Item for wishlist. Brouter “River” improvement.

[P Wat]

Brouter “River” could be a much more useful tool, if it was not for a significant anomaly.

 

Scenario –

When planning certain waterway trips, it is useful to check nominal river distances between potential rest/night/logistical stops.

Unfortunately, the routes generated by Brouter “River” are sometimes not navigable.  This has been confirmed by app testing of familiar waters.

A simple example is routing over weirs (un-navigable) when there is a lock (navigable) nearby.

Of course, in the real world one sees and takes the correct, non-fatal route!

But when planning, cases arise where Brouter River routes via weirs, and thence sometimes via other navigationally-impractical narrow streams, low bridges, etc, even places where a canoe would not go.  The viable route is sometimes relatively distant and longer.

 

Possible solution –

Can Brouter “River” be adjusted:-

(a) to avoid routing over a weir?

(b) to route via a lock when one is present as the obvious alternative to the weir?

 

Examples of navigable and un-navigable routes will be willingly provided if required.

Paul W - (Many years of nautical experience, but little skill in coding)

[Poutnik]

I can help it tailoring the river profile, applying node \aand way tags
waterway*, together with your recent "river navigation" topic in OSMand
mail-list.

I see there will be needed a distinction, what boats are targeted by the
profile, perhaps by "portable_boat" logical flag.

As for big sail boats or motorboats, weirs are not routable.
OTOH, for small portable boats like canoe, routing across weirs
is OK. Small ones can be overcome, big ones can be carried around.
As carrying a small boat is alternative motion mode for canoe,
same as leading a bicycle for bicycle routing.

I also see a trouble is fact, that while roads can be considered as
curved 1-dimensional objects, waterways are rather 2-dimensional
objects, with feature both along and across it.this may bring
difficulties for the routing algorithm. As seeing local 2D context
can be very difficult.

As BRouter does not see the real world, but OSM world.
And not all the OSM world, but a subset of its attributes of nodes, ways
and relations.

I do not see the BRouter use as replacement of route planning
over the waterway maps and water guides. I see it rather as a useful
tool to joint 2 manual waypoint where you do know the water way is routable.

On 10/04/2016 10:07 AM, 'P Wat' via OSM Android bikerouting wrote:
> Brouter “River” could be a much more useful tool, if it was not for a
> significant anomaly.
>
>
>
> Scenario –
>
> When planning certain waterway trips, it is useful to check nominal
> river distances between potential rest/night/logistical stops.
>
> Unfortunately, the routes generated by Brouter “River” are sometimes not
> navigable. This has been confirmed by app testing of familiar waters.
>
> A simple example is routing over weirs (un-navigable) when there is a
> lock (navigable) nearby.
>

> Of course, in the real world one sees and takes the correct,non-fatal route!

>
> But when planning, cases arise where Brouter River routes via weirs, and
> thence sometimes via other navigationally-impractical narrow streams,
> low bridges, etc, even places where a canoe would not go. The viable
> route is sometimes relatively distant and longer.
>
>
>
> Possible solution –
>
> Can Brouter “River” be adjusted:-
>
> (a) to avoid routing over a weir?
>
> (b) to route via a lock when one is present as the obvious alternative
> to the weir?
>
>
>
> Examples of navigable and un-navigable routes will be willingly provided
> if required.
>
> Paul W - (Many years of nautical experience, but little skill in coding)
>

> --
> You received this message because you are subscribed to the Google
> Groups "OSM Android bikerouting" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to osm-android-biker...@googlegroups.com
> <mailto:osm-android-biker...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

[Poutnik]

On 10/04/2016 10:07 AM, 'P Wat' via OSM Android bikerouting wrote:

> Can Brouter “River” be adjusted:-
>
> (a) to avoid routing over a weir?
>
> (b) to route via a lock when one is present as the obvious alternative
> to the weir?
>
> Examples of navigable and un-navigable routes will be willingly provided
> if required.
>
> Paul W - (Many years of nautical experience, but little skill in coding)

The scenario examples as Brouter-web / OSM location links
or location names would be welcome.

Especially where the 2D placement of river features is challenging.

Poutnik

[Poutnik]

Hi Paul,

you may want to try this profile ( joint the line if broken ):

https://raw.githubusercontent.com/poutnikl/Brouter-profiles/master/river-poutnik.brf

It is on alfa stage, so expect anything from a boat blowing up to
sinking down.

Be aware of many incorrect OSM mapping,
leading to e.g Brouter not seeing a dam across Vltava river.

On 10/04/2016 10:07 AM, 'P Wat' via OSM Android bikerouting wrote:

> Brouter “River” could be a much more useful tool, if it was not for a
> significant anomaly.
>
>
>
> Scenario –
>
> When planning certain waterway trips, it is useful to check nominal
> river distances between potential rest/night/logistical stops.
>
> Unfortunately, the routes generated by Brouter “River” are sometimes not
> navigable. This has been confirmed by app testing of familiar waters.
>
> A simple example is routing over weirs (un-navigable) when there is a
> lock (navigable) nearby.
>

> Of course, in the real world one sees and takes the correct,non-fatal route!

