Add XML output format for improved LLM compatibility

Co-authored-by: rohit <[email protected]>

Author: cursoragent

readme.md

@@ -16,6 +16,7 @@ A blazingly fast tool for peeking at codebases. Perfect for loading your codebas
 - 🔗 Web content processing with Markdown conversion
 - 📦 Git repository support
 - 🌐 URL traversal with configurable depth
+- 🏷️ XML output format for better LLM compatibility
 
 ## Installation
 
@@ -115,6 +116,9 @@ glimpse --config_path
 
 # Initialize a .glimpse config file in the current directory
 glimpse --config
+
+# Output in XML format for better LLM compatibility
+glimpse -x /path/to/project
 ```
 
 ## CLI Options
@@ -146,6 +150,7 @@ Options:
       --traverse-links             Traverse links when processing URLs
       --link-depth <DEPTH>         Maximum depth to traverse links (default: 1)
       --pdf <PATH>                 Save output as PDF
+  -x, --xml                        Output in XML format for better LLM compatibility
   -h, --help                       Print help
   -V, --version                    Print version
 ```
@@ -179,6 +184,60 @@ default_excludes = [
 ]
 ```
 
+## XML Output Format
+
+Glimpse supports XML output format designed for better compatibility with Large Language Models (LLMs) like Claude, GPT, and others. When using the `-x` or `--xml` flag, the output is structured with clear XML tags that help LLMs better understand the context and structure of your codebase.
+
+### XML Structure
+
+The XML output wraps all content in a `<context>` tag with the project name:
+
+```xml
+<context name="my_project">
+<tree>
+└── src/
+  └── main.rs
+</tree>
+
+<files>
+<file path="src/main.rs">
+================================================
+fn main() {
+    println!("Hello, World!");
+}
+</file>
+</files>
+
+<summary>
+Total files: 1
+Total size: 45 bytes
+</summary>
+</context>
+```
+
+### Benefits for LLM Usage
+
+- **Clear Context Boundaries**: The `<context>` wrapper helps LLMs understand where your codebase begins and ends
+- **Structured Information**: Separate sections for directory tree, file contents, and summary
+- **Proper Escaping**: XML-safe content that won't confuse parsers
+- **Project Identification**: Automatic project name detection for better context
+
+### Usage Examples
+
+```bash
+# Basic XML output
+glimpse -x /path/to/project
+
+# XML output with file save
+glimpse -x -f project.xml /path/to/project
+
+# XML output to stdout
+glimpse -x --print /path/to/project
+
+# XML output with specific includes
+glimpse -x -i "*.rs,*.py" /path/to/project
+```
+
 ## Token Counting
 
 Glimpse supports two tokenizer backends:

src/analyzer.rs

@@ -64,14 +64,14 @@ pub fn process_directory(args: &Cli) -> Result<()> {
 fn determine_project_name(paths: &[String]) -> String {
     if let Some(first_path) = paths.first() {
         let path = std::path::Path::new(first_path);
-        
+
         // If it's a directory, use its name
         if path.is_dir() {
             if let Some(name) = path.file_name() {
                 return name.to_string_lossy().to_string();
             }
         }
-        
+
         // If it's a file, use the parent directory name
         if path.is_file() {
             if let Some(parent) = path.parent() {
@@ -80,7 +80,7 @@ fn determine_project_name(paths: &[String]) -> String {
                 }
             }
         }
-        
+
         // Fallback to just the path itself
         first_path.clone()
     } else {

src/output.rs

@@ -15,12 +15,20 @@ pub struct FileEntry {
     pub size: u64,
 }
 
-pub fn generate_output(entries: &[FileEntry], format: OutputFormat, xml_format: bool, project_name: Option<String>) -> Result<String> {
+pub fn generate_output(
+    entries: &[FileEntry],
+    format: OutputFormat,
+    xml_format: bool,
+    project_name: Option<String>,
+) -> Result<String> {
     let mut output = String::new();
 
     if xml_format {
         let project_name = project_name.unwrap_or_else(|| "project".to_string());
-        output.push_str(&format!("<context name=\"{}\">\n", xml_escape(&project_name)));
+        output.push_str(&format!(
+            "<context name=\"{}\">\n",
+            xml_escape(&project_name)
+        ));
     }
 
     match format {
@@ -174,7 +182,10 @@ fn generate_files(entries: &[FileEntry], xml_format: bool) -> Result<String> {
 
     for entry in entries {
         if xml_format {
-            output.push_str(&format!("<file path=\"{}\">\n", xml_escape(entry.path.display().to_string().as_str())));
+            output.push_str(&format!(
+                "<file path=\"{}\">\n",
+                xml_escape(entry.path.display().to_string().as_str())
+            ));
             output.push_str(&"=".repeat(48));
             output.push('\n');
             output.push_str(&entry.content);
@@ -379,9 +390,15 @@ mod tests {
     #[test]
     fn test_xml_output() {
         let entries = create_test_entries();
-        
+
         // Test XML tree format
-        let xml_tree_output = generate_output(&entries, OutputFormat::Tree, true, Some("test_project".to_string())).unwrap();
+        let xml_tree_output = generate_output(
+            &entries,
+            OutputFormat::Tree,
+            true,
+            Some("test_project".to_string()),
+        )
+        .unwrap();
         assert!(xml_tree_output.contains("<context name=\"test_project\">"));
         assert!(xml_tree_output.contains("<tree>"));
         assert!(xml_tree_output.contains("</tree>"));
@@ -390,7 +407,13 @@ mod tests {
         assert!(xml_tree_output.contains("</context>"));
 
         // Test XML files format
-        let xml_files_output = generate_output(&entries, OutputFormat::Files, true, Some("test_project".to_string())).unwrap();
+        let xml_files_output = generate_output(
+            &entries,
+            OutputFormat::Files,
+            true,
+            Some("test_project".to_string()),
+        )
+        .unwrap();
         assert!(xml_files_output.contains("<context name=\"test_project\">"));
         assert!(xml_files_output.contains("<files>"));
         assert!(xml_files_output.contains("<file path=\"src/main.rs\">"));
@@ -399,7 +422,13 @@ mod tests {
         assert!(xml_files_output.contains("</context>"));
 
         // Test XML both format
-        let xml_both_output = generate_output(&entries, OutputFormat::Both, true, Some("test_project".to_string())).unwrap();
+        let xml_both_output = generate_output(
+            &entries,
+            OutputFormat::Both,
+            true,
+            Some("test_project".to_string()),
+        )
+        .unwrap();
         assert!(xml_both_output.contains("<context name=\"test_project\">"));
         assert!(xml_both_output.contains("<tree>"));
         assert!(xml_both_output.contains("</tree>"));