From afe84f6129348b493323d9fec429ad6c6c50584d Mon Sep 17 00:00:00 2001
From: Alex Lovell-Troy <alex@lovelltroy.org>
Date: Tue, 2 Jul 2024 16:37:39 -0400
Subject: [PATCH] chore: Clean up URI handling in crawl command

---
 cmd/crawl.go | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/cmd/crawl.go b/cmd/crawl.go
index 87f093b..84fb24e 100644
--- a/cmd/crawl.go
+++ b/cmd/crawl.go
@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"log"
 	"net/url"
+	"strings"
 
 	"github.com/OpenCHAMI/magellan/pkg/crawler"
 	"github.com/spf13/cobra"
@@ -18,10 +19,16 @@ var crawlCmd = &cobra.Command{
 		if err := cobra.ExactArgs(1)(cmd, args); err != nil {
 			return err
 		}
-		_, err := url.ParseRequestURI(args[0])
+		parsedURI, err := url.ParseRequestURI(args[0])
 		if err != nil {
 			return fmt.Errorf("invalid URI specified: %s", args[0])
 		}
+		// Remove any trailing slashes
+		parsedURI.Path = strings.TrimSuffix(parsedURI.Path, "/")
+		// Collapse any doubled slashes
+		parsedURI.Path = strings.ReplaceAll(parsedURI.Path, "//", "/")
+		// Update the URI in the args slice
+		args[0] = parsedURI.String()
 		return nil
 	},
 	Run: func(cmd *cobra.Command, args []string) {