>
> But when planning, cases arise where Brouter River routes via weirs, and
> thence sometimes via other navigationally-impractical narrow streams,
> low bridges, etc, even places where a canoe would not go. The viable
> route is sometimes relatively distant and longer.
>
>
>
> Possible solution –
>
> Can Brouter “River” be adjusted:-
>
> (a) to avoid routing over a weir?
>
> (b) to route via a lock when one is present as the obvious alternative
> to the weir?
>
>
>
> Examples of navigable and un-navigable routes will be willingly provided
> if required.
>
> Paul W - (Many years of nautical experience, but little skill in coding)
>

[P Wat]

Hi Poutnik
You've worked pretty fast, thank you.
Your link

https://raw.githubusercontent.com/poutnikl/Brouter-profiles/master/river-poutnik.brf
opens OK to show about 80 lines of code. - Correct?

Can you remind me what to do to make it operate?  ie is the following correct? -
1 Paste the entire page of code into a text file.
2 Save it with a suitable name eg River2.
3 Change the file extensioin to BRF.
4 Place it in the device in the folder called "Profile2" along with all the other profiles (fast-bike, trekking, etc)
5 Run Brouter in the usual way and select the new profile - (Brouter App, select profile, Server mode, "OK", back to OsmAnd)
Thanks
Paul W

[Poutnik]

On 10/04/2016 06:26 PM, 'P Wat' via OSM Android bikerouting wrote:
> Hi Poutnik
> You've worked pretty fast, thank you.
> Your link

> opens OK to show about 80 lines of code. - Correct?

About 90.

>
> Can you remind me what to do to make it operate? ie is the following
> correct? -
> 1 Paste the entire page of code into a text file.
> 2 Save it with a suitable name eg River2.
> 3 Change the file extensioin to BRF.
> 4 Place it in the device in the folder called "Profile2" along with all
> the other profiles (fast-bike, trekking, etc)
> 5 Run Brouter in the usual way and select the new profile - (Brouter
> App, select profile, Server mode, "OK", back to OsmAnd)

Yes, like that.

But I would rather postpone this usage in Brouter application
until it is verified
working well within available OSM data.

For now, I highly recommend rather testing it with

http://brouter.de/brouter-web/

by pasting and uploading the code to the left bottom text field.

You can involve in the testing the web features,
and also verify and update master OSM data
in
https://www.openstreetmap.org

Poutnik
https://github.com/poutnikl/Brouter-profiles/wiki

[Poutnik]

Be aware of flag
assign portable_boat true

Change it to false for a big boat.

[P Wat]

This really is a quantum leap forward (especially when one remembers to assign Big Boat = False !), but there seems to be at least on anomaly.
Several examples have been found where it manifests itself.  Here is one case-
Go to 51.43859, -0.53765
See part of the River Thamnes, alongside "Bell Weir Lock"?
In Brouter web Client, the custom profile (with assign Big Boat = False) routes over the weir and via the river to the north of the lock. (See uploaded graphic)
In OSM it is seems that the weir is correctly identified.  (See uploaded graphic)
What do you think?

On the bright side, another case of previously false routing has been found to have been resolved by your custom profile. - Thanks.
You'll probably quickly find the error.
Paul W

[Poutnik]

You must mean by "assign Big Boat = False"

assign portable_boat true

Fixed version is be uploaded as v1.0.1

But the error was in bad evaluated boat=no near the weir.
Weir itself is not mapped properly.
 
As it did not occur in the wrong route
at the BRouter-web
either among ways, either among nodes.

Dne 04/10/2016 v 23:11 'P Wat' via OSM Android bikerouting napsal(a):

--
You received this message because you are subscribed to the Google Groups "OSM Android bikerouting" group.

To unsubscribe from this group and stop receiving emails from it, send an email to osm-android-biker...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

-- 
Poutnik ( The Wanderer )

My Brouter profiles 
https://github.com/poutnikl/Brouter-profiles/wiki

[Poutnik]

As the 2 crossing OSM ways waterway=weir and waterway=river
DO NOT intersect. So routing the waterway=river
does not notice there is anything like waterway=weir,
similarly as if 2 roads cross each other at different levels.

Dne 04/10/2016 v 23:57 Poutnik napsal(a):

But the error was in bad evaluated boat=no near the weir.

Weir itself is not mapped properly.
 
As it did not occur in the wrong route
at the BRouter-web
either among ways, either among nodes.

[Poutnik]

There is open question what is the correct route for small portable boats.
I have doubts if they are allowed to go through the lock.
Perhaps the proper route for portable boats would be to carry them
on the bank along the weir and the section with boat=no.

Dne 04/10/2016 v 23:11 'P Wat' via OSM Android bikerouting napsal(a):

[P Wat]

Answering your question "what is the correct route for small portable boats"? -
The convention on the River Thames seems to be to use signposts to encourage canoeists to "Portage" around locks, though one does sometimes see canoes in locks.
The signposted route varies according to the topography of the lock, weir and other local features.
It is, for example, sometimes along the adjacent towpath (footway alongside the waterway originally evolved for horses or people to tow the barges - now a public footpath).

