Add the option to output to XML (#11)

* Add the option to output to XML

* feat: add XML validation and custom marshaling for GitRepo output

Author: imertz

README.md

@@ -34,6 +34,7 @@ By default, your `.git` directory and your `.gitignore` files are ignored. Any f
 * `-o`,  `--output`: Path to the output file. If not specified, will print to standard output.
 * `-e`,  `--estimate`: Estimate the tokens of the output file. If not specified, does not estimate. 
 * `-j`,  `--json`: Output to JSON rather than plain text. Use with `-o` to specify the output file.
+* `-x`,  `--xml`: Output to XML rather than plain text. Use with `-o` to specify the output file.
 * `-i`,  `--ignore`: Path to the `.gptignore` file. If not specified, will look for a `.gptignore` file in the same directory as the `.gitignore` file.
 * `-g`,  `--ignore-gitignore`: Ignore the `.gitignore` file.
 * `-s`,  `--scrub-comments`: Remove comments from the output file to save tokens.

cmd/root.go

@@ -15,6 +15,7 @@ var estimateTokens bool
 var ignoreFilePath string
 var ignoreGitignore bool
 var outputJSON bool
+var outputXML bool
 var debug bool
 var scrubComments bool
 
@@ -54,6 +55,37 @@ var rootCmd = &cobra.Command{
 			}
 			return
 		}
+		if outputXML {
+			output, err := prompt.OutputGitRepoXML(repo, scrubComments)
+			if err != nil {
+				fmt.Printf("Error: %s\n", err)
+				os.Exit(1)
+			}
+			
+			// Validate the XML output
+			if err := prompt.ValidateXML(output); err != nil {
+				fmt.Printf("Error: %s\n", err)
+				os.Exit(1)
+			}
+			
+			if outputFile != "" {
+				// if output file exists, throw error
+				if _, err := os.Stat(outputFile); err == nil {
+					fmt.Printf("Error: output file %s already exists\n", outputFile)
+					os.Exit(1)
+				}
+				err = os.WriteFile(outputFile, []byte(output), 0644)
+				if err != nil {
+					fmt.Printf("Error: could not write to output file %s\n", outputFile)
+					os.Exit(1)
+				}
+			} else {
+				if !debug {
+					fmt.Println(output)
+				}
+			}
+			return
+		}
 		output, err := prompt.OutputGitRepo(repo, preambleFile, scrubComments)
 		if err != nil {
 			fmt.Printf("Error: %s\n", err)
@@ -93,6 +125,8 @@ func init() {
 	rootCmd.Flags().BoolVarP(&ignoreGitignore, "ignore-gitignore", "g", false, "ignore .gitignore file")
 	// output JSON. Should be a bool
 	rootCmd.Flags().BoolVarP(&outputJSON, "json", "j", false, "output JSON")
+	// output XML. Should be a bool
+	rootCmd.Flags().BoolVarP(&outputXML, "xml", "x", false, "output XML")
 	// debug. Should be a bool
 	rootCmd.Flags().BoolVarP(&debug, "debug", "d", false, "debug mode. Do not output to standard output")
 	// scrub comments. Should be a bool

prompt/prompt.go

@@ -3,7 +3,9 @@ package prompt
 import (
 	"bufio"
 	"encoding/json"
+	"encoding/xml"
 	"fmt"
+	"io"
 	"os"
 	"path/filepath"
 	"strings"
@@ -16,16 +18,16 @@ import (
 
 // GitFile is a file in a Git repository
 type GitFile struct {
-	Path     string `json:"path"`     // path to the file relative to the repository root
-	Tokens   int64  `json:"tokens"`   // number of tokens in the file
-	Contents string `json:"contents"` // contents of the file
+	Path     string `json:"path" xml:"path"`     // path to the file relative to the repository root
+	Tokens   int64  `json:"tokens" xml:"tokens"`   // number of tokens in the file
+	Contents string `json:"contents" xml:"contents"` // contents of the file
 }
 
 // GitRepo is a Git repository
 type GitRepo struct {
-	TotalTokens int64     `json:"total_tokens"`
-	Files       []GitFile `json:"files"`
-	FileCount   int       `json:"file_count"`
+	TotalTokens int64     `json:"total_tokens" xml:"total_tokens"`
+	Files       []GitFile `json:"files" xml:"files>file"`
+	FileCount   int       `json:"file_count" xml:"file_count"`
 }
 
 // contains checks if a string is in a slice of strings
@@ -164,6 +166,80 @@ func OutputGitRepo(repo *GitRepo, preambleFile string, scrubComments bool) (stri
 	return output, nil
 }
 
+func OutputGitRepoXML(repo *GitRepo, scrubComments bool) (string, error) {
+    // Prepare XML content
+    if scrubComments {
+        for i, file := range repo.Files {
+            repo.Files[i].Contents = utils.RemoveCodeComments(file.Contents)
+        }
+    }
+    
+    // Add XML header
+    var result strings.Builder
+    result.WriteString("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n")
+    
+    // Use custom marshaling with proper CDATA for code contents
+    result.WriteString("<GitRepo>\n")
+    
+    // Skip the tokens for now
+    result.WriteString("  <total_tokens>PLACEHOLDER</total_tokens>\n")
+    result.WriteString(fmt.Sprintf("  <file_count>%d</file_count>\n", repo.FileCount))
+    result.WriteString("  <files>\n")
+    
+    for _, file := range repo.Files {
+        result.WriteString("    <file>\n")
+        result.WriteString(fmt.Sprintf("      <path>%s</path>\n", escapeXML(file.Path)))
+        result.WriteString(fmt.Sprintf("      <tokens>%d</tokens>\n", file.Tokens))
+        result.WriteString("      <contents><![CDATA[")
+        result.WriteString(file.Contents)
+        result.WriteString("]]></contents>\n")
+        result.WriteString("    </file>\n")
+    }
+    
+    result.WriteString("  </files>\n")
+    result.WriteString("</GitRepo>")
+    
+    // Get the output string
+    outputStr := result.String()
+    
+    // Calculate tokens
+    tokenCount := EstimateTokens(outputStr)
+    repo.TotalTokens = tokenCount
+    
+    // Replace the placeholder with the actual token count
+    outputStr = strings.Replace(outputStr, "<total_tokens>PLACEHOLDER</total_tokens>", 
+                               fmt.Sprintf("<total_tokens>%d</total_tokens>", tokenCount), 1)
+    
+    return outputStr, nil
+}
+
+// escapeXML escapes XML special characters in a string
+func escapeXML(s string) string {
+    s = strings.ReplaceAll(s, "&", "&amp;")
+    s = strings.ReplaceAll(s, "<", "&lt;")
+    s = strings.ReplaceAll(s, ">", "&gt;")
+    s = strings.ReplaceAll(s, "\"", "&quot;")
+    s = strings.ReplaceAll(s, "'", "&apos;")
+    return s
+}
+
+// ValidateXML checks if the given XML string is well-formed
+func ValidateXML(xmlString string) error {
+    decoder := xml.NewDecoder(strings.NewReader(xmlString))
+    for {
+        _, err := decoder.Token()
+        if err == io.EOF {
+            break
+        }
+        if err != nil {
+            return fmt.Errorf("XML validation error: %w", err)
+        }
+    }
+    return nil
+}
+
+
+
 func MarshalRepo(repo *GitRepo, scrubComments bool) ([]byte, error) {
 	// run the output function to get the total tokens
 	_, err := OutputGitRepo(repo, "", scrubComments)