Skip to content

Commit 22ac68c

Browse files
theletterftheletterf
committed
Add format command
1 parent 0d0910d commit 22ac68c

File tree

3 files changed

+200
-0
lines changed

3 files changed

+200
-0
lines changed

‎src/authoring/Elastic.Documentation.Refactor/FormatService.cs‎

Lines changed: 156 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,156 @@
1+
// Licensed to Elasticsearch B.V under one or more agreements.
2+
// Elasticsearch B.V licenses this file to you under the Apache 2.0 License.
3+
// See the LICENSE file in the project root for more information
4+
5+
using System.Buffers;
6+
using System.IO.Abstractions;
7+
using System.Text;
8+
using Elastic.Documentation.Configuration;
9+
using Elastic.Documentation.Diagnostics;
10+
using Elastic.Documentation.Services;
11+
using Microsoft.Extensions.Logging;
12+
13+
namespace Elastic.Documentation.Refactor;
14+
15+
public class FormatService(
16+
ILoggerFactory logFactory
17+
) : IService
18+
{
19+
private readonly ILogger _logger = logFactory.CreateLogger<FormatService>();
20+
21+
// Collection of irregular whitespace characters that may impair Markdown rendering
22+
private static readonly char[] IrregularWhitespaceChars =
23+
[
24+
'\u000B', // Line Tabulation (\v) - <VT>
25+
'\u000C', // Form Feed (\f) - <FF>
26+
'\u00A0', // No-Break Space - <NBSP>
27+
'\u0085', // Next Line
28+
'\u1680', // Ogham Space Mark
29+
'\u180E', // Mongolian Vowel Separator - <MVS>
30+
'\ufeff', // Zero Width No-Break Space - <BOM>
31+
'\u2000', // En Quad
32+
'\u2001', // Em Quad
33+
'\u2002', // En Space - <ENSP>
34+
'\u2003', // Em Space - <EMSP>
35+
'\u2004', // Tree-Per-Em
36+
'\u2005', // Four-Per-Em
37+
'\u2006', // Six-Per-Em
38+
'\u2007', // Figure Space
39+
'\u2008', // Punctuation Space - <PUNCSP>
40+
'\u2009', // Thin Space
41+
'\u200A', // Hair Space
42+
'\u200B', // Zero Width Space - <ZWSP>
43+
'\u2028', // Line Separator
44+
'\u2029', // Paragraph Separator
45+
'\u202F', // Narrow No-Break Space
46+
'\u205F', // Medium Mathematical Space
47+
'\u3000' // Ideographic Space
48+
];
49+
50+
private static readonly SearchValues<char> IrregularWhitespaceSearchValues = SearchValues.Create(IrregularWhitespaceChars);
51+
52+
public async Task<bool> Format(
53+
IDiagnosticsCollector collector,
54+
string? path,
55+
bool? dryRun,
56+
IFileSystem fs,
57+
Cancel ctx
58+
)
59+
{
60+
var isDryRun = dryRun ?? false;
61+
var rootPath = string.IsNullOrEmpty(path) ? fs.Directory.GetCurrentDirectory() : path;
62+
var rootDir = fs.DirectoryInfo.New(rootPath);
63+
64+
if (!rootDir.Exists)
65+
{
66+
collector.EmitError(string.Empty, $"Directory not found: {rootPath}");
67+
return false;
68+
}
69+
70+
_logger.LogInformation("Formatting documentation in: {Path}", rootDir.FullName);
71+
if (isDryRun)
72+
_logger.LogInformation("Running in dry-run mode - no files will be modified");
73+
74+
var markdownFiles = rootDir.GetFiles("*.md", SearchOption.AllDirectories);
75+
var totalFilesProcessed = 0;
76+
var totalFilesModified = 0;
77+
var totalReplacements = 0;
78+
79+
foreach (var file in markdownFiles)
80+
{
81+
if (ctx.IsCancellationRequested)
82+
break;
83+
84+
totalFilesProcessed++;
85+
var (modified, replacements) = await ProcessFile(file, isDryRun, fs);
86+
87+
if (modified)
88+
{
89+
totalFilesModified++;
90+
totalReplacements += replacements;
91+
_logger.LogInformation("Fixed {Count} irregular whitespace(s) in: {File}", replacements, GetRelativePath(rootDir, file));
92+
}
93+
}
94+
95+
_logger.LogInformation("");
96+
_logger.LogInformation("Formatting complete:");
97+
_logger.LogInformation(" Files processed: {Processed}", totalFilesProcessed);
98+
_logger.LogInformation(" Files modified: {Modified}", totalFilesModified);
99+
_logger.LogInformation(" Total replacements: {Replacements}", totalReplacements);
100+
101+
if (isDryRun && totalFilesModified > 0)
102+
{
103+
_logger.LogInformation("");
104+
_logger.LogInformation("Run without --dry-run to apply changes");
105+
}
106+
107+
return true;
108+
}
109+
110+
private static async Task<(bool modified, int replacements)> ProcessFile(IFileInfo file, bool isDryRun, IFileSystem fs)
111+
{
112+
var content = await fs.File.ReadAllTextAsync(file.FullName);
113+
var modified = false;
114+
var replacements = 0;
115+
116+
// Check if file contains any irregular whitespace
117+
if (content.AsSpan().IndexOfAny(IrregularWhitespaceSearchValues) == -1)
118+
return (false, 0);
119+
120+
// Replace irregular whitespace with regular spaces
121+
var sb = new StringBuilder(content.Length);
122+
foreach (var c in content)
123+
{
124+
if (IrregularWhitespaceSearchValues.Contains(c))
125+
{
126+
_ = sb.Append(' ');
127+
replacements++;
128+
modified = true;
129+
}
130+
else
131+
{
132+
_ = sb.Append(c);
133+
}
134+
}
135+
136+
if (modified && !isDryRun)
137+
{
138+
await fs.File.WriteAllTextAsync(file.FullName, sb.ToString());
139+
}
140+
141+
return (modified, replacements);
142+
}
143+
144+
private static string GetRelativePath(IDirectoryInfo root, IFileInfo file)
145+
{
146+
var rootPath = root.FullName.TrimEnd(Path.DirectorySeparatorChar, Path.AltDirectorySeparatorChar);
147+
var filePath = file.FullName;
148+
149+
if (filePath.StartsWith(rootPath, StringComparison.OrdinalIgnoreCase))
150+
{
151+
return filePath.Substring(rootPath.Length).TrimStart(Path.DirectorySeparatorChar, Path.AltDirectorySeparatorChar);
152+
}
153+
154+
return filePath;
155+
}
156+
}

