mirror of
https://github.com/tonytins/cstdotnet.git
synced 2025-03-14 21:51:25 +00:00
Improved normalization algorithm
This commit is contained in:
Tony Bark
parent
84624111ea
commit
d1c589e2c0
4 changed files with 105 additions and 126 deletions
152
CSTNet/CST.cs
152
CSTNet/CST.cs
|
|
@ -8,107 +8,79 @@ namespace CSTNet;
|
|||
|
||||
public static class CST
|
||||
{
|
||||
const char CARET = '^';
|
||||
const string LF = "\u000A";
|
||||
const string CR = "\u000D";
|
||||
const string CRLF = "\u000D\u000A";
|
||||
const string LS = "\u2028";
|
||||
const char CARET = '^';
|
||||
const string LF = "\u000A";
|
||||
const string CR = "\u000D";
|
||||
const string CRLF = "\u000D\u000A";
|
||||
const string LS = "\u2028";
|
||||
|
||||
/// <summary>
|
||||
/// Gets the value from the digit-based key.
|
||||
/// </summary>
|
||||
/// <returns>Returns the entry</returns>
|
||||
public static string Parse(string content, int key) => Parse(content, key.ToString());
|
||||
/// <summary>
|
||||
/// Gets the value from the digit-based key.
|
||||
/// </summary>
|
||||
/// <returns>Returns the entry</returns>
|
||||
public static string Parse(string content, int key) => Parse(content, key.ToString());
|
||||
|
||||
/// <summary>
|
||||
/// Gets the value from the string-based key.
|
||||
/// </summary>
|
||||
/// <returns>Returns the entry</returns>
|
||||
public static string Parse(string content, string key)
|
||||
{
|
||||
var entries = NormalizeEntries(content);
|
||||
return GetEntry(entries, key);
|
||||
}
|
||||
/// <summary>
|
||||
/// Gets the value from the string-based key.
|
||||
/// </summary>
|
||||
/// <returns>Returns the entry</returns>
|
||||
public static string Parse(string content, string key)
|
||||
{
|
||||
var entries = NormalizeEntries(content);
|
||||
return GetEntry(entries, key);
|
||||
}
|
||||
|
||||
#if (NET8_0 && DEBUG)
|
||||
[UnmanagedCallersOnly(EntryPoint = "parse")]
|
||||
public static IntPtr Parse(IntPtr content, IntPtr key)
|
||||
{
|
||||
// => Parse(Marshal.PtrToStringAnsi(content), Marshal.PtrToStringAnsi(key));
|
||||
var entries = NormalizeEntries(Marshal.PtrToStringAnsi(content));
|
||||
return Marshal.StringToHGlobalAnsi(GetEntry(entries, Marshal.PtrToStringAnsi(key)));
|
||||
}
|
||||
[UnmanagedCallersOnly(EntryPoint = "parse")]
|
||||
public static IntPtr Parse(IntPtr content, IntPtr key)
|
||||
{
|
||||
// => Parse(Marshal.PtrToStringAnsi(content), Marshal.PtrToStringAnsi(key));
|
||||
var entries = NormalizeEntries(Marshal.PtrToStringAnsi(content));
|
||||
return Marshal.StringToHGlobalAnsi(GetEntry(entries, Marshal.PtrToStringAnsi(key)));
|
||||
}
|
||||
#endif
|
||||
/// <summary>
|
||||
/// Normalizes the content by replacing various newline characters with Environment.NewLine and filters out comments.
|
||||
/// </summary>
|
||||
/// <param name="content">The content to normalize.</param>
|
||||
/// <returns>An enumerable of normalized lines.</returns>
|
||||
public static IEnumerable<string> NormalizeEntries(string content)
|
||||
{
|
||||
var newLines = new[] { LF, CR, CRLF, LS };
|
||||
|
||||
/// <summary>
|
||||
/// Replaces the document's line endings with the native system line endings.
|
||||
/// </summary>
|
||||
/// <remarks>This stage ensures there are no crashes during parsing.</remarks>
|
||||
/// <param name="content">The content of the document.</param>
|
||||
/// <returns>The document's content with native system line endings.</returns>
|
||||
static IEnumerable<string> NormalizeEntries(string content)
|
||||
{
|
||||
// Check if the document already uses native system line endings.
