add alexa list about daily

[pexcn]

Commands

curl -kLs https://s3.amazonaws.com/alexa-static/top-1m.csv.zip | gunzip > top-1m.csv

# top 1m in world
awk -F ',' '{print $2}' top-1m.csv > top-1m.txt

# top list in china
grep -Fx -f chinalist.txt top-1m.txt > top-cn.txt

# top 5k in china
head -5000 top-cn.txt > top-cn-5000.txt

# top 1k in china
head -1000 top-cn.txt > top-cn-1000.txt

# top 500 in china
head -500 top-cn.txt > top-cn-500.txt

[wyf88]

It seems that the Alexa top list is now restricted to paid customers and the link is no longer updated and incomplete (source).

The Cisco Umbrella Popularity List takes DNS query into account and reflects the most commonly connected domains (including website resources) that are not restricted to those "main domains".

I have conducted some testings and found that the Umbrella list has more common items compared with Alexa list with both gfwlist and chinalist.

• Umbrella: 1661 with gfwlist, 8562 with chinalist
• Alexa: 1444 with gfwlist, 3029 with chinalist

In addition, domains that are only in Umbrella list look much more common than those only in Alexa list. Therefore, I would suggest to replace Alexa list with the Umbrella list. This could better serve the purpose of generating a shorter chinalist and/or gfwlist.

[pexcn]

It seems that the Alexa top list is now restricted to paid customers and the link is no longer updated and incomplete (source).

https://s3.amazonaws.com/alexa-static/top-1m.csv.zip has never stopped updating, and Umbrella list contains too many subdomains, so it is not applicable here.

[wyf88]

Umbrella list contains too many subdomains, so it is not applicable here.

Actually that is one advantage of the Umbrella list, because both the gfwlist and the accelerated-domains.china.conf contain a large number of subdomains, e.g. only a subdomain is blocked, or has server in CN. Therefore compared with Alexa toplist, the Umbrella list can actually match more domains in gfwlist and accelerated-domains.china.conf.

The other thing is that the Alexa list mainly focuses on the main domain of a website and ignores (or puts much less weight on) domains serving static files for that website.

Here are some test chinalist sorted by Umbrella toplist, sorted by Alexa toplist, and sorted by both, generated by today. As you can see, the Umbrella list has 3231 matches with the accelerated-domains.china.conf, whereas the Alexa list has 2642 matches, and there are 5038 matches if both Umbrella and Alexa are used.

Personally I'm using the Umbrella list only and the generated gfwlist and chinalist work well (BTW the matches can be used to generate a short chinalist for mobile apps). But maybe it is worth to using both Umbrella and Alexa? Just for your consideration.

[pexcn]

I'm sorry, I haven't touched this project for too long, I know almost nothing about its logic now.
Maybe I will deal with it later if I have time.

[pexcn]

Reopen.