cleanup: go module changes

cmd/files.go

@@ -8,11 +8,8 @@ import (
 	"time"
 
 	"github.com/spf13/cobra"
-	"github.com/tnypxl/rollup/internal/config"
 )
 
-var cfg *config.Config
-
 var (
 	path            string
 	fileTypes       string
@@ -27,7 +24,7 @@ var filesCmd = &cobra.Command{
 in a given project, current path or a custom path, to a single timestamped markdown file
 whose name is <project-directory-name>-rollup-<timestamp>.md.`,
 	RunE: func(cmd *cobra.Command, args []string) error {
-		return runRollup(cfg)
+		return runRollup()
 	},
 }
 
@@ -90,23 +87,16 @@ func isIgnored(filePath string, patterns []string) bool {
 				return true
 			}
 		} else {
-			// Check if the pattern matches the full path or any part of it
-			if matched, _ := filepath.Match(pattern, filePath); matched {
+			matched, err := filepath.Match(pattern, filepath.Base(filePath))
+			if err == nil && matched {
 				return true
 			}
-			pathParts := strings.Split(filePath, string(os.PathSeparator))
-			for i := range pathParts {
-				partialPath := filepath.Join(pathParts[:i+1]...)
-				if matched, _ := filepath.Match(pattern, partialPath); matched {
-					return true
-				}
-			}
 		}
 	}
 	return false
 }
 
-func runRollup(cfg *config.Config) error {
+func runRollup() error {
 	// Use config if available, otherwise use command-line flags
 	var types, codeGenList, ignoreList []string
 	if cfg != nil && len(cfg.FileTypes) > 0 {

cmd/files_test.go

@@ -21,9 +21,6 @@ func TestMatchGlob(t *testing.T) {
 		{"**/*.go", "dir/subdir/file.go", true},
 		{"dir/*.go", "dir/file.go", true},
 		{"dir/*.go", "otherdir/file.go", false},
-		{"**/test_*.go", "internal/test_helper.go", true},
-		{"docs/**/*.md", "docs/api/endpoints.md", true},
-		{"docs/**/*.md", "src/docs/readme.md", false},
 	}
 
 	for _, test := range tests {
@@ -35,7 +32,7 @@ func TestMatchGlob(t *testing.T) {
 }
 
 func TestIsCodeGenerated(t *testing.T) {
-	patterns := []string{"generated_*.go", "**/auto_*.go", "**/*_gen.go"}
+	patterns := []string{"generated_*.go", "**/auto_*.go"}
 	tests := []struct {
 		path     string
 		expected bool
@@ -44,8 +41,6 @@ func TestIsCodeGenerated(t *testing.T) {
 		{"normal_file.go", false},
 		{"subdir/auto_file.go", true},
 		{"subdir/normal_file.go", false},
-		{"pkg/models_gen.go", true},
-		{"pkg/handler.go", false},
 	}
 
 	for _, test := range tests {
@@ -57,7 +52,7 @@ func TestIsCodeGenerated(t *testing.T) {
 }
 
 func TestIsIgnored(t *testing.T) {
-	patterns := []string{"*.tmp", "**/*.log", ".git/**", "vendor/**"}
+	patterns := []string{"*.tmp", "**/*.log"}
 	tests := []struct {
 		path     string
 		expected bool
@@ -66,10 +61,6 @@ func TestIsIgnored(t *testing.T) {
 		{"file.go", false},
 		{"subdir/file.log", true},
 		{"subdir/file.txt", false},
-		{".git/config", true},
-		{"src/.git/config", true},
-		{"vendor/package/file.go", true},
-		{"internal/vendor/file.go", false},
 	}
 
 	for _, test := range tests {
@@ -90,56 +81,49 @@ func TestRunRollup(t *testing.T) {
 
 	// Create some test files
 	files := map[string]string{
-		"file1.go":             "package main\n\nfunc main() {}\n",
-		"file2.txt":            "This is a text file.\n",
-		"subdir/file3.go":      "package subdir\n\nfunc Func() {}\n",
-		"subdir/file4.json":    "{\"key\": \"value\"}\n",
-		"generated_model.go":   "// Code generated DO NOT EDIT.\n\npackage model\n",
-		"docs/api/readme.md":   "# API Documentation\n",
-		".git/config":          "[core]\n\trepositoryformatversion = 0\n",
-		"vendor/lib/helper.go": "package lib\n\nfunc Helper() {}\n",
+		"file1.go":          "package main\n\nfunc main() {}\n",
+		"file2.txt":         "This is a text file.\n",
+		"subdir/file3.go":   "package subdir\n\nfunc Func() {}\n",
+		"subdir/file4.json": "{\"key\": \"value\"}\n",
 	}
 
 	for name, content := range files {
 		path := filepath.Join(tempDir, name)
-		if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
+		err := os.MkdirAll(filepath.Dir(path), 0755)
+		if err != nil {
 			t.Fatalf("Failed to create directory: %v", err)
 		}
-		if err := os.WriteFile(path, []byte(content), 0o644); err != nil {
+		err = os.WriteFile(path, []byte(content), 0644)
+		if err != nil {
 			t.Fatalf("Failed to write file: %v", err)
 		}
 	}
 
