]> git.ipfire.org Git - people/ms/libloc.git/blob - src/python/location-importer.in
projects / people / ms / libloc.git / blob
? search:


864eab12b329b9082f7279b6fb723f5a645cef02
[people/ms/libloc.git] / src / python / location-importer.in
1 #!/usr/bin/python3
2 ###############################################################################
3 # #
4 # libloc - A library to determine the location of someone on the Internet #
5 # #
6 # Copyright (C) 2020 IPFire Development Team <info@ipfire.org> #
7 # #
8 # This library is free software; you can redistribute it and/or #
9 # modify it under the terms of the GNU Lesser General Public #
10 # License as published by the Free Software Foundation; either #
11 # version 2.1 of the License, or (at your option) any later version. #
12 # #
13 # This library is distributed in the hope that it will be useful, #
14 # but WITHOUT ANY WARRANTY; without even the implied warranty of #
15 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU #
16 # Lesser General Public License for more details. #
17 # #
18 ###############################################################################
19
20 import argparse
21 import ipaddress
22 import logging
23 import math
24 import re
25 import socket
26 import sys
27 import telnetlib
28
29 # Load our location module
30 import location
31 import location.database
32 import location.importer
33 from location.i18n import _
34
35 # Initialise logging
36 log = logging.getLogger("location.importer")
37 log.propagate = 1
38
39 class CLI(object):
40 def parse_cli(self):
41 parser = argparse.ArgumentParser(
42 description=_("Location Importer Command Line Interface"),
43 )
44 subparsers = parser.add_subparsers()
45
46 # Global configuration flags
47 parser.add_argument("--debug", action="store_true",
48 help=_("Enable debug output"))
49 parser.add_argument("--quiet", action="store_true",
50 help=_("Enable quiet mode"))
51
52 # version
53 parser.add_argument("--version", action="version",
54 version="%(prog)s @VERSION@")
55
56 # Database
57 parser.add_argument("--database-host", required=True,
58 help=_("Database Hostname"), metavar=_("HOST"))
59 parser.add_argument("--database-name", required=True,
60 help=_("Database Name"), metavar=_("NAME"))
61 parser.add_argument("--database-username", required=True,
62 help=_("Database Username"), metavar=_("USERNAME"))
63 parser.add_argument("--database-password", required=True,
64 help=_("Database Password"), metavar=_("PASSWORD"))
65
66 # Write Database
67 write = subparsers.add_parser("write", help=_("Write database to file"))
68 write.set_defaults(func=self.handle_write)
69 write.add_argument("file", nargs=1, help=_("Database File"))
70 write.add_argument("--signing-key", nargs="?", type=open, help=_("Signing Key"))
71 write.add_argument("--backup-signing-key", nargs="?", type=open, help=_("Backup Signing Key"))
72 write.add_argument("--vendor", nargs="?", help=_("Sets the vendor"))
73 write.add_argument("--description", nargs="?", help=_("Sets a description"))
74 write.add_argument("--license", nargs="?", help=_("Sets the license"))
75 write.add_argument("--version", type=int, help=_("Database Format Version"))
76
77 # Update WHOIS
78 update_whois = subparsers.add_parser("update-whois", help=_("Update WHOIS Information"))
79 update_whois.set_defaults(func=self.handle_update_whois)
80
81 # Update announcements
82 update_announcements = subparsers.add_parser("update-announcements",
83 help=_("Update BGP Annoucements"))
84 update_announcements.set_defaults(func=self.handle_update_announcements)
85 update_announcements.add_argument("server", nargs=1,
86 help=_("Route Server to connect to"), metavar=_("SERVER"))
87
88 # Update overrides
89 update_overrides = subparsers.add_parser("update-overrides",
90 help=_("Update overrides"),
91 )
92 update_overrides.add_argument(
93 "files", nargs="+", help=_("Files to import"),
94 )
95 update_overrides.set_defaults(func=self.handle_update_overrides)
96
97 # Import countries
98 import_countries = subparsers.add_parser("import-countries",
99 help=_("Import countries"),
100 )
