItem(Context, itemName))
return;
if (TrySetItem(EntityContext.Gen3, itemName))
return;
if (TrySetItem(EntityContext.Gen2, itemName))
return;
InvalidLines.Add($"Unknown Item: {itemName}");
bool TrySetItem(EntityContext context, ReadOnlySpan<char> span)
{
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 items = Strings.GetItemStrings(context);
int item = StringUtil.FindIndexIgnoreCase(items, span);
if (item < 0)
return false;
HeldItem = item;
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
Context = context;
return true;
}
}
private void ParseFirstLineNoItem(ReadOnlySpan<char> line)
{
// Gender Detection
if (line.EndsWith("(M)", StringComparison.Ordinal))
{
line = line[..^3];
Gender = 0;
}
else if (line.EndsWith("(F)", StringComparison.Ordinal))
{
line = line[..^3];
Gender = 1;
}
// Nickname Detection
if (line.IndexOf('(') != -1 && line.IndexOf(')') != -1)
ParseSpeciesNickname(line);
else
ParseSpeciesForm(line);
}
private const string Gmax = "-Gmax";
private bool ParseSpeciesForm(ReadOnlySpan<char> speciesLine)
{
speciesLine = speciesLine.Trim();
if (speciesLine.Length == 0)
return false;
if (speciesLine.EndsWith(Gmax, StringComparison.Ordinal))
{
CanGigantamax = true;
speciesLine = speciesLine[..^Gmax.Length];
}
var speciesIndex = StringUtil.FindIndexIgnoreCase(Strings.specieslist, speciesLine);
if (speciesIndex > 0)
{
// success, nothing else !
Species = (ushort)speciesIndex;
return true;
}
2020-03-14 19:21:42 +00:00
// Form string present.
int end = speciesLine.LastIndexOf('-');
if (end < 0)
return false;
2020-03-14 19:21:42 +00:00
speciesIndex = StringUtil.FindIndexIgnoreCase(Strings.specieslist, speciesLine[..end]);
if (speciesIndex > 0)
{
Species = (ushort)speciesIndex;
FormName = speciesLine[(end + 1)..].ToString();
return true;
}
// failure to parse, check edge cases
foreach (var e in DashedSpecies)
{
var sn = Strings.Species[e];
if (!speciesLine.StartsWith(sn.Replace("♂", "-M").Replace("♀", "-F"), StringComparison.Ordinal))
continue;
Species = e;
FormName = speciesLine[sn.Length..].ToString();
return true;
}
// Version Megas
end = speciesLine[Math.Max(0, end - 1)..].LastIndexOf('-');
if (end < 0)
return false;
speciesIndex = StringUtil.FindIndexIgnoreCase(Strings.specieslist, speciesLine[..end]);
if (speciesIndex > 0)
{
Species = (ushort)speciesIndex;
FormName = speciesLine[(end + 1)..].ToString();
return true;
}
return false;
}
private void ParseSpeciesNickname(ReadOnlySpan<char> line)
{
int index = line.LastIndexOf('(');
Span<char> species = stackalloc char[line.Length];
Span<char> nickname = stackalloc char[line.Length];
if (index > 1) // parenthesis value after: Nickname (Species), correct.