|
||||
if (!content.Contains(Environment.NewLine))
|
||||
{
|
||||
// If not, check for and replace other line ending types.
|
||||
if (content.Contains(LF))
|
||||
content = content.Replace(LF,
|
||||
Environment.NewLine);
|
||||
content = newLines.Aggregate(content, (current, nl) => current.Replace(nl, Environment.NewLine));
|
||||
|
||||
if (content.Contains(CR))
|
||||
content = content.Replace(CR,
|
||||
Environment.NewLine);
|
||||
return content.Split($"{CARET}{Environment.NewLine}", StringSplitOptions.RemoveEmptyEntries)
|
||||
.Where(line => !line.StartsWith("//") && !line.StartsWith('#') && !line.StartsWith("/*") && !line.EndsWith("*/"));
|
||||
}
|
||||
|
||||
if (content.Contains(CRLF))
|
||||
content = content.Replace(CRLF,
|
||||
Environment.NewLine);
|
||||
/// <summary>
|
||||
/// Retrieves the value for the specified key from the given entries.
|
||||
/// </summary>
|
||||
/// <param name="entries">The entries to search through.</param>
|
||||
/// <param name="key">The key to search for.</param>
|
||||
/// <returns>The value for the specified key, or a default string if not found.</returns>
|
||||
static string GetEntry(IEnumerable<string> entries, string key)
|
||||
{
|
||||
// Iterate through the entries.
|
||||
foreach (var entry in entries)
|
||||
{
|
||||
// If the line doesn't start with the key, keep searching.
|
||||
if (!entry.StartsWith(key))
|
||||
continue;
|
||||
|
||||
if (content.Contains(LS))
|
||||
content = content.Replace(LS,
|
||||
Environment.NewLine);
|
||||
}
|
||||
// Locate the index of the caret character.
|
||||
var startIndex = entry.IndexOf(CARET);
|
||||
// Get the line from the caret character to the end of the string.
|
||||
var line = entry[startIndex..];
|
||||
|
||||
// Split the content by the caret and newline characters.
|
||||
var lines = content.Split(new[] { $"{CARET}{Environment.NewLine}" },
|
||||
StringSplitOptions.RemoveEmptyEntries);
|
||||
// Return the line with the caret characters trimmed.
|
||||
return line.TrimStart(CARET).TrimEnd(CARET);
|
||||
}
|
||||
|
||||
// Filter out any lines that start with "//", "#", "/*", or end with "*/".
|
||||
return lines.Where(line =>
|
||||
!line.StartsWith("//") &&
|
||||
!line.StartsWith("#") &&
|
||||
!line.StartsWith("/*") &&
|
||||
!line.EndsWith("*/"))
|
||||
.AsEnumerable();
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Retrieves the value for the specified key from the given entries.
|
||||
/// </summary>
|
||||
/// <param name="entries">The entries to search through.</param>
|
||||
/// <param name="key">The key to search for.</param>
|
||||
/// <returns>The value for the specified key, or a default string if not found.</returns>
|
||||
static string GetEntry(IEnumerable<string> entries, string key)
|
||||
{
|
||||
// Iterate through the entries.
|
||||
foreach (var entry in entries)
|
||||
{
|
||||
// If the line doesn't start with the key, keep searching.
|
||||
if (!entry.StartsWith(key))
|
||||
continue;
|
||||
|
||||
// Locate the index of the caret character.
|
||||
var startIndex = entry.IndexOf(CARET);
|
||||
// Get the line from the caret character to the end of the string.
|
||||
var line = entry[startIndex..];
|
||||
|
||||
// Return the line with the caret characters trimmed.
|
||||
return line.TrimStart(CARET).TrimEnd(CARET);
|
||||
}
|
||||
|
||||
// If no entry is found, return a default string.
|
||||
return "***MISSING***";
|
||||
}
|
||||
// If no entry is found, return a default string.