101 import_countries.add_argument("file", nargs=1, type=argparse.FileType("r"),
102 help=_("File to import"))
103 import_countries.set_defaults(func=self.handle_import_countries)
104
105 args = parser.parse_args()
106
107 # Configure logging
108 if args.debug:
109 location.logger.set_level(logging.DEBUG)
110 elif args.quiet:
111 location.logger.set_level(logging.WARNING)
112
113 # Print usage if no action was given
114 if not "func" in args:
115 parser.print_usage()
116 sys.exit(2)
117
118 return args
119
120 def run(self):
121 # Parse command line arguments
122 args = self.parse_cli()
123
124 # Initialise database
125 self.db = self._setup_database(args)
126
127 # Call function
128 ret = args.func(args)
129
130 # Return with exit code
131 if ret:
132 sys.exit(ret)
133
134 # Otherwise just exit
135 sys.exit(0)
136
137 def _setup_database(self, ns):
138 """
139 Initialise the database
140 """
141 # Connect to database
142 db = location.database.Connection(
143 host=ns.database_host, database=ns.database_name,
144 user=ns.database_username, password=ns.database_password,
145 )
146
147 with db.transaction():
148 db.execute("""
149 -- announcements
150 CREATE TABLE IF NOT EXISTS announcements(network inet, autnum bigint,
151 first_seen_at timestamp without time zone DEFAULT CURRENT_TIMESTAMP,
152 last_seen_at timestamp without time zone DEFAULT CURRENT_TIMESTAMP);
153 CREATE UNIQUE INDEX IF NOT EXISTS announcements_networks ON announcements(network);
154 CREATE INDEX IF NOT EXISTS announcements_family ON announcements(family(network));
155 CREATE INDEX IF NOT EXISTS announcements_search ON announcements USING GIST(network inet_ops);
156
157 -- autnums
158 CREATE TABLE IF NOT EXISTS autnums(number bigint, name text NOT NULL);
159 CREATE UNIQUE INDEX IF NOT EXISTS autnums_number ON autnums(number);
160
161 -- countries
162 CREATE TABLE IF NOT EXISTS countries(
163 country_code text NOT NULL, name text NOT NULL, continent_code text NOT NULL);
164 CREATE UNIQUE INDEX IF NOT EXISTS countries_country_code ON countries(country_code);
165
166 -- networks
167 CREATE TABLE IF NOT EXISTS networks(network inet, country text);
168 CREATE UNIQUE INDEX IF NOT EXISTS networks_network ON networks(network);
169 CREATE INDEX IF NOT EXISTS networks_family ON networks USING BTREE(family(network));
170 CREATE INDEX IF NOT EXISTS networks_search ON networks USING GIST(network inet_ops);
171
172 -- overrides
173 CREATE TABLE IF NOT EXISTS autnum_overrides(
174 number bigint NOT NULL,
175 name text,
176 country text,
177 is_anonymous_proxy boolean,
178 is_satellite_provider boolean,
179 is_anycast boolean
180 );
181 CREATE UNIQUE INDEX IF NOT EXISTS autnum_overrides_number
182 ON autnum_overrides(number);
183
184 CREATE TABLE IF NOT EXISTS network_overrides(
185 network inet NOT NULL,
186 country text,
187 is_anonymous_proxy boolean,
188 is_satellite_provider boolean,
189 is_anycast boolean
190 );
191 CREATE UNIQUE INDEX IF NOT EXISTS network_overrides_network
192 ON network_overrides(network);
193 CREATE INDEX IF NOT EXISTS network_overrides_search
194 ON network_overrides USING GIST(network inet_ops);
195 """)
196
197 return db
198
199 def handle_write(self, ns):
200 """
201 Compiles a database in libloc format out of what is in the database
202 """
203 # Allocate a writer
204 writer = location.Writer(ns.signing_key, ns.backup_signing_key)
205
206 # Set all metadata
207 if ns.vendor:
208 writer.vendor = ns.vendor
209
210 if ns.description:
211 writer.description = ns.description
212
213 if ns.license:
214 writer.license = ns.license
215
216 # Add all Autonomous Systems
217 log.info("Writing Autonomous Systems...")
218
219 # Select all ASes with a name
220 rows = self.db.query("""
221 SELECT
222 autnums.number AS number,
223 COALESCE(
224 (SELECT overrides.name FROM autnum_overrides overrides
225 WHERE overrides.number = autnums.number),
226 autnums.name
227 ) AS name
228 FROM autnums
229 WHERE name <> %s ORDER BY number
230 """, "")
231
232 for row in rows:
233 a = writer.add_as(row.number)
234 a.name = row.name
235
236 # Add all networks
237 log.info("Writing networks...")