{
line[..index].Trim().CopyTo(nickname);
line[index..].Trim().CopyTo(species);
RemoveAll(ref species, ParenJunk); // Trim out excess data
}
else // parenthesis value before: (Species) Nickname, incorrect
{
int start = index + 1;
int end = line.IndexOf(')');
var tmp = line[start..end];
if (end < line.Length - 2)
{
line[(end + 2)..].CopyTo(nickname);
tmp.CopyTo(species);
}
else // (Species), or garbage
{
tmp.CopyTo(species);
nickname = Span<char>.Empty;
}
}
species = species.TrimEnd('\0');
nickname = nickname.TrimEnd('\0');
if (ParseSpeciesForm(species))
Nickname = nickname.ToString();
else if (ParseSpeciesForm(nickname))
Nickname = species.ToString();
}
private ReadOnlySpan<char> ParseLineMove(ReadOnlySpan<char> line)
{
var startSearch = line[1] == ' ' ? 2 : 1;
var option = line.IndexOf('/');
line = option != -1 ? line[startSearch..option] : line[startSearch..];
var moveString = line.Trim();
var hiddenPowerName = Strings.Move[(int)Move.HiddenPower];
if (!moveString.StartsWith(hiddenPowerName, StringComparison.OrdinalIgnoreCase))
return moveString; // regular move
if (moveString.Length == hiddenPowerName.Length)
return hiddenPowerName;
// Defined Hidden Power
Span<char> type = stackalloc char[moveString.Length - hiddenPowerName.Length];
moveString[hiddenPowerName.Length..].CopyTo(type);
RemoveAll(ref type, ParenJunk); // Trim out excess data
type = type.Trim();
int hpVal = StringUtil.FindIndexIgnoreCase(Strings.types.AsSpan(1), type); // Get HP Type
HiddenPowerType = hpVal;
if (!Array.TrueForAll(IVs, z => z == 31))
{
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 (!HiddenPower.SetIVsForType(hpVal, IVs, Context))
InvalidLines.Add($"Invalid IVs for Hidden Power Type: {type}");
}
else if (hpVal >= 0)
{
Refactoring: Move Source (Legality) (#3560) Rewrites a good amount of legality APIs pertaining to: * Legal moves that can be learned * Evolution chains & cross-generation paths * Memory validation with forgotten moves In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data. The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space. The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation. * `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game. * `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`). * Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
2022-08-03 23:15:27 +00:00
HiddenPower.SetIVs(hpVal, IVs, Context); // Alter IVs
}
else
{
InvalidLines.Add($"Invalid Hidden Power Type: {type}");
}
return hiddenPowerName;
}
private bool ParseLineEVs(ReadOnlySpan<char> line)
{
int start = 0;
while (true)
{
var chunk = line[start..];
var separator = chunk.IndexOf('/');
var len = separator == -1 ? chunk.Length : separator;
var tuple = chunk[..len].Trim();
if (!AbsorbValue(tuple))
InvalidLines.Add($"Invalid EV tuple: {tuple}");
if (separator == -1)
break; // no more stats
start += separator + 1;
}
return true;
bool AbsorbValue(ReadOnlySpan<char> text)
{
var space = text.IndexOf(' ');
if (space == -1)
return false;
var stat = text[(space + 1)..].Trim();
var statIndex = StringUtil.FindIndexIgnoreCase(StatNames, stat);
if (statIndex == -1)
return false;
var value = text[..space].Trim();
if (!ushort.TryParse(value, out var statValue))
return false;
EVs[statIndex] = statValue;
return true;
}
}
private bool ParseLineIVs(ReadOnlySpan<char> line)
{
int start = 0;
while (true)
{
var chunk = line[start..];
var separator = chunk.IndexOf('/');
var len = separator == -1 ? chunk.Length : separator;
var tuple = chunk[..len].Trim();
if (!AbsorbValue(tuple))
InvalidLines.Add($"Invalid IV tuple: {tuple}");
if (separator == -1)
break; // no more stats
start += separator + 1;
}
return true;
bool AbsorbValue(ReadOnlySpan<char> text)
{
var space = text.IndexOf(' ');
if (space == -1)
return false;
var stat = text[(space + 1)..].Trim();
var statIndex = StringUtil.FindIndexIgnoreCase(StatNames, stat);
if (statIndex == -1)
return false;
var value = text[..space].Trim();
if (!byte.TryParse(value, out var statValue))
return false;
IVs[statIndex] = statValue;
return true;
}
}
private static void RemoveAll(ref Span<char> buffer, ReadOnlySpan<char> remove)
{
int ctr = 0;
for (var i = 0; i < buffer.Length; i++)
{
char c = buffer[i];
if (!remove.Contains(c))
buffer[ctr++] = c;
}
if (ctr != buffer.Length)
buffer = buffer[..ctr];
}
}