-from requests import get
-from hashlib import sha256
-import sqlite3
-from bs4 import BeautifulSoup
-
-headers = {
- "user-agent": "fedi-block-api (https://gitlab.com/EnjuAihara/fedi-block-api)"
-}
-
-
-def get_mastodon_blocks(domain: str) -> dict:
- blocks = {
- "Suspended servers": [],
- "Filtered media": [],
- "Limited servers": [],
- "Silenced servers": [],
- }
-
- translations = {
- "Gesperrte Server": "Suspended servers",
- "Gefilterte Medien": "Filtered media",
- "Stummgeschaltete Server": "Silenced servers",
- "停止済みのサーバー": "Suspended servers",
- "メディアを拒否しているサーバー": "Filtered media",
- "サイレンス済みのサーバー": "Silenced servers",
- "Serveurs suspendus": "Suspended servers",
- "Médias filtrés": "Filtered media",
- "Serveurs limités": "Silenced servers",
- }
-
- try:
- doc = BeautifulSoup(
- get(f"https://{domain}/about/more", headers=headers, timeout=5).text,
- "html.parser",
- )
- except:
- return {}
-
- for header in doc.find_all("h3"):
- for line in header.find_next_siblings("table")[0].find_all("tr")[1:]:
- header_text = header.text
- if header_text in translations:
- header_text = translations[header_text]
- if header_text in blocks:
- blocks[header_text].append(
- {
- "domain": line.find("span").text,
- "hash": line.find("span")["title"][9:],
- "reason": line.find_all("td")[1].text.strip(),
- }
- )
- return {
- "reject": blocks["Suspended servers"],
- "media_removal": blocks["Filtered media"],
- "federated_timeline_removal": blocks["Limited servers"]
- + blocks["Silenced servers"],
- }
-
-
-def get_hash(domain: str) -> str:
- return sha256(domain.encode("utf-8")).hexdigest()
-
-
-def get_type(domain: str) -> str:
- try:
- res = get(f"https://{domain}/nodeinfo/2.1.json", headers=headers, timeout=5)
- if res.status_code == 404:
- res = get(f"https://{domain}/nodeinfo/2.0.json", headers=headers, timeout=5)
- if res.ok and "text/html" in res.headers["content-type"]:
- res = get(f"https://{domain}/nodeinfo/2.1", headers=headers, timeout=5)
- if res.ok:
- return res.json()["software"]["name"]
- elif res.status_code == 404:
- res = get(f"https://{domain}/api/v1/instance", headers=headers, timeout=5)
- if res.ok:
- return "mastodon"
- except:
- return None
-
-
-conn = sqlite3.connect("blocks.db")
-c = conn.cursor()
-
-c.execute(
- "select domain, software from instances where software in ('pleroma', 'mastodon')"
+#!/usr/bin/python3
+# -*- coding: utf-8 -*-
+
+# Fedi API Block - An aggregator for fetching blocking data from fediverse nodes
+# Copyright (C) 2023 Free Software Foundation
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published
+# by the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program. If not, see <https://www.gnu.org/licenses/>.
+
+import reqto
+import time
+import bs4
+import itertools
+import re
+import fba
+
+fba.cursor.execute(
+ "SELECT domain, software, origin, nodeinfo_url FROM instances WHERE software IN ('pleroma', 'mastodon', 'friendica', 'misskey', 'gotosocial', 'bookwyrm', 'takahe') AND (last_blocked IS NULL OR last_blocked < ?) ORDER BY rowid DESC", [time.time() - fba.config["recheck_block"]]
)
-for blocker, software in c.fetchall():
+rows = fba.cursor.fetchall()
+print(f"INFO: Checking {len(rows)} entries ...")
+for blocker, software, origin, nodeinfo_url in rows:
+ # NOISY-DEBUG: print("DEBUG: BEFORE blocker,software,origin,nodeinfo_url:", blocker, software, origin, nodeinfo_url)
+ blockdict = []
+ blocker = fba.tidyup(blocker)
+ # NOISY-DEBUG: print("DEBUG: AFTER blocker,software:", blocker, software)
+
+ if blocker == "":
+ print("WARNING: blocker is now empty!")