238
239 # Select all known networks
240 rows = self.db.query("""
241 -- Return a list of those networks enriched with all
242 -- other information that we store in the database
243 SELECT
244 DISTINCT ON (network)
245 network,
246 autnum,
247
248 -- Country
249 COALESCE(
250 (
251 SELECT country FROM network_overrides overrides
252 WHERE networks.network <<= overrides.network
253 ORDER BY masklen(overrides.network) DESC
254 LIMIT 1
255 ),
256 (
257 SELECT country FROM autnum_overrides overrides
258 WHERE networks.autnum = overrides.number
259 ),
260 networks.country
261 ) AS country,
262
263 -- Flags
264 COALESCE(
265 (
266 SELECT is_anonymous_proxy FROM network_overrides overrides
267 WHERE networks.network <<= overrides.network
268 ORDER BY masklen(overrides.network) DESC
269 LIMIT 1
270 ),
271 (
272 SELECT is_anonymous_proxy FROM autnum_overrides overrides
273 WHERE networks.autnum = overrides.number
274 ),
275 FALSE
276 ) AS is_anonymous_proxy,
277 COALESCE(
278 (
279 SELECT is_satellite_provider FROM network_overrides overrides
280 WHERE networks.network <<= overrides.network
281 ORDER BY masklen(overrides.network) DESC
282 LIMIT 1
283 ),
284 (
285 SELECT is_satellite_provider FROM autnum_overrides overrides
286 WHERE networks.autnum = overrides.number
287 ),
288 FALSE
289 ) AS is_satellite_provider,
290 COALESCE(
291 (
292 SELECT is_anycast FROM network_overrides overrides
293 WHERE networks.network <<= overrides.network
294 ORDER BY masklen(overrides.network) DESC
295 LIMIT 1
296 ),
297 (
298 SELECT is_anycast FROM autnum_overrides overrides
299 WHERE networks.autnum = overrides.number
300 ),
301 FALSE
302 ) AS is_anycast
303 FROM (
304 SELECT
305 known_networks.network AS network,
306 announcements.autnum AS autnum,
307 networks.country AS country,
308
309 -- Must be part of returned values for ORDER BY clause
310 masklen(announcements.network) AS sort_a,
311 masklen(networks.network) AS sort_b
312 FROM (
313 SELECT network FROM announcements
314 UNION ALL
315 SELECT network FROM networks
316 UNION ALL
317 SELECT network FROM network_overrides
318 ) known_networks
319 LEFT JOIN
320 announcements ON known_networks.network <<= announcements.network
321 LEFT JOIN
322 networks ON known_networks.network <<= networks.network
323 ORDER BY
324 known_networks.network,
325 sort_a DESC,
326 sort_b DESC
327 ) networks
328 """)
329
330 for row in rows:
331 network = writer.add_network(row.network)
332
333 # Save country
334 if row.country:
335 network.country_code = row.country
336
337 # Save ASN
338 if row.autnum:
339 network.asn = row.autnum
340
341 # Set flags
342 if row.is_anonymous_proxy:
343 network.set_flag(location.NETWORK_FLAG_ANONYMOUS_PROXY)
344
345 if row.is_satellite_provider:
346 network.set_flag(location.NETWORK_FLAG_SATELLITE_PROVIDER)
347
348 if row.is_anycast:
349 network.set_flag(location.NETWORK_FLAG_ANYCAST)
350
351 # Add all countries
352 log.info("Writing countries...")
353 rows = self.db.query("SELECT * FROM countries ORDER BY country_code")
354
355 for row in rows:
356 c = writer.add_country(row.country_code)
357 c.continent_code = row.continent_code
358 c.name = row.name
359
360 # Write everything to file
361 log.info("Writing database to file...")
