refactor: Simplify documentation generation process and enhance content handling

Author: 239573049

src/KoalaWiki/KernelFactory.cs

@@ -26,7 +26,7 @@ public static Kernel GetKernel(string chatEndpoint,
         string apiKey,
         string gitPath,
         string model, bool isCodeAnalysis = true,
-        List<string>? files = null)
+        List<string>? files = null, Action<IKernelBuilder>? kernelBuilderAction = null)
     {
         using var activity = Activity.Current?.Source.StartActivity();
         activity?.SetTag("model", model);
@@ -99,7 +99,7 @@ public static Kernel GetKernel(string chatEndpoint,
         }
 
         // 添加文件函数
-        var fileFunction = new FileFunction(gitPath,files);
+        var fileFunction = new FileFunction(gitPath, files);
         kernelBuilder.Plugins.AddFromObject(fileFunction);
         kernelBuilder.Plugins.AddFromType<AgentFunction>();
         activity?.SetTag("plugins.file_function", "loaded");
@@ -111,6 +111,8 @@ public static Kernel GetKernel(string chatEndpoint,
             activity?.SetTag("plugins.code_analyze_function", "loaded");
         }
 
+        kernelBuilderAction?.Invoke(kernelBuilder);
+        
         var kernel = kernelBuilder.Build();
         kernel.FunctionInvocationFilters.Add(new FunctionResultInterceptor());
 

src/KoalaWiki/KoalaWarehouse/DocumentPending/DocsFunction.cs

@@ -0,0 +1,123 @@
+using System.ComponentModel;
+
+namespace KoalaWiki.KoalaWarehouse.DocumentPending;
+
+public class DocsFunction
+{
+    /// <summary>
+    /// 写入内容
+    /// </summary>
+    /// <returns></returns>
+    [KernelFunction("Write"), Description("""
+                                          Generate the content for the document.
+                                          Usage:
+                                          - This tool will overwrite the existing content.
+                                          - Always edit the existing content first. Do not overwrite it unless explicitly required. 
+                                          - Use emojis only when the user explicitly requests it. Avoid adding emojis to the document unless specifically asked to do so.
+                                          """)]
+    public string Write(
+        [Description("The content to write")] string content)
+    {
+        Content = content;
+        if (string.IsNullOrEmpty(Content))
+        {
+            return "<system-reminder>Content cannot be empty.</system-reminder>";
+        }
+
+        Content = Content.Trim();
+        return @$"<system-reminder>Write successful</system-reminder>";
+    }
+
+    [KernelFunction("Edit"), Description("""
+                                         Perform precise string replacement operations in the generated document.
+                                         Usage:
+                                         - Before making any edits, you must use the `Read` tool at least once in the conversation. If you attempt to edit without reading the file, the tool will report an error.
+                                         - When editing the text output from the `Read` tool, make sure to retain its exact indentation (tabs/spaces), that is, the form that appears after the line number prefix. The line number prefix format is: space + line number + tab. Everything after that tab is the actual file content and must match it. Do not include any components of the line number prefix in the old string or new string.
+                                         - Always prioritize editing existing files in the code repository. Do not overwrite the content unless explicitly required.
+                                         - Use emojis only when the user explicitly requests it. Do not add emojis to the file unless required.
+                                         - If the `oldString` is not unique in the file, the edit will fail. Either provide a longer string with more context to make it unique, or use `replaceAll` to change all instances of the "old string".
+                                         - Use `replaceAll` to replace and rename strings throughout the file. This parameter is very useful when renaming variables, etc.
+                                         """)]
+    public string Edit(
+        [Description("The text to replace")]
+        string oldString,
+        [Description("The text to replace it with (must be different from old_string)")]
+        string newString,
+        [Description("Replace all occurences of old_string (default false)")]
+        bool replaceAll = false)
+    {
+        if (string.IsNullOrEmpty(Content))
+        {
+            return "<system-reminder>Document content is empty, please write content first.</system-reminder>";
+        }
+
+        if (string.IsNullOrEmpty(oldString))
+        {
+            return "<system-reminder>Old string cannot be empty.</system-reminder>";
+        }
+
+        if (oldString == newString)
+        {
+            return "<system-reminder>New string must be different from old string.</system-reminder>";
+        }
+
+        if (!Content.Contains(oldString))
+        {
+            return "<system-reminder>Old string not found in document.</system-reminder>";
+        }
+
+        if (!replaceAll && Content.Split(new[] { oldString }, StringSplitOptions.None).Length > 2)
+        {
+            return "<system-reminder>Old string is not unique in the document. Use replaceAll=true to replace all occurrences or provide a longer string with more context.</system-reminder>";
+        }
+
+        if (replaceAll)
+        {
+            Content = Content.Replace(oldString, newString);
+        }
+        else
+        {
+            int index = Content.IndexOf(oldString);
+            Content = Content.Substring(0, index) + newString + Content.Substring(index + oldString.Length);
+        }
+        
+        return @$"<system-reminder>Edit successful</system-reminder>";
+    }
+
+    [KernelFunction("Read"), Description("""
+                                         To read the current generated document content, please note that this method can only read the content of the generated document.
+                                         Usage:
+                                         - By default, it reads up to 2000 lines from the beginning of the file.
+                                         - You can choose to specify the line offset and limit, but it is recommended not to provide these parameters to read the entire file.
+                                         - Any lines exceeding 2000 characters will be truncated.
+                                         - If the file you are reading exists but is empty, you will receive a system warning instead of the file content.
+                                         """)]
+    public string Read(
+        [Description("The line number to start reading from. Only provide if the file is too large to read at once")]
+        int offset,
+        [Description("The number of lines to read. Only provide if the file is too large to read at once.")]
+        int limit = 2000)
+    {
+        var lines = Content.Split(new[] { '\n', '\r' }, StringSplitOptions.RemoveEmptyEntries);
+
+        if (offset < 0 || offset >= lines.Length)
+        {
+            // 读取所有
+            return string.Join("\n", lines);
+        }
+
+        if (limit <= 0 || offset + limit > lines.Length)
+        {
+            // 读取到结尾
+            return string.Join("\n", lines.Skip(offset));
+        }
+
+        // 读取指定范围
+        return string.Join("\n", lines.Skip(offset).Take(limit));
+    }
+
+    /// <summary>
+    /// 内容
+    /// </summary>
+    public string? Content { get; private set; }
+}

