added archive, megamillions, and powerball site logic

cmd/browser/main.go

@@ -0,0 +1,79 @@
+package main
+
+import (
+	"context"
+	"fmt"
+	"io"
+	"os"
+
+	"github.com/urfave/cli/v3"
+
+	"gitea.stevedudenhoeffer.com/steve/go-extractor"
+	"gitea.stevedudenhoeffer.com/steve/go-extractor/cmd/browser/pkg/browser"
+)
+
+func deferClose(cl io.Closer) {
+	_ = cl.Close()
+}
+func main() {
+	cmd := &cli.Command{
+		Name:  "browser",
+		Flags: browser.Flags,
+		Usage: "<url>",
+		Action: func(ctx context.Context, cli *cli.Command) error {
+			target := cli.Args().First()
+			if target == "" {
+				return fmt.Errorf("no url specified")
+			}
+
+			b, err := browser.FromCommand(ctx, cli)
+			if err != nil {
+				return err
+			}
+
+			defer deferClose(b)
+
+			// now open the user specified url
+			doc, err := b.Open(ctx, target, extractor.OpenPageOptions{})
+			if err != nil {
+				return err
+			}
+
+			defer deferClose(doc)
+
+			article, err := extractor.Readability(ctx, doc)
+
+			if err != nil {
+				return err
+			}
+
+			content := ""
+
+			if article.Content != "" {
+
+				if len(article.Content) > 32 {
+					content = article.Content[:32] + "..."
+				} else {
+					content = article.Content
+				}
+			}
+
+			fmt.Println("Title:", article.Title)
+			fmt.Println("Byline:", article.Byline)
+			fmt.Println("Site:", article.SiteName)
+			fmt.Println("Published:", article.PublishedTime)
+			fmt.Println("Excerpt:", article.Excerpt)
+			fmt.Println("Length:", article.Length)
+			fmt.Println("Lang:", article.Lang)
+			fmt.Println("Content:", content)
+			fmt.Println("TextContent:", article.TextContent)
+			return nil
+		},
+	}
+
+	err := cmd.Run(context.Background(), os.Args)
+
+	if err != nil {
+		panic(err)
+	}
+}