+ continue
+ elif fba.is_blacklisted(blocker):
+ print(f"WARNING: blocker='{blocker}' is blacklisted now!")
+ continue
+
+ # NOISY-DEBUG: print(f"DEBUG: blocker='{blocker}'")
+ fba.update_last_blocked(blocker)
+
if software == "pleroma":
- print(blocker)
+ print("INFO: blocker:", blocker)
try:
# Blocks
- federation = get(
- f"https://{blocker}/nodeinfo/2.1.json", headers=headers, timeout=5
- ).json()["metadata"]["federation"]
+ json = fba.fetch_nodeinfo(blocker, nodeinfo_url)
+ if json is None:
+ print("WARNING: Could not fetch nodeinfo from blocker:", blocker)
+ continue
+
+ print("DEBUG: Updating nodeinfo:", blocker)
+ fba.update_last_nodeinfo(blocker)
+
+ federation = json["metadata"]["federation"]
+
+ if "enabled" in federation:
+ # NOISY-DEBUG: print("DEBUG: Instance has no block list to analyze:", blocker)
+ continue
+
if "mrf_simple" in federation:
for block_level, blocks in (
{**federation["mrf_simple"],
**{"quarantined_instances": federation["quarantined_instances"]}}
).items():
+ # NOISY-DEBUG: print("DEBUG: block_level, blocks():", block_level, len(blocks))
+ block_level = fba.tidyup(block_level)
+ # NOISY-DEBUG: print("DEBUG: BEFORE block_level:", block_level)
+
+ if block_level == "":
+ print("WARNING: block_level is now empty!")
+ continue
+
for blocked in blocks:
+ # NOISY-DEBUG: print("DEBUG: BEFORE blocked:", blocked)
+ blocked = fba.tidyup(blocked)
+ # NOISY-DEBUG: print("DEBUG: AFTER blocked:", blocked)
+
if blocked == "":
+ print("WARNING: blocked is empty after fba.tidyup():", blocker, block_level)
continue
- c.execute(
- "select domain from instances where domain = ?", (blocked,)
- )
- if c.fetchone() == None:
- c.execute(
- "insert into instances select ?, ?, ?",
- (blocked, get_hash(blocked), get_type(blocked)),
+
+ if blocked.count("*") > 1:
+ # -ACK!-oma also started obscuring domains without hash
+ fba.cursor.execute(
+ "SELECT domain, nodeinfo_url FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", [blocked.replace("*", "_")]
)
- c.execute(
- "select * from blocks where blocker = ? and blocked = ? and block_level = ?",
- (blocker, blocked, block_level),
+ searchres = fba.cursor.fetchone()
+ # NOISY-DEBUG: print("DEBUG: searchres[]:", type(searchres))
+ if searchres != None:
+ blocked = searchres[0]
+ nodeinfo_url = searchres[1]
+ # NOISY-DEBUG: print("DEBUG: Looked up domain:", blocked)
+
+ # NOISY-DEBUG: print("DEBUG: Looking up instance by domain:", blocked)
+ if not fba.is_instance_registered(blocked):
+ # NOISY-DEBUG: print("DEBUG: Domain wasn't found, adding:", blocked, blocker)
+ fba.add_instance(blocked, blocker, origin, nodeinfo_url)
+
+ fba.cursor.execute(
+ "SELECT * FROM blocks WHERE blocker = ? AND blocked = ? AND block_level = ? LIMIT 1",
+ (
+ blocker,
+ blocked,
+ block_level
+ ),
)
- if c.fetchone() == None:
- c.execute(
- "insert into blocks select ?, ?, '', ?",
- (blocker, blocked, block_level),
- )
- conn.commit()
+
+ if fba.cursor.fetchone() == None:
+ # NOISY-DEBUG: print("DEBUG: Blocking:", blocker, blocked, block_level)
+ fba.block_instance(blocker, blocked, "unknown", block_level)
+
+ if block_level == "reject":