 	// Set up test configuration
 	cfg = &config.Config{
-		FileTypes:     []string{"go", "txt", "md"},
-		Ignore:        []string{"*.json", ".git/**", "vendor/**"},
-		CodeGenerated: []string{"generated_*.go"},
+		FileTypes:     []string{"go", "txt"},
+		Ignore:        []string{"*.json"},
+		CodeGenerated: []string{},
 	}
-
-	// Change working directory to the temp directory
-	originalWd, _ := os.Getwd()
-	os.Chdir(tempDir)
-	defer os.Chdir(originalWd)
+	path = tempDir
 
 	// Run the rollup
-	if err := runRollup(cfg); err != nil {
+	err = runRollup()
+	if err != nil {
 		t.Fatalf("runRollup() failed: %v", err)
 	}
 
 	// Check if the output file was created
-	outputFiles, err := filepath.Glob("*.rollup.md")
+	outputFiles, err := filepath.Glob(filepath.Join(tempDir, "*.rollup.md"))
 	if err != nil {
-		t.Fatalf("Error globbing for output file: %v", err)
+		t.Fatalf("Failed to glob output files: %v", err)
 	}
-	if len(outputFiles) == 0 {
-		allFiles, _ := filepath.Glob("*")
-		t.Fatalf("No rollup.md file found. Files in directory: %v", allFiles)
+	if len(outputFiles) != 1 {
+		t.Fatalf("Expected 1 output file, got %d", len(outputFiles))
 	}
-	outputFile := outputFiles[0]
 
 	// Read the content of the output file
-	content, err := os.ReadFile(outputFile)
+	content, err := os.ReadFile(outputFiles[0])
 	if err != nil {
 		t.Fatalf("Failed to read output file: %v", err)
 	}
@@ -149,8 +133,6 @@ func TestRunRollup(t *testing.T) {
 		"# File: file1.go",
 		"# File: file2.txt",
 		"# File: subdir/file3.go",
-		"# File: docs/api/readme.md",
-		"# File: generated_model.go (Code-generated, Read-only)",
 	}
 	for _, expected := range expectedContent {
 		if !strings.Contains(string(content), expected) {
@@ -158,15 +140,8 @@ func TestRunRollup(t *testing.T) {
 		}
 	}
 
-	// Check if the ignored files are not included
-	ignoredContent := []string{
-		"file4.json",
-		".git/config",
-		"vendor/lib/helper.go",
-	}
-	for _, ignored := range ignoredContent {
-		if strings.Contains(string(content), ignored) {
-			t.Errorf("Output file contains ignored file: %s", ignored)
-		}
+	// Check if the ignored file is not included
+	if strings.Contains(string(content), "file4.json") {
+		t.Errorf("Output file contains ignored file: file4.json")
 	}
 }

cmd/root.go

@@ -7,6 +7,7 @@ import (
 
 var (
 	configFile string
+	cfg        *config.Config
 	verbose    bool
 )
 
@@ -18,10 +19,10 @@ Use subcommands to perform specific operations.`,
 }
 
 func Execute(conf *config.Config) error {
-	if conf == nil {
-		conf = &config.Config{} // Use an empty config if none is provided
+	cfg = conf
+	if cfg == nil {
+		cfg = &config.Config{} // Use an empty config if none is provided
 	}
-	cfg = conf // Set the cfg variable in cmd/files.go
 	return rootCmd.Execute()
 }
 

cmd/web.go

@@ -64,7 +64,7 @@ func runWeb(cmd *cobra.Command, args []string) error {
                 OutputAlias:      site.OutputAlias,
                 PathOverrides:    convertPathOverrides(site.PathOverrides),
             }
-            logger.Printf("Site %d configuration: BaseURL=%s, CSSLocator=%s, MaxDepth=%d, AllowedPaths=%v",
+            logger.Printf("Site %d configuration: BaseURL=%s, CSSLocator=%s, MaxDepth=%d, AllowedPaths=%v", 
                        i+1, site.BaseURL, site.CSSLocator, site.MaxDepth, site.AllowedPaths)
         }
     } else {
@@ -77,7 +77,7 @@ func runWeb(cmd *cobra.Command, args []string) error {
                 ExcludeSelectors: excludeSelectors,
                 MaxDepth:         depth,
             }
-            logger.Printf("URL %d configuration: BaseURL=%s, CSSLocator=%s, MaxDepth=%d",
+            logger.Printf("URL %d configuration: BaseURL=%s, CSSLocator=%s, MaxDepth=%d", 
                        i+1, u, includeSelector, depth)
         }
     }
@@ -110,7 +110,7 @@ func runWeb(cmd *cobra.Command, args []string) error {
             BurstLimit:        burstLimit,
         },
     }
-    logger.Printf("Scraper configuration: OutputType=%s, RequestsPerSecond=%f, BurstLimit=%d",
+    logger.Printf("Scraper configuration: OutputType=%s, RequestsPerSecond=%f, BurstLimit=%d", 
                outputType, requestsPerSecond, burstLimit)
 