src/KoalaWiki/KoalaWarehouse/DocumentPending/DocumentPendingService.cs

@@ -236,14 +236,16 @@ private static async Task<DocumentFileItem> ProcessCatalogueItems(DocumentCatalo
         string gitRepository, string branch, string path, ClassifyType? classify, List<string> files)
     {
         DocumentContext.DocumentStore = new DocumentStore();
+
+        var docs = new DocsFunction();
         // 为每个文档处理创建独立的Kernel实例，避免状态管理冲突
         var documentKernel = KernelFactory.GetKernel(
             OpenAIOptions.Endpoint,
             OpenAIOptions.ChatApiKey,
             path,
             OpenAIOptions.ChatModel,
             false, // 文档生成不需要代码分析功能
-            files
+            files, (builder => { builder.Plugins.AddFromObject(docs); })
         );
 
         var chat = documentKernel.Services.GetService<IChatCompletionService>();
@@ -288,7 +290,7 @@ private static async Task<DocumentFileItem> ProcessCatalogueItems(DocumentCatalo
         // 保存原始内容，防止精炼失败时丢失
         var originalContent = sr.ToString();
 
-        if (string.IsNullOrEmpty(originalContent) && count < 3)
+        if (string.IsNullOrEmpty(docs.Content) && count < 3)
         {
             count++;
             goto reset;
@@ -371,6 +373,8 @@ 7. ENSURE all enhancements are based on the code files analyzed in the original
                 history.AddUserMessage(refineContents);
 
                 var refinedContent = new StringBuilder();
+                int reset1 = 1;
+                reset1:
                 await foreach (var item in chat.GetStreamingChatMessageContentsAsync(history, settings, documentKernel))
                 {
                     if (!string.IsNullOrEmpty(item.Content))
@@ -379,6 +383,12 @@ 7. ENSURE all enhancements are based on the code files analyzed in the original
                     }
                 }
 
+                if (string.IsNullOrEmpty(docs.Content) && reset1 < 3)
+                {
+                    reset1++;
+                    goto reset1;
+                }
+
                 // 检查精炼后的内容是否有效
                 if (!string.IsNullOrWhiteSpace(refinedContent.ToString()))
                 {
@@ -398,43 +408,10 @@ 7. ENSURE all enhancements are based on the code files analyzed in the original
             }
         }
 
-        // 删除内容中所有的<thinking>内的内容，可能存在多个<thinking>标签,
-        var thinkingRegex = new Regex(@"<thinking>.*?</thinking>", RegexOptions.Singleline);
-        sr = new StringBuilder(thinkingRegex.Replace(sr.ToString(), string.Empty));
-
-
-        // 使用正则表达式将<blog></blog>中的内容提取
-        var regex = new Regex(@"<blog>(.*?)</blog>", RegexOptions.Singleline);
-
-        var match = regex.Match(sr.ToString());
-
-        if (match.Success)
-        {
-            // 提取到的内容
-            var extractedContent = match.Groups[1].Value;
-            sr.Clear();
-            sr.Append(extractedContent);
-        }
-
-        var content = sr.ToString().Trim();
-
-        // 删除所有的所有的<think></think>
-        var thinkRegex = new Regex(@"<think>(.*?)</think>", RegexOptions.Singleline);
-        content = thinkRegex.Replace(content, string.Empty);
-
-        // 从docs提取
-        var docsRegex = new Regex(@"<docs>(.*?)</docs>", RegexOptions.Singleline);
-        var docsMatch = docsRegex.Match(content);
-        if (docsMatch.Success)
-        {
-            // 提取到的内容
-            var extractedDocs = docsMatch.Groups[1].Value;
-            content = content.Replace(docsMatch.Value, extractedDocs);
-        }
 
         var fileItem = new DocumentFileItem()
         {
-            Content = content,
+            Content = docs.Content,
             DocumentCatalogId = catalog.Id,
             Description = string.Empty,
             Extra = new Dictionary<string, string>(),