Closed sensescape closed 6 years ago
These ones are the "good" networks that need to be normalized to better values:
CA:transcanada
when nat_name
= Trans-Canada Highway
CA:NB2
when shield_text is >= 100 but < 200.CA:NB3
when shield_text is >= 200.my:federal
with MY:federal
no:Fylkesvei
and no:Fylkesvei
with NO:Fylkesvei
pe:national
with PE:national
za:national
with ZA:national
za:regional
with ZA:regional
TH-roads
with TH:road
TH-network
with TH:network
ua:international
with UA:international
ua:national
with UA:national
ua:regional
with UA:regional
ua:territorial
with UA:territorial
TR-roads
with TR:road
TR-road
with TR:road
(think this is the same as above)TR-motorway
with TR:motorway
TR:DY
for refs that begin with D
.VN-roads
with VN:roads
I'll followup with "funky" per country refs that need to get networks...
These are out of the "funky" client side scene file hacks:
AR:RP
BR:BR
IT:B-road
when network is IT:*
and not IT:A-road
or ref begins with SS
? I think the SS should probably receive a different network that's aliased client side, what do you think? Similar for SR
and SP
.MY:expressway
network or similar. Scene file is confusing if all these are valid ref tests: E
, B
, J
, K
, P
.RO:DX
network (why not DN?)PE:PE
PE:{region_code}
. Partial list: AR, AY, CU, HV, IC, MO, TA, PU, LM. Shield text shouldn't include that prefix.RP
should return ES:AP
networkN
or N-
should return ES:N
network. Note some of these get shield text now, some don't.NA
should return ES:AA
network. (yellow)ES:{region_code}
. Partial list: AS, BU, C-, EI, M-, PA, TF and AG, AS, BI, C-, CM, Ma, MA, Me, EI, EX, FV, GC, GI, LO, LR, LZ, M-, PA, RM, TF, V-. Shield text shouldn't include that prefix.SG:Expressway
network?ZA:M
Not sure difference between M and MO.TR:DY
TR:O-roads
or drop the 0- based on OSM tag prevenance?VN:QL
network (eg: us route)VN:CT
network (eg: us interstate)VI-roads
should we normalize that? I think it should be client alias to VN:QL artwork.Ho Chi Minh Highway
should get VN:QL network instead of it's VI-roads network? maybe just change the data after confirming?CH:motorway
versus CH:national
versus CH:nationalstrasse
expressway
versus motorway
Australia M, B, C roads should keep their prefix (eg M1 instead of 1) in the shield_text calculation. See Melbourne.
Natural Earth e-road
shield_text needs to include the E (see zoom 7).
New Zealand State Highway isn't coming thru as NZ:SH
network:
Not all Canadian shields are drawing correctly: https://github.com/tangrams/tangram/issues/662
Not all Canadian shields have default network of CA
.
Chile Panamerican when network is CL:national
and shield_text is 5
than network should be CL:panamerican
instead.
Indonesia National should be ID:national
(alternatively ID:motorway
) but is just ID
. Test for motorway and assign?
Not all Norway shields have default network of NO
(very common). Similar to to https://github.com/tilezen/vector-datasource/issues/1491#issuecomment-414106052.
Also odd that the e-road network
's shield_text
values here include the space which is removed in other countries.
Thailand seems broken (just TH
with no more detail).
For Mexico, to disambiguate between the region of Mexico and the federal routes let's use MX:MX
instead of MX:MEX
(reserving the 3-char for the capital region).
India national highways are sometimes not receiving a shield_text value: https://www.openstreetmap.org/edit?way=180205706#map=17/20.92767/72.99029.
Blue Italian networks generally aren't being set:
Examples:
Goals:
IT:SR
Regional roadIT:SS
Strade Statali (same artwork as IT:SR) - State highwaysIT:SP
strada provinciale (same artwork as IT:SR) - provincial roadNetherlands roads could be set looking at the ref prefix, and we have artwork:
NL:A-road
national motorway route, ‘A’ prefixNL:N-road
provincial route, ‘N’ prefixNL:S-road
city route, ‘S’ prefixAustria is mostly working for AT:A-road
relations, but there are many roads not in relations starting with A ref that could be assigned that network value. Similarly B refs could be given AT:B-road
.
Belgium could use similar cleanup as Netherlands.
For the most part this all looks good. I took the last few QA comments and moved them into https://github.com/tilezen/vector-datasource/issues/1561 for the next milestone.
We should fix the Australia ref prefixes in this release, that's in https://github.com/tilezen/vector-datasource/issues/1562.
Followup to https://github.com/tilezen/vector-datasource/issues/1062... this will be running tally.