362 for file in ns.file:
363 writer.write(file)
364
365 def handle_update_whois(self, ns):
366 downloader = location.importer.Downloader()
367
368 # Download all sources
369 with self.db.transaction():
370 # Create some temporary tables to store parsed data
371 self.db.execute("""
372 CREATE TEMPORARY TABLE _autnums(number integer, organization text)
373 ON COMMIT DROP;
374 CREATE UNIQUE INDEX _autnums_number ON _autnums(number);
375
376 CREATE TEMPORARY TABLE _organizations(handle text, name text NOT NULL)
377 ON COMMIT DROP;
378 CREATE UNIQUE INDEX _organizations_handle ON _organizations(handle);
379
380 CREATE TEMPORARY TABLE _rirdata(network inet NOT NULL, country text NOT NULL)
381 ON COMMIT DROP;
382 CREATE INDEX _rirdata_search ON _rirdata USING BTREE(family(network), masklen(network));
383 CREATE UNIQUE INDEX _rirdata_network ON _rirdata(network);
384 """)
385
386 # Remove all previously imported content
387 self.db.execute("""
388 TRUNCATE TABLE networks;
389 """)
390
391 for source in location.importer.WHOIS_SOURCES:
392 with downloader.request(source, return_blocks=True) as f:
393 for block in f:
394 self._parse_block(block)
395
396 # Process all parsed networks from every RIR we happen to have access to,
397 # insert the largest network chunks into the networks table immediately...
398 families = self.db.query("SELECT DISTINCT family(network) AS family FROM _rirdata ORDER BY family(network)")
399
400 for family in (row.family for row in families):
401 smallest = self.db.get("SELECT MIN(masklen(network)) AS prefix FROM _rirdata WHERE family(network) = %s", family)
402
403 self.db.execute("INSERT INTO networks(network, country) \
404 SELECT network, country FROM _rirdata WHERE masklen(network) = %s AND family(network) = %s", smallest.prefix, family)
405
406 # ... determine any other prefixes for this network family, ...
407 prefixes = self.db.query("SELECT DISTINCT masklen(network) AS prefix FROM _rirdata \
408 WHERE family(network) = %s ORDER BY masklen(network) ASC OFFSET 1", family)
409
410 # ... and insert networks with this prefix in case they provide additional
411 # information (i. e. subnet of a larger chunk with a different country)
412 for prefix in (row.prefix for row in prefixes):
413 self.db.execute("""
414 WITH candidates AS (
415 SELECT
416 _rirdata.network,
417 _rirdata.country
418 FROM
419 _rirdata
420 WHERE
421 family(_rirdata.network) = %s
422 AND
423 masklen(_rirdata.network) = %s
424 ),
425 filtered AS (
426 SELECT
427 DISTINCT ON (c.network)
428 c.network,
429 c.country,
430 masklen(networks.network),
431 networks.country AS parent_country
432 FROM
433 candidates c
434 LEFT JOIN
435 networks
436 ON
437 c.network << networks.network
438 ORDER BY
439 c.network,
440 masklen(networks.network) DESC NULLS LAST
441 )
442 INSERT INTO
443 networks(network, country)
444 SELECT
445 network,
446 country
447 FROM
448 filtered
449 WHERE
450 parent_country IS NULL
451 OR
452 country <> parent_country
453 ON CONFLICT DO NOTHING""",
454 family, prefix,
455 )
456
457 self.db.execute("""
458 INSERT INTO autnums(number, name)
459 SELECT _autnums.number, _organizations.name FROM _autnums
460 JOIN _organizations ON _autnums.organization = _organizations.handle
461 ON CONFLICT (number) DO UPDATE SET name = excluded.name;
462 """)
463
464 # Download all extended sources
465 for source in location.importer.EXTENDED_SOURCES:
466 with self.db.transaction():
467 # Download data
468 with downloader.request(source) as f:
469 for line in f:
470 self._parse_line(line)
471
472 def _check_parsed_network(self, network):
473 """
474 Assistive function to detect and subsequently sort out parsed
475 networks from RIR data (both Whois and so-called "extended sources"),
476 which are or have...
477
478 (a) not globally routable (RFC 1918 space, et al.)
479 (b) covering a too large chunk of the IP address space (prefix length
480 is < 7 for IPv4 networks, and < 10 for IPv6)
481 (c) "0.0.0.0" or "::" as a network address
482 (d) are too small for being publicly announced (we have decided not to
483 process them at the moment, as they significantly enlarge our
484 database without providing very helpful additional information)
485
486 This unfortunately is necessary due to brain-dead clutter across
487 various RIR databases, causing mismatches and eventually disruptions.
488
489 We will return False in case a network is not suitable for adding
490 it to our database, and True otherwise.
