From 667a55b51cce96863c327ef0978c5af5ff8fbb4a Mon Sep 17 00:00:00 2001
From: curben <2809763-curben@users.noreply.gitlab.com>
Date: Thu, 2 Apr 2020 22:46:28 +0100
Subject: [PATCH] fix: revert to Umbrella list

- Tranco doesn't include subdomain
- Closes #12
- Closes #11
- Closes #10
- Closes #9
- Revert bb9e8ca7cf3009f51290e96ac0731bda8d181c48
---
 script.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/script.sh b/script.sh
index 5b36dc38..17017613 100644
--- a/script.sh
+++ b/script.sh
@@ -8,7 +8,7 @@ mkdir -p "tmp/" && cd "tmp/"
 
 ## Prepare datasets
 curl -L "https://urlhaus.abuse.ch/downloads/csv/" -o "urlhaus.zip"
-curl -L "https://tranco-list.eu/top-1m.csv.zip" -o "top-1m.csv.zip"
+curl -L "https://s3-us-west-1.amazonaws.com/umbrella-static/top-1m.csv.zip" -o "top-1m.csv.zip"
 
 cp "../src/exclude.txt" "."
 
@@ -49,7 +49,7 @@ cut -f 1 -d ":" | \
 sort -u > "urlhaus-domains-online.txt"
 
 
-## Parse the Tranco 1 Million
+## Parse the Umbrella 1 Million
 unzip -p "top-1m.csv.zip" | \
 dos2unix | \
 # Parse domains only