feat: Add FileReader tool

.gitignore

@@ -367,4 +367,4 @@ appsettings.Local.json
 docker/ollama-cache
 docker/vllm-cache
 *.xlsx
-TODO.md
+*.pdf

src/Cellm/Cellm.csproj

@@ -34,6 +34,7 @@
     <PackageReference Include="Microsoft.Extensions.Logging.Debug" Version="8.0.0" />
     <PackageReference Include="Microsoft.Extensions.Options" Version="8.0.2" />
     <PackageReference Include="Microsoft.Extensions.Options.ConfigurationExtensions" Version="8.0.0" />
+    <PackageReference Include="PdfPig" Version="0.1.9" />
     <PackageReference Include="Sentry.Extensions.Logging" Version="4.10.2" />
     <PackageReference Include="Sentry.Profiling" Version="4.10.2" />
     <Content Include="appsettings.json">

src/Cellm/Models/Client.cs

@@ -57,6 +57,10 @@ public async Task<Prompt> Send(Prompt prompt, string? provider, Uri? baseAddress
         {
             throw new CellmException($"Method not supported: {ex.Message}", ex);
         }
+        catch (FileReaderException ex)
+        {
+            throw new CellmException($"File could not be read: {ex.Message}", ex);
+        }
         catch (NullReferenceException ex)
         {
             throw new CellmException($"Null reference error: {ex.Message}", ex);

src/Cellm/Services/ServiceLocator.cs

@@ -9,6 +9,7 @@
 using Cellm.Models.PipelineBehavior;
 using Cellm.Services.Configuration;
 using Cellm.Tools;
+using Cellm.Tools.FileReader;
 using Cellm.Tools.Glob;
 using ExcelDna.Integration;
 using MediatR;
@@ -98,11 +99,13 @@ private static IServiceCollection ConfigureServices(IServiceCollection services)
             .AddMemoryCache()
             .AddSingleton<Cache>();
 
-
         // Tools
         services
             .AddSingleton<ToolRunner>()
-            .AddSingleton<ToolFactory>();
+            .AddSingleton<ToolFactory>()
+            .AddSingleton<FileReaderFactory>()
+            .AddSingleton<IFileReader, PdfReader>()
+            .AddSingleton<IFileReader, TextReader>();
 
         // Model Providers
         var rateLimiterConfiguration = configuration.GetRequiredSection(nameof(RateLimiterConfiguration)).Get<RateLimiterConfiguration>()

src/Cellm/Tools/FileReader/Exceptions/FileReaderException.cs

@@ -0,0 +1,11 @@
+public class FileReaderException : Exception
+{
+    public FileReaderException(string message) : base(message)
+    {
+    }
+
+    public FileReaderException(string message, Exception innerException)
+        : base(message, innerException)
+    {
+    }
+}

src/Cellm/Tools/FileReader/FileReaderFactory.cs

@@ -0,0 +1,17 @@
+namespace Cellm.Tools.FileReader;
+
+internal class FileReaderFactory
+{
+    private readonly IEnumerable<IFileReader> _readers;
+
+    public FileReaderFactory(IEnumerable<IFileReader> readers)
+    {
+        _readers = readers;
+    }
+
+    public IFileReader GetReader(string filePath)
+    {
+        return _readers.FirstOrDefault(r => r.CanRead(filePath))
+            ?? throw new NotSupportedException($"No reader found for file: {filePath}");
+    }
+}

src/Cellm/Tools/FileReader/FileReaderRequest.cs

@@ -0,0 +1,6 @@
+using System.ComponentModel;
+using MediatR;
+
+namespace Cellm.Tools.FileReader;
+
+internal record FileReaderRequest([Description("The absolute path to the file.")] string FilePath) : IRequest<FileReaderResponse>;

src/Cellm/Tools/FileReader/FileReaderRequestHandler.cs

@@ -0,0 +1,35 @@
+using MediatR;
+
+namespace Cellm.Tools.FileReader;
+
+internal class FileReaderRequestHandler : IRequestHandler<FileReaderRequest, FileReaderResponse>
+{
+    private readonly FileReaderFactory _fileReaderFactory;
+
+    public FileReaderRequestHandler(FileReaderFactory fileReaderFactory)
+    {
+        _fileReaderFactory = fileReaderFactory;
+    }
+
+    public async Task<FileReaderResponse> Handle(FileReaderRequest request, CancellationToken cancellationToken)
+    {
+        try
+        {
+            var reader = _fileReaderFactory.GetReader(request.FilePath);
+            var content = await reader.ReadContent(request.FilePath, cancellationToken);
+            return new FileReaderResponse(content);
+        }
+        catch (ArgumentException ex)
+        {
+            throw new FileReaderException($"Failed to read file: {request.FilePath}", ex);
+        }
+        catch (FileNotFoundException ex)
+        {
+            throw new FileReaderException($"Failed to read file: {request.FilePath}", ex);
+        }
+        catch (NotSupportedException ex)
+        {
+            throw new FileReaderException($"Failed to read file: {request.FilePath}", ex);
+        }
+    }
+}

src/Cellm/Tools/FileReader/FileReaderResponse.cs

@@ -0,0 +1,5 @@
+using System.ComponentModel;
+
+namespace Cellm.Tools.FileReader;
+
+internal record FileReaderResponse([Description("The content of the file")] string FileContent);