491 """
492
493 if not network or not (isinstance(network, ipaddress.IPv4Network) or isinstance(network, ipaddress.IPv6Network)):
494 return False
495
496 if not network.is_global:
497 log.warning("Skipping non-globally routable network: %s" % network)
498 return False
499
500 if network.version == 4:
501 if network.prefixlen < 7:
502 log.warning("Skipping too big IP chunk: %s" % network)
503 return False
504
505 if network.prefixlen > 24:
506 log.debug("Skipping network too small to be publicly announced: %s" % network)
507 return False
508
509 if str(network.network_address) == "0.0.0.0":
510 log.warning("Skipping network based on 0.0.0.0: %s" % network)
511 return False
512
513 elif network.version == 6:
514 if network.prefixlen < 10:
515 log.warning("Skipping too big IP chunk: %s" % network)
516 return False
517
518 if network.prefixlen > 48:
519 log.debug("Skipping network too small to be publicly announced: %s" % network)
520 return False
521
522 if str(network.network_address) == "::":
523 log.warning("Skipping network based on '::': %s" % network)
524 return False
525
526 else:
527 # This should not happen...
528 log.warning("Skipping network of unknown family, this should not happen: %s" % network)
529 return False
530
531 # In case we have made it here, the network is considered to
532 # be suitable for libloc consumption...
533 return True
534
535 def _parse_block(self, block):
536 # Get first line to find out what type of block this is
537 line = block[0]
538
539 # aut-num
540 if line.startswith("aut-num:"):
541 return self._parse_autnum_block(block)
542
543 # inetnum
544 if line.startswith("inet6num:") or line.startswith("inetnum:"):
545 return self._parse_inetnum_block(block)
546
547 # organisation
548 elif line.startswith("organisation:"):
549 return self._parse_org_block(block)
550
551 def _parse_autnum_block(self, block):
552 autnum = {}
553 for line in block:
554 # Split line
555 key, val = split_line(line)
556
557 if key == "aut-num":
558 m = re.match(r"^(AS|as)(\d+)", val)
559 if m:
560 autnum["asn"] = m.group(2)
561
562 elif key == "org":
563 autnum[key] = val
564
565 # Skip empty objects
566 if not autnum:
567 return
568
569 # Insert into database
570 self.db.execute("INSERT INTO _autnums(number, organization) \
571 VALUES(%s, %s) ON CONFLICT (number) DO UPDATE SET \
572 organization = excluded.organization",
573 autnum.get("asn"), autnum.get("org"),
574 )
575
576 def _parse_inetnum_block(self, block):
577 log.debug("Parsing inetnum block:")
578
579 inetnum = {}
580 for line in block:
581 log.debug(line)
582
583 # Split line
584 key, val = split_line(line)
585
586 # Filter any inetnum records which are only referring to IP space
587 # not managed by that specific RIR...
588 if key == "netname":
589 if re.match(r"(ERX-NETBLOCK|(AFRINIC|ARIN|LACNIC|RIPE)-CIDR-BLOCK|IANA-NETBLOCK-\d{1,3}|NON-RIPE-NCC-MANAGED-ADDRESS-BLOCK)", val.strip()):
590 log.warning("Skipping record indicating historic/orphaned data: %s" % val.strip())
591 return
592
593 if key == "inetnum":
594 start_address, delim, end_address = val.partition("-")
595
596 # Strip any excess space
597 start_address, end_address = start_address.rstrip(), end_address.strip()
598
599 # Convert to IP address
600 try:
601 start_address = ipaddress.ip_address(start_address)
602 end_address = ipaddress.ip_address(end_address)
603 except ValueError:
604 log.warning("Could not parse line: %s" % line)
605 return
606
607 # Set prefix to default
608 prefix = 32
609
610 # Count number of addresses in this subnet
611 num_addresses = int(end_address) - int(start_address)
612 if num_addresses:
613 prefix -= math.log(num_addresses, 2)
614
615 inetnum["inetnum"] = "%s/%.0f" % (start_address, prefix)
616
617 elif key == "inet6num":
618 inetnum[key] = val
619
620 elif key == "country":
621 inetnum[key] = val.upper()
622
623 # Skip empty objects
624 if not inetnum or not "country" in inetnum:
625 return
626
627 # Skip objects with bogus country code 'ZZ'
628 if inetnum.get("country") == "ZZ":
629 log.warning("Skipping network with bogus country 'ZZ': %s" % \