‎src/tooling/docs-builder/Commands/FormatCommand.cs‎

Lines changed: 43 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,43 @@
1+
// Licensed to Elasticsearch B.V under one or more agreements.
2+
// Elasticsearch B.V licenses this file to you under the Apache 2.0 License.
3+
// See the LICENSE file in the project root for more information
4+
5+
using System.IO.Abstractions;
6+
using ConsoleAppFramework;
7+
using Elastic.Documentation.Configuration;
8+
using Elastic.Documentation.Diagnostics;
9+
using Elastic.Documentation.Refactor;
10+
using Elastic.Documentation.Services;
11+
using Microsoft.Extensions.Logging;
12+
13+
namespace Documentation.Builder.Commands;
14+
15+
internal sealed class FormatCommand(
16+
ILoggerFactory logFactory,
17+
IDiagnosticsCollector collector
18+
)
19+
{
20+
/// <summary>
21+
/// Format documentation files by fixing common issues like irregular whitespace
22+
/// </summary>
23+
/// <param name="path"> -p, Path to the documentation folder, defaults to pwd</param>
24+
/// <param name="dryRun">Preview changes without modifying files</param>
25+
/// <param name="ctx"></param>
26+
[Command("")]
27+
public async Task<int> Format(
28+
string? path = null,
29+
bool? dryRun = null,
30+
Cancel ctx = default
31+
)
32+
{
33+
await using var serviceInvoker = new ServiceInvoker(collector);
34+
35+
var service = new FormatService(logFactory);
36+
var fs = new FileSystem();
37+
38+
serviceInvoker.AddCommand(service, (path, dryRun, fs),
39+
async static (s, collector, state, ctx) => await s.Format(collector, state.path, state.dryRun, state.fs, ctx)
40+
);
41+
return await serviceInvoker.InvokeAsync(ctx);
42+
}
43+
}

‎src/tooling/docs-builder/Program.cs‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -38,6 +38,7 @@
3838
app.Add<MoveCommand>("mv");
3939
app.Add<ServeCommand>("serve");
4040
app.Add<IndexCommand>("index");
41+
app.Add<FormatCommand>("format");
4142

4243
//assembler commands
4344

0 commit comments

Comments
 (0)