From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter =?utf-8?q?M=C3=BCller?= To: location@lists.ipfire.org Subject: Re: [PATCH v2] location-importer.in: track original countries as well Date: Wed, 26 May 2021 20:11:57 +0200 Message-ID: In-Reply-To: <20210522125758.28770-1-peter.mueller@ipfire.org> MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="===============4705939918598463732==" List-Id: --===============4705939918598463732== Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Hello Michael, this one should be ready for merging as well. Or do you want me to rebase it against current master branch? Thanks, and best regards, Peter M=C3=BCller > This helps us to determine how many network objects have more than one > country set, and what their original country code set looked like. >=20 > The second version of this patch uses ALTER TABLE to add the column for > original countries, preventing existing SQL setups from breaking. >=20 > Signed-off-by: Peter M=C3=BCller > --- > src/python/location-importer.in | 52 ++++++++++++++++++++------------- > 1 file changed, 32 insertions(+), 20 deletions(-) >=20 > diff --git a/src/python/location-importer.in b/src/python/location-importer= .in > index e5f55af..c3f908a 100644 > --- a/src/python/location-importer.in > +++ b/src/python/location-importer.in > @@ -165,6 +165,7 @@ class CLI(object): > =20 > -- networks > CREATE TABLE IF NOT EXISTS networks(network inet, country text); > + ALTER TABLE networks ADD COLUMN IF NOT EXISTS original_countries text[= ]; > CREATE UNIQUE INDEX IF NOT EXISTS networks_network ON networks(network= ); > CREATE INDEX IF NOT EXISTS networks_family ON networks USING BTREE(fam= ily(network)); > CREATE INDEX IF NOT EXISTS networks_search ON networks USING GIST(netw= ork inet_ops); > @@ -377,7 +378,7 @@ class CLI(object): > ON COMMIT DROP; > CREATE UNIQUE INDEX _organizations_handle ON _organizations(handle); > =20 > - CREATE TEMPORARY TABLE _rirdata(network inet NOT NULL, country text NO= T NULL) > + CREATE TEMPORARY TABLE _rirdata(network inet NOT NULL, country text NO= T NULL, original_countries text[] NOT NULL) > ON COMMIT DROP; > CREATE INDEX _rirdata_search ON _rirdata USING BTREE(family(network), = masklen(network)); > CREATE UNIQUE INDEX _rirdata_network ON _rirdata(network); > @@ -407,8 +408,8 @@ class CLI(object): > for family in (row.family for row in families): > smallest =3D self.db.get("SELECT MIN(masklen(network)) AS prefix FROM = _rirdata WHERE family(network) =3D %s", family) > =20 > - self.db.execute("INSERT INTO networks(network, country) \ > - SELECT network, country FROM _rirdata WHERE masklen(network) =3D %s A= ND family(network) =3D %s", smallest.prefix, family) > + self.db.execute("INSERT INTO networks(network, country, original_count= ries) \ > + SELECT network, country, original_countries FROM _rirdata WHERE maskl= en(network) =3D %s AND family(network) =3D %s", smallest.prefix, family) > =20 > # ... determine any other prefixes for this network family, ... > prefixes =3D self.db.query("SELECT DISTINCT masklen(network) AS prefix= FROM _rirdata \ > @@ -421,7 +422,8 @@ class CLI(object): > WITH candidates AS ( > SELECT > _rirdata.network, > - _rirdata.country > + _rirdata.country, > + _rirdata.original_countries > FROM > _rirdata > WHERE > @@ -434,6 +436,7 @@ class CLI(object): > DISTINCT ON (c.network) > c.network, > c.country, > + c.original_countries, > masklen(networks.network), > networks.country AS parent_country > FROM > @@ -447,10 +450,11 @@ class CLI(object): > masklen(networks.network) DESC NULLS LAST > ) > INSERT INTO > - networks(network, country) > + networks(network, country, original_countries) > SELECT > network, > - country > + country, > + original_countries > FROM > filtered > WHERE > @@ -617,28 +621,36 @@ class CLI(object): > inetnum[key] =3D [ipaddress.ip_network(val, strict=3DFalse)] > =20 > elif key =3D=3D "country": > - inetnum[key] =3D val.upper() > + # Catch RIR data objects with more than one country code... > + if not key in inetnum.keys(): > + inetnum[key] =3D [] > + else: > + if val.upper() in inetnum.get("country"): > + # ... but keep this list distinct... > + continue > + > + inetnum[key].append(val.upper()) > =20 > # Skip empty objects > if not inetnum or not "country" in inetnum: > return > =20 > + # Prepare skipping objects with unknown country codes... > + invalidcountries =3D [singlecountry for singlecountry in inetnum.get("co= untry") if singlecountry not in validcountries] > + > # Iterate through all networks enumerated from above, check them for pla= usibility and insert > # them into the database, if _check_parsed_network() succeeded > for single_network in inetnum.get("inet6num") or inetnum.get("inetnum"): > if self._check_parsed_network(single_network): > - > - # Skip objects with unknown country codes - to avoid log spam for inva= lid or too small > - # networks, this check is - kinda ugly - done at this point > - if validcountries and inetnum.get("country") not in validcountries: > - log.warning("Skipping network with bogus country '%s': %s" % \ > - (inetnum.get("country"), inetnum.get("inet6num") or inetnum.get("ine= tnum"))) > + # Skip objects with unknown country codes if they are valid to avoid l= og spam... > + if validcountries and invalidcountries: > + log.warning("Skipping network with bogus countr(y|ies) %s (original c= ountries: %s): %s" % \ > + (invalidcountries, inetnum.get("country"), inetnum.get("inet6num") o= r inetnum.get("inetnum"))) > break > =20 > - # Everything is fine here, run INSERT statement... > - self.db.execute("INSERT INTO _rirdata(network, country) \ > - VALUES(%s, %s) ON CONFLICT (network) DO UPDATE SET country =3D exclud= ed.country", > - "%s" % single_network, inetnum.get("country"), > + self.db.execute("INSERT INTO _rirdata(network, country, original_count= ries) \ > + VALUES(%s, %s, %s) ON CONFLICT (network) DO UPDATE SET country =3D ex= cluded.country", > + "%s" % single_network, inetnum.get("country")[0], inetnum.get("countr= y"), > ) > =20 > def _parse_org_block(self, block): > @@ -729,10 +741,10 @@ class CLI(object): > if not self._check_parsed_network(network): > return > =20 > - self.db.execute("INSERT INTO networks(network, country) \ > - VALUES(%s, %s) ON CONFLICT (network) DO \ > + self.db.execute("INSERT INTO networks(network, country, original_countri= es) \ > + VALUES(%s, %s, %s) ON CONFLICT (network) DO \ > UPDATE SET country =3D excluded.country", > - "%s" % network, country, > + "%s" % network, country, [country], > ) > =20 > def handle_update_announcements(self, ns): >=20 --===============4705939918598463732==--