|
||||
return "***MISSING***";
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -1,25 +1,25 @@
|
|||
<Project Sdk="Microsoft.NET.Sdk">
|
||||
|
||||
<PropertyGroup>
|
||||
<TargetFrameworks>net6.0;net8.0</TargetFrameworks>
|
||||
<Version>2.1.100</Version>
|
||||
<Nullable>enable</Nullable>
|
||||
<LangVersion>latest</LangVersion>
|
||||
<ImplicitUsings>enable</ImplicitUsings>
|
||||
<Authors>Tony Bark</Authors>
|
||||
<PackageDescription>
|
||||
Caret-Separated Text (or CST) is a key-value pair format represented by digits or words
|
||||
as keys and the value as text enclosed between carets. ([key] ^[value]^)
|
||||
<PropertyGroup>
|
||||
<TargetFrameworks>net6.0;net8.0</TargetFrameworks>
|
||||
<Version>2.1.101-alpha</Version>
|
||||
<Nullable>enable</Nullable>
|
||||
<LangVersion>latest</LangVersion>
|
||||
<ImplicitUsings>enable</ImplicitUsings>
|
||||
<Authors>Tony Bark</Authors>
|
||||
<PackageDescription>
|
||||
Caret-Separated Text (or CST) is a key-value pair format represented by digits or words
|
||||
as keys and the value as text enclosed between carets. ([key] ^[value]^)
|
||||
|
||||
CSTNet provides you the framework for parsing the CST format.
|
||||
</PackageDescription>
|
||||
<RepositoryUrl>https://github.com/tonytins/cstdotnet</RepositoryUrl>
|
||||
<PackageLicenseExpression>BSD-3-Clause</PackageLicenseExpression>
|
||||
</PropertyGroup>
|
||||
CSTNet provides you the framework for parsing the CST format.
|
||||
</PackageDescription>
|
||||
<RepositoryUrl>https://github.com/tonytins/cstdotnet</RepositoryUrl>
|
||||
<PackageLicenseExpression>BSD-3-Clause</PackageLicenseExpression>
|
||||
</PropertyGroup>
|
||||
|
||||
<!-- Support AOT on .NET 8+ -->
|
||||
<PropertyGroup Condition=" '$(TargetFramework)' == 'net8.0' ">
|
||||
<IsAotCompatible>true</IsAotCompatible>
|
||||
</PropertyGroup>
|
||||
<!-- Support AOT on .NET 8+ -->
|
||||
<PropertyGroup Condition=" '$(TargetFramework)' == 'net8.0' ">
|
||||
<IsAotCompatible>true</IsAotCompatible>
|
||||
</PropertyGroup>
|
||||
|
||||
</Project>
|
||||
|
|
@ -4,10 +4,10 @@ namespace CSTNet;
|
|||
|
||||
public class UIText : IUIText
|
||||
{
|
||||
/// <summary>
|
||||
/// The language of the text.
|
||||
/// </summary>
|
||||
string Language { get; set; } = "english";
|
||||
/// <summary>
|
||||
/// The language of the text.
|
||||
/// </summary>
|
||||
string Language { get; set; } = "english";
|
||||
|
||||
/// <summary>
|
||||
/// The base directory for the language files.
|
||||
|
|
|
|||
33
README.md
33
README.md
|
|
@ -1,17 +1,10 @@
|
|||
# CST.NET
|
||||
# [CST.NET](http://CST.NET)
|
||||
|
||||
<p align="center">
|
||||
<a href="https://github.com/tonytins/cstdotnet/blob/main/LICENSE"><img src="https://img.shields.io/github/license/tonytins/cstdotnet" alt="GitHub license"></a>
|
||||
<a href="https://github.com/tonytins/cstdotnet/actions?query=workflow%3Abuild.yml"><img src="https://img.shields.io/github/actions/workflow/status/tonytins/cstdotnet/build.yml" alt="GitHub Workflow Status"></a>
|
||||
<img src="https://img.shields.io/github/commit-activity/w/tonytins/cstdotnet" alt="GitHub commit activity">
|
||||
<a href="code_of_conduct.md"></br>
|
||||
<img src="https://img.shields.io/codeclimate/maintainability-percentage/tonytins/cstdotnet" alt="Code Climate maintainability">
|
||||
<img src="https://img.shields.io/nuget/dt/CSTNet" alt="NuGet Downloads"> <a href="https://www.nuget.org/packages/tonybark.updatetools"><img src="https://img.shields.io/nuget/v/cstnet.svg" /></a></br><img src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg" alt="Contributor Covenant"></a></br>
|
||||
</p>
|
||||