+ # NOISY-DEBUG: print("DEBUG: Adding to blockdict:", blocked)
+ blockdict.append(
+ {
+ "blocked": blocked,
+ "reason" : None
+ })
+ else:
+ # NOISY-DEBUG: print("DEBUG: Updating last_seen:", blocker, blocked, block_level)
+ fba.update_last_seen(blocker, blocked, block_level)
+
+ fba.connection.commit()
+
# Reasons
if "mrf_simple_info" in federation:
+ # NOISY-DEBUG: print("DEBUG: Found mrf_simple_info:", blocker)
for block_level, info in (
{**federation["mrf_simple_info"],
**(federation["quarantined_instances_info"]
if "quarantined_instances_info" in federation
else {})}
).items():
+ # NOISY-DEBUG: print("DEBUG: block_level, info.items():", block_level, len(info.items()))
+ block_level = fba.tidyup(block_level)
+ # NOISY-DEBUG: print("DEBUG: BEFORE block_level:", block_level)
+
+ if block_level == "":
+ print("WARNING: block_level is now empty!")
+ continue
+
for blocked, reason in info.items():
- c.execute(
- "update blocks set reason = ? where blocker = ? and blocked = ? and block_level = ?",
- (reason["reason"], blocker, blocked, block_level),
- )
- conn.commit()
+ # NOISY-DEBUG: print("DEBUG: BEFORE blocked:", blocked)
+ blocked = fba.tidyup(blocked)
+ # NOISY-DEBUG: print("DEBUG: AFTER blocked:", blocked)
+
+ if blocked == "":
+ print("WARNING: blocked is empty after fba.tidyup():", blocker, block_level)
+ continue
+ elif blocked.count("*") > 1:
+ # same domain guess as above, but for reasons field
+ fba.cursor.execute(
+ "SELECT domain FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", [blocked.replace("*", "_")]
+ )
+ searchres = fba.cursor.fetchone()
+
+ if searchres != None:
+ blocked = searchres[0]
+
+ # NOISY-DEBUG: print("DEBUG: Updating block reason:", blocker, blocked, reason["reason"])
+ fba.update_block_reason(reason["reason"], blocker, blocked, block_level)
+
+ for entry in blockdict:
+ if entry["blocked"] == blocked:
+ # NOISY-DEBUG: print("DEBUG: Updating entry reason:", blocked)
+ entry["reason"] = reason["reason"]
+
+ fba.connection.commit()
except Exception as e:
- print("error:", e, blocker)
+ print(f"ERROR: blocker='{blocker}',software='{software}',exception[{type(e)}]:'{str(e)}'")
elif software == "mastodon":
- print(blocker)
+ print("INFO: blocker:", blocker)
try:
- json = get_mastodon_blocks(blocker)
+ # json endpoint for newer mastodongs
+ try:
+ json = {
+ "reject" : [],
+ "media_removal" : [],
+ "followers_only": [],
+ "report_removal": []
+ }
+
+ # handling CSRF, I've saw at least one server requiring it to access the endpoint
+ # NOISY-DEBUG: print("DEBUG: Fetching meta:", blocker)
+ meta = bs4.BeautifulSoup(
+ reqto.get(f"https://{blocker}/about", headers=fba.headers, timeout=(fba.config["connection_timeout"], config["read_timeout"])).text,
+ "html.parser",
+ )
+ try:
+ csrf = meta.find("meta", attrs={"name": "csrf-token"})["content"]
+ # NOISY-DEBUG: print("DEBUG: Adding CSRF token:", blocker, csrf)
+ reqheaders = {**fba.api_headers, **{"X-CSRF-Token": csrf}}
+ except:
+ # NOISY-DEBUG: print("DEBUG: No CSRF token found, using normal headers:", blocker)
+ reqheaders = fba.api_headers
+
+ # NOISY-DEBUG: print("DEBUG: Quering API domain_blocks:", blocker)