630 (inetnum.get("inet6num") or inetnum.get("inetnum")))
631 return
632
633 network = ipaddress.ip_network(inetnum.get("inet6num") or inetnum.get("inetnum"), strict=False)
634
635 if not self._check_parsed_network(network):
636 return
637
638 self.db.execute("INSERT INTO _rirdata(network, country) \
639 VALUES(%s, %s) ON CONFLICT (network) DO UPDATE SET country = excluded.country",
640 "%s" % network, inetnum.get("country"),
641 )
642
643 def _parse_org_block(self, block):
644 org = {}
645 for line in block:
646 # Split line
647 key, val = split_line(line)
648
649 if key in ("organisation", "org-name"):
650 org[key] = val
651
652 # Skip empty objects
653 if not org:
654 return
655
656 self.db.execute("INSERT INTO _organizations(handle, name) \
657 VALUES(%s, %s) ON CONFLICT (handle) DO \
658 UPDATE SET name = excluded.name",
659 org.get("organisation"), org.get("org-name"),
660 )
661
662 def _parse_line(self, line):
663 # Skip version line
664 if line.startswith("2"):
665 return
666
667 # Skip comments
668 if line.startswith("#"):
669 return
670
671 try:
672 registry, country_code, type, line = line.split("|", 3)
673 except:
674 log.warning("Could not parse line: %s" % line)
675 return
676
677 # Skip any lines that are for stats only
678 if country_code == "*":
679 return
680
681 if type in ("ipv6", "ipv4"):
682 return self._parse_ip_line(country_code, type, line)
683
684 def _parse_ip_line(self, country, type, line):
685 try:
686 address, prefix, date, status, organization = line.split("|")
687 except ValueError:
688 organization = None
689
690 # Try parsing the line without organization
691 try:
692 address, prefix, date, status = line.split("|")
693 except ValueError:
694 log.warning("Unhandled line format: %s" % line)
695 return
696
697 # Skip anything that isn't properly assigned
698 if not status in ("assigned", "allocated"):
699 return
700
701 # Cast prefix into an integer
702 try:
703 prefix = int(prefix)
704 except:
705 log.warning("Invalid prefix: %s" % prefix)
706 return
707
708 # Fix prefix length for IPv4
709 if type == "ipv4":
710 prefix = 32 - int(math.log(prefix, 2))
711
712 # Try to parse the address
713 try:
714 network = ipaddress.ip_network("%s/%s" % (address, prefix), strict=False)
715 except ValueError:
716 log.warning("Invalid IP address: %s" % address)
717 return
718
719 if not self._check_parsed_network(network):
720 return
721
722 self.db.execute("INSERT INTO networks(network, country) \
723 VALUES(%s, %s) ON CONFLICT (network) DO \
724 UPDATE SET country = excluded.country",
725 "%s" % network, country,
726 )
727
728 def handle_update_announcements(self, ns):
729 server = ns.server[0]
730
731 with self.db.transaction():
732 if server.startswith("/"):
733 self._handle_update_announcements_from_bird(server)
734 else:
735 self._handle_update_announcements_from_telnet(server)
736
737 # Purge anything we never want here
738 self.db.execute("""
739 -- Delete default routes
740 DELETE FROM announcements WHERE network = '::/0' OR network = '0.0.0.0/0';
741
742 -- Delete anything that is not global unicast address space
743 DELETE FROM announcements WHERE family(network) = 6 AND NOT network <<= '2000::/3';
744
745 -- DELETE "current network" address space
746 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '0.0.0.0/8';
747
748 -- DELETE local loopback address space
749 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '127.0.0.0/8';
750
751 -- DELETE RFC 1918 address space
752 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '10.0.0.0/8';
753 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '172.16.0.0/12';
754 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '192.168.0.0/16';
755
756 -- DELETE test, benchmark and documentation address space
757 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '192.0.0.0/24';
758 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '192.0.2.0/24';
759 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '198.18.0.0/15';
760 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '198.51.100.0/24';
761 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '203.0.113.0/24';