<p align="center"><a href="https://github.com/tonytins/cstdotnet/blob/main/LICENSE"><img src="https://img.shields.io/github/license/tonytins/cstdotnet" alt="GitHub license"></a> <a href="https://github.com/tonytins/cstdotnet/actions?query=workflow%3Abuild.yml"><img src="https://img.shields.io/github/actions/workflow/status/tonytins/cstdotnet/build.yml" alt="GitHub Workflow Status"></a> <img src="https://img.shields.io/github/commit-activity/w/tonytins/cstdotnet" alt="GitHub commit activity"> <a href="code_of_conduct.md"></br> <img src="https://img.shields.io/codeclimate/maintainability-percentage/tonytins/cstdotnet" alt="Code Climate maintainability"> <img src="https://img.shields.io/nuget/dt/CSTNet" alt="NuGet Downloads"> <a href="https://www.nuget.org/packages/tonybark.updatetools"><img src="https://img.shields.io/nuget/v/cstnet.svg" /></a></br><img src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg" alt="Contributor Covenant"></a></br> </p>
|
||||
|
||||
CST.NET is a library for parsing Maxis' key-value pair format. It can be used in conjunction with your own custom frameworks, or the original ``UIText`` APIs.
|
||||
CST.NET is a library for parsing Maxis' key-value pair format. It can be used in conjunction with your own custom frameworks, or the original `UIText` APIs.
|
||||
|
||||
Caret-Separated Text (or CST) is a key-value pair format represented by digits or words as keys and the value as text enclosed between carets. (e.g. ``<key> ^<text>^``) Any text which is not enclosed with carets is considered a comment and ignored. Neither strings nor comments may use the caret character.
|
||||
Caret-Separated Text (or CST) is a key-value pair format represented by digits or words as keys and the value as text enclosed between carets. (e.g. `<key> ^<text>^`) Any text which is not enclosed with carets is considered a comment and ignored. Neither strings nor comments may use the caret character.
|
||||
|
||||
## Changelog
|
||||
|
||||
|
|
@ -23,7 +16,7 @@ See [/docs](./doc/README.md).
|
|||
|
||||
## To-do
|
||||
|
||||
- [ ] Support for parameters (e.g. ``%1``)
|
||||
- [ ] Support for parameters (e.g. `%1`)
|
||||
|
||||
## Known issues
|
||||
|
||||
|
|
@ -38,6 +31,20 @@ See [/docs](./doc/README.md).
|
|||
- [.NET Interactive](https://github.com/dotnet/interactive/blob/main/README.md) for notebooks (optional).
|
||||
- [VSCode Extension](https://marketplace.visualstudio.com/items?itemName=ms-dotnettools.dotnet-interactive-vscode) or [nteract](https://nteract.io/).
|
||||
|
||||
|
||||
# Contributing
|
||||
|
||||
You can contribute to CST.NET by testing cutting edge features in the latest releases, filing bugs, and joining in the discussion on our forums!
|
||||
|
||||
* [Getting Started](https://github.com/tonytins/cstdotnet/wiki)
|
||||
* [Project Structure](https://github.com/tonytins/cstdotnet/wiki/Project-structure)
|
||||
* [Coding Standards](https://github.com/tonytins/cstdotnet/wiki/Coding-standards)
|
||||
* [Pull Requests](https://github.com/tonytins/cstdotnet/pulls): [Open](https://github.com/tonytins/cstdotnet/pulls)/[Closed](https://github.com/tonytins/cstdotnet/issues?q=is%3Apr+is%3Aclosed)
|
||||
|
||||
Looking for something to do? Check out the issues tagged as [help wanted](https://github.com/tonytins/cstdotnet/labels/help%20wanted) to get started.
|
||||
|
||||
Regarding translations, full object and UI translations should currently be released on the forums. This is far from perfect and is due to be reworked. Stay tuned!
|
||||
|
||||
## License
|
||||
|
||||
I license this project under the BSD-3-Clause license - see [LICENSE](LICENSE) for details.
|
||||
I license this project under the BSD-3-Clause license - see [LICENSE](LICENSE) for details.
|
||||
Loading…
Add table
Reference in a new issue