print improvements, fix unicode char parsing

parse.go

@@ -5,7 +5,6 @@ import (
 	"fmt"
 	"github.com/PuerkitoBio/goquery"
 	"io/ioutil"
-	"path/filepath"
 	"strings"
 )
 
@@ -19,7 +18,7 @@ func parse(dir, pathPrefix string) []Link {
 
 	// parse md
 	var links []Link
-	fmt.Printf("[Parsing note] %s\n", trim(dir, pathPrefix, ".md"))
+	fmt.Printf("[Parsing note] %s => ", trim(dir, pathPrefix, ".md"))
 
 	var buf bytes.Buffer
 	if err := md.Convert(source, &buf); err != nil {
@@ -35,22 +34,18 @@ func parse(dir, pathPrefix string) []Link {
 			target = "#"
 		}
 
-		target = strings.Replace(target, "%20", " ", -1)
-		target = strings.Split(processTarget(target), "#")[0]
-		target = strings.TrimSpace(target)
-		target = strings.Replace(target, " ", "-", -1)
+		target = processTarget(target)
+		source := processSource(trim(dir, pathPrefix, ".md"))
 
-		source := filepath.ToSlash(hugoPathTrim(trim(dir, pathPrefix, ".md")))
-
-		fmt.Printf("  '%s' => %s\n", text, target)
+		// fmt.Printf("  '%s' => %s\n", source, target)
 		links = append(links, Link{
-			Source: UnicodeSanitize(source),
-			Target: UnicodeSanitize(target),
+			Source: source,
+			Target: target,
 			Text:   text,
 		})
 		n++
 	})
-	fmt.Printf("  Found: %d links\n", n)
+	fmt.Printf("found: %d links\n", n)
 
 	return links
 }

util.go

@@ -2,6 +2,7 @@ package main
 
 import (
 	"fmt"
+	"net/url"
 	"path/filepath"
 	"strings"
 	"unicode"
@@ -22,7 +23,16 @@ func processTarget(source string) string {
 	if strings.HasPrefix(source, "/") {
 		return strings.TrimSuffix(source, ".md")
 	}
-	return "/" + strings.TrimSuffix(strings.TrimSuffix(source, ".html"), ".md")
+	res := "/" + strings.TrimSuffix(strings.TrimSuffix(source, ".html"), ".md")
+	res = strings.Split(res, "#")[0]
+	res = strings.TrimSpace(res)
+	return UnicodeSanitize(res)
+}
+
+func processSource(source string) string {
+	res := filepath.ToSlash(hugoPathTrim(source))
+	res = UnicodeSanitize(res)
+	return strings.ReplaceAll(url.PathEscape(res), "%2F", "/")
 }
 
 func isInternal(link string) bool {
@@ -80,4 +90,3 @@ func filter(links []Link) (res []Link) {
 	fmt.Printf("Removed %d external and non-markdown links\n", len(links)-len(res))
 	return res
 }
-

walk.go

@@ -8,6 +8,7 @@ import (
 	"os"
 	"path/filepath"
 	"strings"
+	"time"
 )
 
 // recursively walk directory and return all files with given extension
@@ -19,6 +20,8 @@ func walk(root, ext string, index bool, ignorePaths map[string]struct{}) (res []
 	m.Handle("---", front.YAMLHandler)
 	nPrivate := 0
 
+	start := time.Now()
+
 	err := filepath.WalkDir(root, func(fp string, d fs.DirEntry, e error) error {
 		if e != nil {
 			return e
@@ -50,8 +53,10 @@ func walk(root, ext string, index bool, ignorePaths map[string]struct{}) (res []
 				// check if page is private
 				if parsedPrivate, ok := frontmatter["draft"]; !ok || !parsedPrivate.(bool) {
 					info, _ := os.Stat(s)
-					adjustedPath := UnicodeSanitize(strings.Replace(hugoPathTrim(trim(s, root, ".md")), " ", "-", -1))
-					i[adjustedPath] = Content{
+					source := processSource(trim(s, root, ".md"))
+
+					// adjustedPath := UnicodeSanitize(strings.Replace(hugoPathTrim(trim(s, root, ".md")), " ", "-", -1))
+					i[source] = Content{
 						LastModified: info.ModTime(),
 						Title:        title,
 						Content:      body,
@@ -67,6 +72,10 @@ func walk(root, ext string, index bool, ignorePaths map[string]struct{}) (res []
 	if err != nil {
 		panic(err)
 	}
+
+	end := time.Now()
+
+	fmt.Printf("[DONE] in %s\n", end.Sub(start).Round(time.Millisecond))
 	fmt.Printf("Ignored %d private files \n", nPrivate)
 	fmt.Printf("Parsed %d total links \n", len(res))
 	return res, i