762
763 -- DELETE CGNAT address space (RFC 6598)
764 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '100.64.0.0/10';
765
766 -- DELETE link local address space
767 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '169.254.0.0/16';
768
769 -- DELETE IPv6 to IPv4 (6to4) address space
770 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '192.88.99.0/24';
771
772 -- DELETE multicast and reserved address space
773 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '224.0.0.0/4';
774 DELETE FROM announcements WHERE family(network) = 4 AND network <<= '240.0.0.0/4';
775
776 -- Delete networks that are too small to be in the global routing table
777 DELETE FROM announcements WHERE family(network) = 6 AND masklen(network) > 48;
778 DELETE FROM announcements WHERE family(network) = 4 AND masklen(network) > 24;
779
780 -- Delete any non-public or reserved ASNs
781 DELETE FROM announcements WHERE NOT (
782 (autnum >= 1 AND autnum <= 23455)
783 OR
784 (autnum >= 23457 AND autnum <= 64495)
785 OR
786 (autnum >= 131072 AND autnum <= 4199999999)
787 );
788
789 -- Delete everything that we have not seen for 14 days
790 DELETE FROM announcements WHERE last_seen_at <= CURRENT_TIMESTAMP - INTERVAL '14 days';
791 """)
792
793 def _handle_update_announcements_from_bird(self, server):
794 # Pre-compile the regular expression for faster searching
795 route = re.compile(b"^\s(.+?)\s+.+?\[AS(.*?).\]$")
796
797 log.info("Requesting routing table from Bird (%s)" % server)
798
799 # Send command to list all routes
800 for line in self._bird_cmd(server, "show route"):
801 m = route.match(line)
802 if not m:
803 log.debug("Could not parse line: %s" % line.decode())
804 continue
805
806 # Fetch the extracted network and ASN
807 network, autnum = m.groups()
808
809 # Insert it into the database
810 self.db.execute("INSERT INTO announcements(network, autnum) \
811 VALUES(%s, %s) ON CONFLICT (network) DO \
812 UPDATE SET autnum = excluded.autnum, last_seen_at = CURRENT_TIMESTAMP",
813 network.decode(), autnum.decode(),
814 )
815
816 def _handle_update_announcements_from_telnet(self, server):
817 # Pre-compile regular expression for routes
818 route = re.compile(b"^\*[\s\>]i([^\s]+).+?(\d+)\si\r\n", re.MULTILINE|re.DOTALL)
819
820 with telnetlib.Telnet(server) as t:
821 # Enable debug mode
822 #if ns.debug:
823 # t.set_debuglevel(10)
824
825 # Wait for console greeting
826 greeting = t.read_until(b"> ", timeout=30)
827 if not greeting:
828 log.error("Could not get a console prompt")
829 return 1
830
831 # Disable pagination
832 t.write(b"terminal length 0\n")
833
834 # Wait for the prompt to return
835 t.read_until(b"> ")
836
837 # Fetch the routing tables
838 for protocol in ("ipv6", "ipv4"):
839 log.info("Requesting %s routing table" % protocol)
840
841 # Request the full unicast routing table
842 t.write(b"show bgp %s unicast\n" % protocol.encode())
843
844 # Read entire header which ends with "Path"
845 t.read_until(b"Path\r\n")
846
847 while True:
848 # Try reading a full entry
849 # Those might be broken across multiple lines but ends with i
850 line = t.read_until(b"i\r\n", timeout=5)
851 if not line:
852 break
853
854 # Show line for debugging
855 #log.debug(repr(line))
856
857 # Try finding a route in here
858 m = route.match(line)
859 if m:
860 network, autnum = m.groups()
861
862 # Convert network to string
863 network = network.decode()
864
865 # Append /24 for IPv4 addresses
866 if not "/" in network and not ":" in network:
867 network = "%s/24" % network
868
869 # Convert AS number to integer
870 autnum = int(autnum)
871
872 log.info("Found announcement for %s by %s" % (network, autnum))
873
874 self.db.execute("INSERT INTO announcements(network, autnum) \
875 VALUES(%s, %s) ON CONFLICT (network) DO \
876 UPDATE SET autnum = excluded.autnum, last_seen_at = CURRENT_TIMESTAMP",
877 network, autnum,
878 )
879
880 log.info("Finished reading the %s routing table" % protocol)
881
882 def _bird_cmd(self, socket_path, command):
883 # Connect to the socket
884 s = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM)
