Merge branch 'claude/fix-bugs-and-gaps-01DvJSzruQh49DU6XK5AykQU' (#4)

cmd/root.go

@@ -1,8 +1,10 @@
 package cmd
 
 import (
+	"log"
+
 	"github.com/spf13/cobra"
-	config "github.com/tnypxl/rollup/internal/config"
+	"github.com/tnypxl/rollup/internal/config"
 )
 
 var (
@@ -15,13 +17,31 @@ var rootCmd = &cobra.Command{
 	Short: "Rollup is a tool for combining and processing files",
 	Long: `Rollup is a versatile tool that can combine and process files in various ways.
 Use subcommands to perform specific operations.`,
+	PersistentPreRunE: func(cmd *cobra.Command, args []string) error {
+		// Skip config loading for generate and help commands
+		if cmd.Name() == "generate" || cmd.Name() == "help" {
+			return nil
+		}
+
+		// Determine config path
+		configPath := configFile
+		if configPath == "" {
+			configPath = "rollup.yml"
+		}
+
+		// Load configuration
+		var err error
+		cfg, err = config.Load(configPath)
+		if err != nil {
+			log.Printf("Warning: Failed to load configuration from %s: %v", configPath, err)
+			cfg = &config.Config{} // Use empty config if loading fails
+		}
+
+		return nil
+	},
 }
 
-func Execute(conf *config.Config) error {
-	if conf == nil {
-		conf = &config.Config{} // Use an empty config if none is provided
-	}
-	cfg = conf // Set the cfg variable in cmd/files.go
+func Execute() error {
 	return rootCmd.Execute()
 }
 

cmd/web.go

@@ -28,7 +28,19 @@ var webCmd = &cobra.Command{
 	Use:   "web",
 	Short: "Scrape main content from webpages and convert to Markdown",
 	Long:  `Scrape the main content from one or more webpages, ignoring navigational elements, ads, and other UI aspects. Convert the content to a well-structured Markdown file.`,
-	RunE:  runWeb,
+	PreRunE: func(cmd *cobra.Command, args []string) error {
+		// Initialize Playwright for web scraping
+		if err := scraper.InitPlaywright(); err != nil {
+			return fmt.Errorf("failed to initialize Playwright: %w", err)
+		}
+		return nil
+	},
+	RunE: runWeb,
+	PostRunE: func(cmd *cobra.Command, args []string) error {
+		// Clean up Playwright resources
+		scraper.ClosePlaywright()
+		return nil
+	},
 }
 
 func init() {
@@ -144,95 +156,6 @@ func runWeb(cmd *cobra.Command, args []string) error {
 	return nil
 }
 
-func writeSingleFile(content map[string]string) error {
-	outputFile := generateDefaultFilename()
-	file, err := os.Create(outputFile)
-	if err != nil {
-		return fmt.Errorf("error creating output file: %v", err)
-	}
-	defer file.Close()
-
-	for url, c := range content {
-		_, err = fmt.Fprintf(file, "# ::: Content from %s\n\n%s\n\n---\n\n", url, c)
-		if err != nil {
-			return fmt.Errorf("error writing content to file: %v", err)
-		}
-	}
-
-	fmt.Printf("Content has been extracted from %d URL(s) and saved to %s\n", len(content), outputFile)
-	return nil
-}
-
-func writeMultipleFiles(content map[string]string) error {
-	for url, c := range content {
-		filename, err := getFilenameFromContent(c, url)
-		if err != nil {
-			return fmt.Errorf("error generating filename for %s: %v", url, err)
-		}
-
-		file, err := os.Create(filename)
-		if err != nil {
-			return fmt.Errorf("error creating output file %s: %v", filename, err)
-		}
-
-		_, err = file.WriteString(fmt.Sprintf("# ::: Content from %s\n\n%s\n", url, c))
-		if err != nil {
-			file.Close()
-			return fmt.Errorf("error writing content to file %s: %v", filename, err)
-		}
-
-		file.Close()
-		fmt.Printf("Content from %s has been saved to %s\n", url, filename)
-	}
-
-	return nil
-}
-
-func generateDefaultFilename() string {
-	timestamp := time.Now().Format("20060102-150405")
-	return fmt.Sprintf("web-%s.rollup.md", timestamp)
-}
-
-func scrapeURL(urlStr string) (string, error) {
-	content, err := testExtractAndConvertContent(urlStr)
-	if err != nil {
-		return "", err
-	}
-
-	return content, nil
-}
-
-var (
-	testExtractAndConvertContent = extractAndConvertContent
-)
-
-func extractAndConvertContent(urlStr string) (string, error) {
-	content, err := scraper.FetchWebpageContent(urlStr)
-	if err != nil {
-		return "", fmt.Errorf("error fetching webpage content: %v", err)
-	}
-
-	if includeSelector != "" {
-		content, err = scraper.ExtractContentWithCSS(content, includeSelector, excludeSelectors)
-		if err != nil {
-			return "", fmt.Errorf("error extracting content with CSS: %v", err)
-		}
-	}
-
-	markdown, err := scraper.ProcessHTMLContent(content, scraper.Config{})
-	if err != nil {
-		return "", fmt.Errorf("error processing HTML content: %v", err)
-	}
-
-	parsedURL, err := url.Parse(urlStr)
-	if err != nil {
-		return "", fmt.Errorf("error parsing URL: %v", err)
-	}
-	header := fmt.Sprintf("# ::: Content from %s\n\n", parsedURL.String())
-
-	return header + markdown + "\n\n", nil
-}
-
 func getFilenameFromContent(content, urlStr string) (string, error) {
 	// Try to extract title from content
 	titleStart := strings.Index(content, "<title>")