+ blocks = reqto.get(f"https://{blocker}/api/v1/instance/domain_blocks", headers=reqheaders, timeout=(fba.config["connection_timeout"], config["read_timeout"])).json()
+
+ # NOISY-DEBUG: print("DEBUG: blocks():", len(blocks))
+ for block in blocks:
+ entry = {
+ 'domain': block['domain'],
+ 'hash' : block['digest'],
+ 'reason': block['comment']
+ }
+
+ # NOISY-DEBUG: print("DEBUG: severity,domain,hash,comment:", block['severity'], block['domain'], block['digest'], block['comment'])
+ if block['severity'] == 'suspend':
+ json['reject'].append(entry)
+ elif block['severity'] == 'silence':
+ json['followers_only'].append(entry)
+ elif block['severity'] == 'reject_media':
+ json['media_removal'].append(entry)
+ elif block['severity'] == 'reject_reports':
+ json['report_removal'].append(entry)
+ else:
+ print("WARNING: Unknown severity:", block['severity'], block['domain'])
+ except:
+ # NOISY-DEBUG: print("DEBUG: Failed, Trying mastodon-specific fetches:", blocker)
+ json = fba.get_mastodon_blocks(blocker)
+
+ # NOISY-DEBUG: print("DEBUG: json.items():", blocker, len(json.items()))
for block_level, blocks in json.items():
+ # NOISY-DEBUG: print("DEBUG: blocker,block_level,blocks():", blocker, block_level, len(blocks))
+ block_level = fba.tidyup(block_level)
+ # NOISY-DEBUG: print("DEBUG: AFTER-block_level:", block_level)
+ if block_level == "":
+ print("WARNING: block_level is empty, blocker:", blocker)
+ continue
+
for instance in blocks:
blocked, blocked_hash, reason = instance.values()
- if blocked.count("*") <= 1:
- c.execute(
- "select hash from instances where hash = ?", (blocked_hash,)
+ # NOISY-DEBUG: print("DEBUG: blocked,hash,reason:", blocked, blocked_hash, reason)
+ blocked = fba.tidyup(blocked)
+ # NOISY-DEBUG: print("DEBUG: AFTER-blocked:", blocked)
+
+ if blocked == "":
+ print("WARNING: blocked is empty:", blocker)
+ continue
+ elif blocked.count("*") < 1:
+ # No obsfucation for this instance
+ fba.cursor.execute(
+ "SELECT hash FROM instances WHERE domain = ? LIMIT 1", [blocked]
)
- if c.fetchone() == None:
- c.execute(
- "insert into instances select ?, ?, ?",
- (blocked, get_hash(blocked), get_type(blocked)),
- )
- c.execute(
- "select * from blocks where blocker = ? and blocked = ? and block_level = ?",
- (blocker, blocked, block_level),
+
+ if fba.cursor.fetchone() == None:
+ # NOISY-DEBUG: print("DEBUG: Hash wasn't found, adding:", blocked, blocker)
+ fba.add_instance(blocked, blocker, origin)
+ else:
+ # Doing the hash search for instance names as well to tidy up DB
+ fba.cursor.execute(
+ "SELECT domain FROM instances WHERE hash = ? LIMIT 1", [blocked_hash]
+ )
+ searchres = fba.cursor.fetchone()
+
+ if searchres != None:
+ # NOISY-DEBUG: print("DEBUG: Updating domain: ", searchres[0])
+ blocked = searchres[0]
+
+ fba.cursor.execute(
+ "SELECT * FROM blocks WHERE blocker = ? AND blocked = ? AND block_level = ? LIMIT 1",
+ (
+ blocker,
+ blocked if blocked.count("*") <= 1 else blocked_hash,
+ block_level
+ ),
)
- if c.fetchone() == None:
- c.execute(
- "insert into blocks select ?, ?, ?, ?",
- (
- blocker,
- blocked if blocked.count("*") <= 1 else blocked_hash,
- reason,
- block_level,
- ),
+
+ if fba.cursor.fetchone() == None:
