Update parser

scripts/rsw/internal/cmd/page.go

@@ -4,7 +4,7 @@ import (
 	"fmt"
 	"strings"
 
-	"github.com/runescape-wiki/rsw/internal/extract"
+	"github.com/runescape-wiki/rsw/internal/htmlconv"
 	"github.com/runescape-wiki/rsw/internal/render"
 	"github.com/runescape-wiki/rsw/internal/wiki"
 	"github.com/spf13/cobra"
@@ -27,80 +27,54 @@ Examples:
 			title := args[0]
 			client := wiki.NewClient(GameBaseURL())
 
-			page, err := client.GetPage(title)
+			if Raw() {
+				page, err := client.GetPage(title)
+				if err != nil {
+					return fmt.Errorf("failed to fetch page: %w", err)
+				}
+				fmt.Println(page.Wikitext)
+				return nil
+			}
+
+			page, err := client.GetPageHTML(title)
 			if err != nil {
 				return fmt.Errorf("failed to fetch page: %w", err)
 			}
 
-			wikitext := page.Wikitext
-
 			if pageSection != "" {
-				idx := wiki.FindSectionIndex(page.Sections, pageSection)
-				if idx == -1 {
-					needle := strings.ToLower(pageSection)
-					// Case-insensitive exact match
-					for _, s := range page.Sections {
-						if strings.ToLower(s.Line) == needle {
-							fmt.Sscanf(s.Index, "%d", &idx)
-							break
-						}
-					}
-					// Case-insensitive prefix match (e.g. "Location" → "Locations")
-					if idx == -1 {
-						for _, s := range page.Sections {
-							if strings.HasPrefix(strings.ToLower(s.Line), needle) {
-								fmt.Sscanf(s.Index, "%d", &idx)
-								break
-							}
-						}
-					}
-					// Case-insensitive contains match
-					if idx == -1 {
-						for _, s := range page.Sections {
-							if strings.Contains(strings.ToLower(s.Line), needle) {
-								fmt.Sscanf(s.Index, "%d", &idx)
-								break
-							}
-						}
-					}
-				}
-				if idx == -1 {
+				body := htmlconv.ExtractSection(page.HTML, pageSection)
+				if body == "" {
+					sections := htmlconv.ListSections(page.HTML)
 					return fmt.Errorf("section %q not found. Available sections: %s",
-						pageSection, listSections(page.Sections))
+						pageSection, formatSectionNames(sections))
 				}
-				sectionPage, err := client.GetPageSection(title, idx)
-				if err != nil {
-					return fmt.Errorf("failed to fetch section: %w", err)
-				}
-				wikitext = sectionPage.Wikitext
-			}
-
-			if Raw() {
-				fmt.Println(wikitext)
+				md := render.New()
+				md.H1(page.Title)
+				md.Line(body)
+				fmt.Print(md.String())
 				return nil
 			}
 
 			md := render.New()
 			md.H1(page.Title)
 
-			if pageSection == "" && len(page.Sections) > 0 {
+			sections := htmlconv.ListSections(page.HTML)
+			if len(sections) > 0 {
 				md.H2("Sections")
-				for _, s := range page.Sections {
+				for _, s := range sections {
 					indent := ""
-					if s.Level == "3" {
+					if s.Level == 3 {
 						indent = "  "
-					} else if s.Level == "4" {
+					} else if s.Level >= 4 {
 						indent = "    "
 					}
-					md.Line(fmt.Sprintf("%s- %s", indent, s.Line))
+					md.Line(fmt.Sprintf("%s- %s", indent, s.Name))
 				}
 				md.Newline()
 				md.HR()
 			}
 
-			plain := extract.ExtractPlainText(wikitext)
-			md.P(plain)
-
+			md.Line(htmlconv.Convert(page.HTML))
 			fmt.Print(md.String())
 			return nil
 		},
@@ -110,10 +84,10 @@ Examples:
 	return cmd
 }
 
-func listSections(sections []wiki.Section) string {
+func formatSectionNames(sections []htmlconv.SectionInfo) string {
 	names := make([]string, len(sections))
 	for i, s := range sections {
-		names[i] = s.Line
+		names[i] = s.Name
 	}
 	return strings.Join(names, ", ")
 }