885 s.connect(socket_path)
886
887 # Allocate some buffer
888 buffer = b""
889
890 # Send the command
891 s.send(b"%s\n" % command.encode())
892
893 while True:
894 # Fill up the buffer
895 buffer += s.recv(4096)
896
897 while True:
898 # Search for the next newline
899 pos = buffer.find(b"\n")
900
901 # If we cannot find one, we go back and read more data
902 if pos <= 0:
903 break
904
905 # Cut after the newline character
906 pos += 1
907
908 # Split the line we want and keep the rest in buffer
909 line, buffer = buffer[:pos], buffer[pos:]
910
911 # Look for the end-of-output indicator
912 if line == b"0000 \n":
913 return
914
915 # Otherwise return the line
916 yield line
917
918 def handle_update_overrides(self, ns):
919 with self.db.transaction():
920 # Drop all data that we have
921 self.db.execute("""
922 TRUNCATE TABLE autnum_overrides;
923 TRUNCATE TABLE network_overrides;
924 """)
925
926 for file in ns.files:
927 log.info("Reading %s..." % file)
928
929 with open(file, "rb") as f:
930 for type, block in location.importer.read_blocks(f):
931 if type == "net":
932 network = block.get("net")
933 # Try to parse and normalise the network
934 try:
935 network = ipaddress.ip_network(network, strict=False)
936 except ValueError as e:
937 log.warning("Invalid IP network: %s: %s" % (network, e))
938 continue
939
940 # Prevent that we overwrite all networks
941 if network.prefixlen == 0:
942 log.warning("Skipping %s: You cannot overwrite default" % network)
943 continue
944
945 self.db.execute("""
946 INSERT INTO network_overrides(
947 network,
948 country,
949 is_anonymous_proxy,
950 is_satellite_provider,
951 is_anycast
952 ) VALUES (%s, %s, %s, %s, %s)
953 ON CONFLICT (network) DO NOTHING""",
954 "%s" % network,
955 block.get("country"),
956 self._parse_bool(block, "is-anonymous-proxy"),
957 self._parse_bool(block, "is-satellite-provider"),
958 self._parse_bool(block, "is-anycast"),
959 )
960
961 elif type == "aut-num":
962 autnum = block.get("aut-num")
963
964 # Check if AS number begins with "AS"
965 if not autnum.startswith("AS"):
966 log.warning("Invalid AS number: %s" % autnum)
967 continue
968
969 # Strip "AS"
970 autnum = autnum[2:]
971
972 self.db.execute("""
973 INSERT INTO autnum_overrides(
974 number,
975 name,
976 country,
977 is_anonymous_proxy,
978 is_satellite_provider,
979 is_anycast
980 ) VALUES(%s, %s, %s, %s, %s, %s)
981 ON CONFLICT DO NOTHING""",
982 autnum,
983 block.get("name"),
984 block.get("country"),
985 self._parse_bool(block, "is-anonymous-proxy"),
986 self._parse_bool(block, "is-satellite-provider"),
987 self._parse_bool(block, "is-anycast"),
988 )
989
990 else:
991 log.warning("Unsupport type: %s" % type)
992
993 @staticmethod
994 def _parse_bool(block, key):
995 val = block.get(key)
996
997 # There is no point to proceed when we got None
998 if val is None:
999 return
1000
1001 # Convert to lowercase
1002 val = val.lower()
1003
1004 # True
1005 if val in ("yes", "1"):
1006 return True
1007
1008 # False
1009 if val in ("no", "0"):
1010 return False
1011
1012 # Default to None
1013 return None
1014
1015 def handle_import_countries(self, ns):
1016 with self.db.transaction():
1017 # Drop all data that we have
1018 self.db.execute("TRUNCATE TABLE countries")
1019
1020 for file in ns.file:
1021 for line in file:
1022 line = line.rstrip()
1023
1024 # Ignore any comments
1025 if line.startswith("#"):
1026 continue
1027
1028 try:
1029 country_code, continent_code, name = line.split(maxsplit=2)
1030 except:
1031 log.warning("Could not parse line: %s" % line)
1032 continue
1033
1034 self.db.execute("INSERT INTO countries(country_code, name, continent_code) \
1035 VALUES(%s, %s, %s) ON CONFLICT DO NOTHING", country_code, name, continent_code)
1036
1037
1038 def split_line(line):
1039 key, colon, val = line.partition(":")
1040
1041 # Strip any excess space
1042 key = key.strip()
1043 val = val.strip()
1044
1045 return key, val
1046
1047 def main():
1048 # Run the command line interface
1049 c = CLI()
1050 c.run()
1051
1052 main()
A library to determine the location of someone on the Internet