+ fba.block_instance(blocker, blocked if blocked.count("*") <= 1 else blocked_hash, reason, block_level)
+
+ if block_level == "reject":
+ blockdict.append(
+ {
+ "blocked": blocked,
+ "reason" : reason
+ })
+ else:
+ fba.update_last_seen(blocker, blocked if blocked.count("*") <= 1 else blocked_hash, block_level)
+
+ if reason != "":
+ # NOISY-DEBUG: print("DEBUG: Updating block reason:", blocker, blocked, reason)
+ fba.update_block_reason(reason, blocker, blocked if blocked.count("*") <= 1 else blocked_hash, block_level)
+
+ fba.connection.commit()
+ except Exception as e:
+ print(f"ERROR: blocker='{blocker}',software='{software}',exception[{type(e)}]:'{str(e)}'")
+ elif software == "friendica" or software == "misskey" or software == "bookwyrm" or software == "takahe":
+ print("INFO: blocker:", blocker)
+ try:
+ if software == "friendica":
+ json = fba.get_friendica_blocks(blocker)
+ elif software == "misskey":
+ json = fba.get_misskey_blocks(blocker)
+ elif software == "bookwyrm":
+ print("WARNING: bookwyrm is not fully supported for fetching blacklist!", blocker)
+ #json = fba.get_bookwyrm_blocks(blocker)
+ elif software == "takahe":
+ print("WARNING: takahe is not fully supported for fetching blacklist!", blocker)
+ #json = fba.get_takahe_blocks(blocker)
+
+ for block_level, blocks in json.items():
+ # NOISY-DEBUG: print("DEBUG: blocker,block_level,blocks():", blocker, block_level, len(blocks))
+ block_level = fba.tidyup(block_level)
+ # NOISY-DEBUG: print("DEBUG: AFTER-block_level:", block_level)
+ if block_level == "":
+ print("WARNING: block_level is empty, blocker:", blocker)
+ continue
+
+ for instance in blocks:
+ blocked, reason = instance.values()
+ # NOISY-DEBUG: print("DEBUG: BEFORE blocked:", blocked)
+ blocked = fba.tidyup(blocked)
+ # NOISY-DEBUG: print("DEBUG: AFTER blocked:", blocked)
+
+ if blocked == "":
+ print("WARNING: blocked is empty:", blocker)
+ continue
+ if blocked.count("*") > 0:
+ # Some friendica servers also obscure domains without hash
+ fba.cursor.execute(
+ "SELECT domain FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", [blocked.replace("*", "_")]
+ )
+ searchres = fba.cursor.fetchone()
+ if searchres != None:
+ blocked = searchres[0]
+
+ if blocked.count("?") > 0:
+ # Some obscure them with question marks, not sure if that's dependent on version or not
+ fba.cursor.execute(
+ "SELECT domain, origin, nodeinfo_url FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", [blocked.replace("?", "_")]
+ )
+ searchres = fba.cursor.fetchone()
+ if searchres != None:
+ blocked = searchres[0]
+ origin = searchres[1]
+ nodeinfo_url = searchres[2]
+
+ # NOISY-DEBUG: print("DEBUG: AFTER-blocked:", blocked)
+ if not fba.is_instance_registered(blocked):
+ # NOISY-DEBUG: print("DEBUG: Hash wasn't found, adding:", blocked, blocker)
+ fba.add_instance(blocked, blocker, origin, nodeinfo_url)
+
+ fba.cursor.execute(
+ "SELECT * FROM blocks WHERE blocker = ? AND blocked = ?",
+ (blocker, blocked),
+ )
+
+ if fba.cursor.fetchone() == None:
+ fba.block_instance(blocker, blocked, reason, block_level)
+
+ if block_level == "reject":
+ blockdict.append(
+ {