RNA002
002
, expectedA002
AR:provincial
), but they do have aRP
prefix in he ref.RP21
21
AR:provincial
addr:country
=AU
(and evenaddr:state
=QLD
), which could be used like we haveoperator
country assignments now.AU:N-highway
national highway (slowly being replaced with A-road and M-road)AU:A-road
AU:M-road
AU:B-road
AU:C-road
AU:N-route
for national routeAU:S-route
AU:Metro-road
AU:T-drive
[x] Brazil
BR
should sort first for Brazil, thenBR:Trans-Amazonian
, than the variousBR:*
state networks.BR:Trans-Amazonian
Trans-Amazonian Highwayhttps://www.openstreetmap.org/way/472168229
ref
=PE-050
source
=IBGE
BR:DF
BR:AC
BR:AL
BR:AP
BR:AM
BR:BA
BR:CE
BR:ES
BR:GO
BR:MA
BR:MT
BR:MS
BR:MG
BR:PA
BR:PB
BR:PR
BR:PE
BR:PI
BR:RJ
BR:RN
BR:RS
BR:RO
BR:RR
BR:SC
BR:SP
BR:SE
BR:TO
CA:transcanada
highway network should sort beforeCA:yellowhead
, before otherCA:*
https://www.openstreetmap.org/way/95904761 https://www.openstreetmap.org/relation/1121378 https://www.openstreetmap.org/relation/1307245CN-expressways
network should sort beforeCN-expressways-regional
, beforeAsianHighway
, before otherCN-roads
. (AsianHighway is like Europe E-roads at end of the lists, same is true for Russia for instance.)G42
5
, expectedG42
AsianHighway
, expectedCN-expressways
CN-expressways-regional
is a subset ofCN-expressways
network tagged features when the ref begins with anS
S20
S20
CN-expressways
, expectedCN-expressways-regional
E05
, expectedE5
e-road
FR:D-roads
network when no relation is presentD *
andD*
roads that collide without a spatial check.D983
FR:D-road
FR:##:D-road
withFR:D-road
networkFR:01:D-road
,FR:52:D-road
,FR:51:D-road
,FR:55:D-road
,FR:08:D-road
so we could do a replace, but a general pattern match would be better.D17
FR:01:D-road
, expectedFR:D-road
operator
-based fallbacknetwork
logic to dedup &/or not trigger unless no other network, and sort order should beFR:A-road
,FR:N-road
,FR:D-road
,FR
,e-road
.DIRNO
E03
,A84
e-road
, expectedFR:A-road
e-road, FR, e-road, FR:A-road
, expectedFR:A-road, e-road, e-road
(noFR
at all)BAB
network roads should receive aDE:BAB
network value instead to enforce country code prefixes.A1
BAB
, expectedDE:BAB
Via Solutions Südwest
to map to networkDE:BAB
.Via Solutions Südwest
A5
DE:BAB
autobahnplus A8 GmbH
to map to networkDE:BAB
.autobahnplus A8 GmbH
A8
DE:BAB
Bundesstraßen*
(federal road) network value roads should receive aDE:BS
network value instead of genericDE
based on operator fallback.Bundesrepublik Deutschland
B6
DE:BS
(or at leastDE
with old logic)Bundesstraßen*
(federal road) without a network should receive aDE:BS
network value instead based in theB*
ref pattern match, and should strip the space inB #
B 9
should be:B9
DE:BS
Landesstraßen*
(state/region road) network value roads should receive aDE:Landesstraßen
network value instead.Landesstraßen NRW
,Landesstraßen Niedersachsen
,Landesstraßen 2. Ordnung im Saarland
,Landesstraßen Brandenburg
,Landesstraßen 1. Ordnung im Saarland
, but we should use match on general pattern.L84
Landesstraßen NRW
, expectedDE:LS
DE:LS
network value based onL*
ref pattern matching (limited to Germany bbox), more basic example.L39
null
, expectedDE:LS
DE:KS
network value based on network value subsitution.K203
Kreisstraßen Hildesheim
, expectedDE:KS
DE:KS
network value based onK*
ref pattern matching (limited to Germany bbox, watch out for Malaysia overlap)K40
DE:KS
JP:national
network roads often missing but can be added looking at way name.国道163号
(国道
means national route with the postifx号
indicating that the number is a route number, but sometimes there's additional postfixes)163
null
, expectedJP:national
jp:prefectural
network should be replaced withJP:prefectural
305
jp:prefectural
, expectedJP:prefectural
MEX
(note space), no networkMEX 15D
null
, should be:15D
null
, expectedMX:MEX
MEX
(note space), no networkMEX610
null
, should be:610
null
, expectedMX:MEX
MEX
need to be prefixed with country code:MX:MEX
(substitution)120
MEX
, expectedMX:MEX
KR:national
relations (perhaps those should beKR:expressway
though?), but there should be many...KR:expressway
for https://en.wikipedia.org/wiki/Expressways_in_South_Korea added in capitalKR:national
https://en.wikipedia.org/wiki/National_highways_of_South_Korea added in capitalKR:local
https://en.wikipedia.org/wiki/Local_highways_of_South_Korea added in JejuAH
network should substitute out forAsianHighway
:1
AH
, expectedAsianHighway
GB
network based onHighways England
operator should substitute out forGB:M-road
when ref starts with M:M25
GB
, expectedGB:M-road
GB
network based onHighways England
operator should substitute out forGB:A-road
when ref starts with A:A3
GB
, expectedGB:A-road
GB:A-road-green
??? when ref starts with A in the bounding box of the United Kingodm, but there seem to be 2 flavours of A-road, namely GREEN whenhighway=motorway
???:A30
null
, expectedGB:A-road-green
GB:A-road-white
when ref starts with A in the bounding box of the United Kingodm, but there seem to be 2 flavours of A-road, namely WHITE when not highway=motorway ???:A312
null
, expectedGB:A-road-white
GR:national
networks need to strip off the leadingEO
on the shield_text.ΕΟ8
should be:8
GR:national
GR
networks (based on operator) should sub outGR:motorway
(another operator overriding network problem?)A6
GR
, should be:GR:motorway