From 234773b8f0e3a5f4eb005b7ff1c6479fd9dcb77d Mon Sep 17 00:00:00 2001
From: setop <setop@zoocoop.com>
Date: Fri, 25 Nov 2022 18:10:45 +0100
Subject: [PATCH] build graph from mobilizon instances and followers

---
 .tool-versions          |  1 +
 cmd.txt                 | 17 +++++++++++++
 fedcrawler_followers.py | 53 +++++++++++++++++++++++++++++++++++++++++
 template.dot            | 13 ++++++++++
 4 files changed, 84 insertions(+)
 create mode 100644 .tool-versions
 create mode 100644 cmd.txt
 create mode 100644 fedcrawler_followers.py
 create mode 100644 template.dot

diff --git a/.tool-versions b/.tool-versions
new file mode 100644
index 0000000..c6e5191
--- /dev/null
+++ b/.tool-versions
@@ -0,0 +1 @@
+python 3.9.9
diff --git a/cmd.txt b/cmd.txt
new file mode 100644
index 0000000..8a78972
--- /dev/null
+++ b/cmd.txt
@@ -0,0 +1,17 @@
+. .venv/bin/activate
+
+pip install scrapy
+
+wget https://framagit.org/-/snippets/6539/raw/main/pp.awk
+
+curl -fsS 'https://instances.joinmobilizon.org/api/v1/instances?start=0&count=1000' | jq -r '.data[].host' instances.json > instances.txt
+
+scrapy runspider -o out.csv:csv fedcrawler_followers.py
+
+dos2unix out.csv
+
+uniq0 out.csv >| out.u.csv
+
+bash <(awk -f pp.awk template.dot) < out.u.csv >| out.dot
+
+neato -Tsvg -o out.svg out.dot
diff --git a/fedcrawler_followers.py b/fedcrawler_followers.py
new file mode 100644
index 0000000..c14a4d6
--- /dev/null
+++ b/fedcrawler_followers.py
@@ -0,0 +1,53 @@
+# -*- coding: utf-8 -*-
+import logging
+
+from urllib.parse import urlparse
+
+from scrapy.spiders import Spider
+from scrapy import Request
+
+
+logger = logging.getLogger(__name__)
+
+
+
+class Fedator(Spider):
+	name = 'fedator'
+	custom_settings = {
+		"USER_AGENT" : 'Mozilla/5.0 (X11; Linux x86_64; rv:100.0) Gecko/20100101 Firefox/100.0',
+		"ROBOTSTXT_OBEY" : False,
+		"REFERER_ENABLED" : False,
+		"COOKIES_ENABLED" : False,
+		"TELNETCONSOLE_ENABLED" : False,
+		"HTTPCACHE_ENABLED" : True,
+		"DEFAULT_REQUEST_HEADERS" : {
+			'Accept': 'application/activity+json',
+		},
+	}
+	
+	MAX_DEPTH = 10
+
+	visited = set()
+	
+	def  start_requests(self):
+		for line in open("instances.txt").readlines():
+			host = line[:-1]
+			self.visited.add(host)
+			yield Request(f'https://{host}/@relay/followers?page=1', meta= {"dst":host, "depth" : 0, "page":1 })
+		
+
+	def parse(self, response):
+		followers = response.json()["orderedItems"]  # "https://mobilizon.sans-nuage.fr/relay"
+		if len(followers)>0:
+			dst = response.request.meta["dst"]
+			page = response.request.meta["page"] + 1
+			yield Request(f'https://{dst}/@relay/followers?page={page}', meta= {"dst":dst, "depth" : 0, "page": page})
+			depth = response.request.meta["depth"]
+			for follower in followers:
+				host = follower.split("/")[2]
+				yield {"src":  host, "dst" : dst }
+				if host not in self.visited and depth < self.MAX_DEPTH:
+					self.visited.add(host)
+					yield Request(f'https://{host}/@relay/followers?page=1', meta= {"dst":host, "depth" : depth+1, "page":1  })
+				else:
+					logger.debug(f"already visited of maxdepth ({depth+1})")
diff --git a/template.dot b/template.dot
new file mode 100644
index 0000000..8c406c2
--- /dev/null
+++ b/template.dot
@@ -0,0 +1,13 @@
+digraph fediverse {
+graph [overlap=false]
+#!
+IFS=","
+while read src dst
+do
+	if [ ! $src == "src" ];
+	then
+		echo "\"${src}\" -> \"${dst}\""
+	fi
+done
+#!
+}
\ No newline at end of file