     logger.Println("Starting scraping process")
@@ -131,7 +131,7 @@ func runWeb(cmd *cobra.Command, args []string) error {
 }
 
 func writeSingleFile(content map[string]string) error {
-	outputFile := generateDefaultFilename()
+	outputFile := generateDefaultFilename(urls)
 	file, err := os.Create(outputFile)
 	if err != nil {
 		return fmt.Errorf("error creating output file: %v", err)
@@ -139,7 +139,7 @@ func writeSingleFile(content map[string]string) error {
 	defer file.Close()
 
 	for url, c := range content {
-		_, err = fmt.Fprintf(file, "# Content from %s\n\n%s\n\n---\n\n", url, c)
+		_, err = file.WriteString(fmt.Sprintf("# 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)
 		}
@@ -174,7 +174,7 @@ func writeMultipleFiles(content map[string]string) error {
 	return nil
 }
 
-func generateDefaultFilename() string {
+func generateDefaultFilename(urls []string) string {
 	timestamp := time.Now().Format("20060102-150405")
 	return fmt.Sprintf("web-%s.rollup.md", timestamp)
 }
@@ -261,11 +261,11 @@ func getFilenameFromContent(content, urlStr string) (string, error) {
 	if err != nil {
 		return "", fmt.Errorf("invalid URL: %v", err)
 	}
-
+	
 	if parsedURL.Host == "" {
 		return "", fmt.Errorf("invalid URL: missing host")
 	}
-
+	
 	filename := parsedURL.Host
 	if parsedURL.Path != "" && parsedURL.Path != "/" {
 		filename += strings.TrimSuffix(parsedURL.Path, "/")

internal/scraper/scraper.go

@@ -339,16 +339,6 @@ func ClosePlaywright() {
 	}
 }
 
-// InitBrowser initializes the browser
-func InitBrowser() error {
-	return InitPlaywright()
-}
-
-// CloseBrowser closes the browser
-func CloseBrowser() {
-	ClosePlaywright()
-}
-
 // FetchWebpageContent retrieves the content of a webpage using Playwright
 func FetchWebpageContent(urlStr string) (string, error) {
 	logger.Printf("Fetching webpage content for URL: %s\n", urlStr)
@@ -528,9 +518,7 @@ func ExtractLinks(urlStr string) ([]string, error) {
 
 	var result []string
 	for _, link := range links.([]interface{}) {
-		// Normalize URL by removing trailing slash
-		normalizedLink := strings.TrimRight(link.(string), "/")
-		result = append(result, normalizedLink)
+		result = append(result, link.(string))
 	}
 
 	logger.Printf("Extracted %d links\n", len(result))
@@ -564,23 +552,6 @@ func ExtractContentWithCSS(content, includeSelector string, excludeSelectors []s
 		return "", fmt.Errorf("error extracting content with CSS selector: %v", err)
 	}
 
-	// Trim leading and trailing whitespace
-	selectedContent = strings.TrimSpace(selectedContent)
-
-	// Normalize newlines
-	selectedContent = strings.ReplaceAll(selectedContent, "\r\n", "\n")
-	selectedContent = strings.ReplaceAll(selectedContent, "\r", "\n")
-
-	// Remove indentation while preserving structure
-	lines := strings.Split(selectedContent, "\n")
-	for i, line := range lines {
-		lines[i] = strings.TrimSpace(line)
-	}
-	selectedContent = strings.Join(lines, "\n")
-
-	// Remove any leading or trailing newlines
-	selectedContent = strings.Trim(selectedContent, "\n")
-
 	logger.Printf("Extracted content length: %d\n", len(selectedContent))
 	return selectedContent, nil
 }

internal/scraper/scraper_test.go

@@ -5,9 +5,6 @@ import (
 	"net/http"
 	"net/http/httptest"
 	"strings"
-	"reflect"
-	"log"
-	"io/ioutil"
 )
 
 func TestIsAllowedURL(t *testing.T) {
@@ -72,9 +69,6 @@ func TestGetOverrides(t *testing.T) {
 }
 
 func TestExtractContentWithCSS(t *testing.T) {
-	// Initialize logger for testing
-	logger = log.New(ioutil.Discard, "", 0)
-
 	html := `
 		<html>
 			<body>
@@ -144,12 +138,6 @@ This is a **test** paragraph.
 }
 
 func TestExtractLinks(t *testing.T) {
-	// Initialize Playwright before running the test
-	if err := InitPlaywright(); err != nil {
-		t.Fatalf("Failed to initialize Playwright: %v", err)
-	}
-	defer ClosePlaywright()
-
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		w.Header().Set("Content-Type", "text/html")
 		w.Write([]byte(`