cmd/web_test.go

@@ -97,8 +97,3 @@ func TestGetFilenameFromContent(t *testing.T) {
 		}
 	}
 }
-
-// Mock functions for testing
-func mockExtractAndConvertContent(urlStr string) (string, error) {
-	return "Mocked content for " + urlStr, nil
-}

internal/config/config.go

@@ -92,6 +92,10 @@ func (c *Config) Validate() error {
 		return fmt.Errorf("file_extensions or sites must be specified")
 	}
 
+	if c.OutputType != "" && c.OutputType != "single" && c.OutputType != "separate" {
+		return fmt.Errorf("output_type must be 'single' or 'separate'")
+	}
+
 	if c.RequestsPerSecond != nil && *c.RequestsPerSecond <= 0 {
 		return fmt.Errorf("requests_per_second must be positive")
 	}

internal/config/config_test.go

@@ -136,6 +136,30 @@ func TestValidate(t *testing.T) {
 			},
 			wantErr: true,
 		},
+		{
+			name: "Valid output type single",
+			config: Config{
+				FileExtensions: []string{"go"},
+				OutputType:     "single",
+			},
+			wantErr: false,
+		},
+		{
+			name: "Valid output type separate",
+			config: Config{
+				FileExtensions: []string{"go"},
+				OutputType:     "separate",
+			},
+			wantErr: false,
+		},
+		{
+			name: "Invalid output type",
+			config: Config{
+				FileExtensions: []string{"go"},
+				OutputType:     "invalid",
+			},
+			wantErr: true,
+		},
 	}
 
 	for _, tt := range tests {

main.go

@@ -2,40 +2,13 @@ package main
 
 import (
 	"fmt"
-	"log"
 	"os"
 
 	"github.com/tnypxl/rollup/cmd"
-	"github.com/tnypxl/rollup/internal/config"
-	"github.com/tnypxl/rollup/internal/scraper"
 )
 
 func main() {
-	// Check if the command is "help"
-	isHelpCommand := len(os.Args) > 1 && (os.Args[1] == "help" || os.Args[1] == "--help" || os.Args[1] == "-h")
-
-	var cfg *config.Config
-	var err error
-
-	if !isHelpCommand {
-		configPath := "rollup.yml"
-		cfg, err = config.Load(configPath)
-		if err != nil {
-			log.Printf("Warning: Failed to load configuration: %v", err)
-			// Continue execution without a config file
-		}
-
-		// Initialize the scraper logger with default verbosity (false)
-		scraper.SetupLogger(false)
-
-		err = scraper.InitPlaywright()
-		if err != nil {
-			log.Fatalf("Failed to initialize Playwright: %v", err)
-		}
-		defer scraper.ClosePlaywright()
-	}
-
-	if err := cmd.Execute(cfg); err != nil {
+	if err := cmd.Execute(); err != nil {
 		fmt.Println(err)
 		os.Exit(1)
 	}