+ "blocked": blocked,
+ "reason" : reason
+ })
+ else:
+ fba.update_last_seen(blocker, blocked, block_level)
+
+ if reason != '':
+ # NOISY-DEBUG: print("DEBUG: Updating block reason:", blocker, blocked, reason)
+ fba.update_block_reason(reason, blocker, blocked, block_level)
+
+ fba.connection.commit()
+ except Exception as e:
+ print(f"ERROR: blocker='{blocker}',software='{software}',exception[{type(e)}]:'{str(e)}'")
+ elif software == "gotosocial":
+ print("INFO: blocker:", blocker)
+ try:
+ # Blocks
+ federation = reqto.get(f"https://{blocker}{get_peers_url}?filter=suspended", headers=fba.api_headers, timeout=(fba.config["connection_timeout"], config["read_timeout"])).json()
+
+ if (federation == None):
+ print("WARNING: No valid response:", blocker);
+ elif "error" in federation:
+ print("WARNING: API returned error:", federation["error"])
+ else:
+ # NOISY-DEBUG: print("DEBUG: Checking fenderation():", len(federation))
+ for peer in federation:
+ blocked = peer["domain"].lower()
+ # NOISY-DEBUG: print("DEBUG: BEFORE blocked:", blocked)
+ blocked = fba.tidyup(blocked)
+ # NOISY-DEBUG: print("DEBUG: AFTER blocked:", blocked)
+
+ if blocked == "":
+ print("WARNING: blocked is empty:", blocker)
+ continue
+ elif blocked.count("*") > 0:
+ # GTS does not have hashes for obscured domains, so we have to guess it
+ fba.cursor.execute(
+ "SELECT domain, origin, nodeinfo_url FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", [blocked.replace("*", "_")]
)
- conn.commit()
+ searchres = fba.cursor.fetchone()
+
+ if searchres != None:
+ blocked = searchres[0]
+ origin = searchres[1]
+ nodeinfo_url = searchres[2]
+
+ if not fba.is_instance_registered(blocked):
+ # NOISY-DEBUG: print("DEBUG: Domain wasn't found, adding:", blocked, blocker)
+ fba.add_instance(blocked, blocker, origin, nodeinfo_url)
+
+ fba.cursor.execute(
+ "SELECT * FROM blocks WHERE blocker = ? AND blocked = ? AND block_level = ? LIMIT 1",
+ (
+ blocker,
+ blocked,
+ "reject"
+ ),
+ )
+
+ if fba.cursor.fetchone() == None:
+ # NOISY-DEBUG: print(f"DEBUG: blocker='{blocker}' is blocking '{blocked}' for unknown reason at this point")
+ fba.block_instance(blocker, blocked, "unknown", "reject")
+
+ blockdict.append(
+ {
+ "blocked": blocked,
+ "reason" : None
+ })
+ else:
+ fba.update_last_seen(blocker, blocked, "reject")
+
+ if "public_comment" in peer:
+ # NOISY-DEBUG: print("DEBUG: Updating block reason:", blocker, blocked, peer["public_comment"])
+ fba.update_block_reason(peer["public_comment"], blocker, blocked, "reject")
+
+ for entry in blockdict:
+ if entry["blocked"] == blocked:
+ # NOISY-DEBUG: print(f"DEBUG: Setting block reason for blocked='{blocked}':'{peer['public_comment']}'")
+ entry["reason"] = peer["public_comment"]
+
+ fba.connection.commit()
except Exception as e:
- print("error:", e, blocker)
-conn.close()
+ print(f"ERROR: blocker='{blocker}',software='{software}',exception[{type(e)}]:'{str(e)}'")
+ else:
+ print("WARNING: Unknown software:", blocker, software)
+
+ if fba.config["bot_enabled"] and len(blockdict) > 0:
+ send_bot_post(blocker, blockdict)
+
+ blockdict = []
+
+fba.connection.close()
projects / fba.git / blobdiff