src/Cellm/Tools/FileReader/IFileReader.cs

@@ -0,0 +1,6 @@
+internal interface IFileReader
+{
+    public bool CanRead(string filePath);
+
+    public Task<string> ReadContent(string filePath, CancellationToken cancellationToken);
+}

src/Cellm/Tools/FileReader/PdfReader.cs

@@ -0,0 +1,50 @@
+using System.Text;
+using UglyToad.PdfPig;
+using UglyToad.PdfPig.Content;
+using UglyToad.PdfPig.DocumentLayoutAnalysis.PageSegmenter;
+using UglyToad.PdfPig.DocumentLayoutAnalysis.ReadingOrderDetector;
+using UglyToad.PdfPig.DocumentLayoutAnalysis.WordExtractor;
+
+internal class PdfReader : IFileReader
+{
+    public bool CanRead(string filePath)
+    {
+        if (string.IsNullOrWhiteSpace(filePath))
+        {
+            throw new ArgumentException("File path cannot be null or empty.", nameof(filePath));
+        }
+
+        if (!File.Exists(filePath))
+        {
+            throw new FileNotFoundException($"File not found: {filePath}");
+        }
+
+        return Path.GetExtension(filePath).ToLower() is ".pdf";
+    }
+
+    public Task<string> ReadContent(string filePath, CancellationToken cancellationToken)
+    {
+        var stringBuilder = new StringBuilder();
+
+        using (PdfDocument document = PdfDocument.Open(filePath))
+        {
+            foreach (Page page in document.GetPages())
+            {
+                var pageSegmenter = DocstrumBoundingBoxes.Instance;
+                var textBlocks = pageSegmenter.GetBlocks(page.GetWords());
+
+                var readingOrder = UnsupervisedReadingOrderDetector.Instance;
+                var orderedTextBlocks = readingOrder.Get(textBlocks);
+
+                foreach (var orderedTextBlock in orderedTextBlocks)
+                {
+                    var text = orderedTextBlock.Text.Normalize(NormalizationForm.FormKC);
+                    stringBuilder.Append(text);
+                    stringBuilder.AppendLine();
+                }
+            }
+        }
+
+        return Task.FromResult(stringBuilder.ToString());
+    }
+}

src/Cellm/Tools/FileReader/TextReader.cs

@@ -0,0 +1,35 @@
+using System.Text;
+
+internal class TextReader : IFileReader
+{
+    private readonly List<string> _extensions;
+
+    public TextReader()
+    {
+        _extensions = new List<string> { ".c", ".cpp", ".cs", ".csv", ".cxx", ".h", ".hxx", ".html", ".java", ".json", ".jsonl", ".md", ".php", ".py", ".rb", ".txt", ".xml" };
+    }
+
+    public bool CanRead(string filePath)
+    {
+        if (string.IsNullOrWhiteSpace(filePath))
+        {
+            throw new ArgumentException("File path cannot be null or empty.", nameof(filePath));
+        }
+
+        if (!File.Exists(filePath))
+        {
+            throw new FileNotFoundException($"File not found: {filePath}");
+        }
+
+        return _extensions.Contains(Path.GetExtension(filePath).ToLowerInvariant());
+    }
+
+    public async Task<string> ReadContent(string filePath, CancellationToken cancellationToken)
+    {
+        using (var stream = File.OpenRead(filePath))
+        using (var reader = new StreamReader(stream, Encoding.UTF8, true))
+        {
+            return await reader.ReadToEndAsync();
+        }
+    }
+}

src/Cellm/Tools/ToolRunner.cs

@@ -1,5 +1,6 @@
 using Cellm.Models;
 using Cellm.Prompts;
+using Cellm.Tools.FileReader;
 using Cellm.Tools.Glob;
 using MediatR;
 
@@ -17,7 +18,10 @@ public ToolRunner(ISender sender, Serde serde, ToolFactory toolFactory)
         _sender = sender;
         _serde = serde;
         _toolFactory = toolFactory;
-        _toolTypes = new List<Type>() { typeof(GlobRequest) };
+        _toolTypes = new List<Type>() {
+            typeof(GlobRequest),
+            typeof(FileReaderRequest)
+        };
     }
 
     public List<Tool> GetTools()
@@ -30,6 +34,7 @@ public async Task<string> Run(ToolCall toolCall)
         return toolCall.Name switch
         {
             nameof(GlobRequest) => await Run<GlobRequest>(toolCall.Arguments),
+            nameof(FileReaderRequest) => await Run<FileReaderRequest>(toolCall.Arguments),
             _ => throw new ArgumentException($"Unsupported tool: {toolCall.Name}")
         };
     }

src/Cellm/packages.lock.json

@@ -171,6 +171,12 @@
           "Microsoft.Extensions.Primitives": "8.0.0"
         }
       },
+      "PdfPig": {
+        "type": "Direct",
+        "requested": "[0.1.9, )",
+        "resolved": "0.1.9",
+        "contentHash": "VU27oq5O0rpD+zWiD639xXkSHtsNXqbPpYV0pzYg9VPhDibH1cq1i67f9DuW2WnzDVEti0HipLcZxDhdayIzGg=="
+      },
       "Sentry.Extensions.Logging": {
         "type": "Direct",
         "requested": "[4.10.2, )",