[P Wat]

Testing again - carefully
Amended line 3 to read "assign portable_boat false".
Uploaded to Web client.  Located same lock & weir.  Entered start & end points. Set profile "custom".
Aha!  Routing is correct for "large" boat.
Looks like I owe you an apology - Maybe yesterday I messed up the "Assign....."

Your code is better than it seemed!
I'll extend my testing, then feed back.
Thanks
Paul W
=====================

On Tuesday, October 4, 2016 at 10:57:28 PM UTC+1, Poutnik wrote:

You must mean by "assign Big Boat = False"

assign portable_boat true

Fixed version is be uploaded as v1.0.1

But the error was in bad evaluated boat=no near the weir.
Weir itself is not mapped properly.
 
As it did not occur in the wrong route
at the BRouter-web
either among ways, either among nodes.

Dne 04/10/2016 v 23:11 'P Wat' via OSM Android bikerouting napsal(a):

This really is a quantum leap forward (especially when one remembers to assign Big Boat = False !), but there seems to be at least on anomaly.
Several examples have been found where it manifests itself.  Here is one case-
Go to 51.43859, -0.53765
See part of the River Thamnes, alongside "Bell Weir Lock"?
In Brouter web Client, the custom profile (with assign Big Boat = False) routes over the weir and via the river to the north of the lock. (See uploaded graphic)
In OSM it is seems that the weir is correctly identified.  (See uploaded graphic)
What do you think?

On the bright side, another case of previously false routing has been found to have been resolved by your custom profile. - Thanks.
You'll probably quickly find the error.
Paul W

--
You received this message because you are subscribed to the Google Groups "OSM Android bikerouting" group.

To unsubscribe from this group and stop receiving emails from it, send an email to osm-android-bikerouting+unsub...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[poutnik]

Be aware the major problem will be the *proper* osm mapping of river features. As they usually are placed in the map correctly, but they are often misconnected a/o misrelated.
A waterway=weir not connected with a waterway=river means no weir on the river.
A bridge not connected with river means no bridge over the river. And if there is no waterway attribute about max height, there is no hint the boat may not pass.

5. října 2016 9:51:37 CEST, 'P Wat' via OSM Android bikerouting <osm-android...@googlegroups.com> napsal:

--
Sent from my phone via Android email client K-9.
Please, forgive my brevity.

[P Wat]

I think I understand the logic of what you say here.  If so, correcting/CONNECTING large numbers of weirs to their rivers would be a task to far.

Meanwhile, I've tested the entire non-tidal Thames. Feedback coming soon for your perusal.
PW
 ===========

[P Wat]

Testing again - carefully
Line 3 set to read "assign portable_boat false". Ie Routing for non-portable boat.
Examining a 107km non-tidal length of River Thames.

Locks checked:-

Mapledurham = OK

Caversham = OK

Sonning = Failed – Goes through fish channel? – It would be worth discovering why.

Shiplake = OK

Marsh = OK

Hambleden = OK

Hurley = OK

Temple = OK

Marlow = OK

Cookham = OK

Boulters = OK

Bray = OK

Boveney = OK

Romney = OK

Old Windsor = OK

Bell Weir = OK

Penton Hook = OK

Shepperton = Failed – Goes long route over weir about 500 metres to south. – It would be worth discovering why.

Sunbury = Avoids weir OK, but routes via non-functioning old lock. – Minor navigational issue.

Molesey = OK

Teddington = OK

Richmond = Contentious – This is a moveable weir lifted at high tide to allow navigation – The lock is a few metres to the north.  Sig= nposting is clear.  Navigational distance is unaffected.  Coding adjustment is not recommended at this stage. Any discrepancy is negligible.

 

New start & finish points set:-

Locks checked:-.

St John’s = OK

Buscot = Failed - Routed vie weir to the south

Grafton = OK

Radcot = OK

Rushey = OK

Shifford = OK

Northmoor = Failed - Routed vie weir to the south-east.

Pinkhill = OK

Eynsham = OK

King’s = OK

Godstow = OK

Osney = Failed - Routed vie weir to the south-west.

Iffley = Fairly significant failure – Routed via minor channel non-navigable.

Sandford = Failed – Avoided several weirs then routed over small weir and millstream a few meters to the east.

Abingdon = OK

Culham = OK

Clifton = OK

Day’s = OK

Benson  = OK

Cleeve = Failed = Routed via weir to the east.

Goring = OK

Whitchurch = OK.

 

That’s all 44 locks on the non-tidal Thames (plus Richmond).

Paul W

[Poutnik]

I will have time for deeper review probably not sooner
than in the evening. just few notes.

What is esseintial is , what is displayed in tha data tab
( the link just about the uploaded code ), optionally
opened in Ms Excel or LibreOffice as a table.

What is not there, either
is not in OSM
OR
is in OSM, but is properly connected a/o related
OR
is in OSM, properly connected a/o related
BUt is not exposed for Brouter profiles
in Brouter RD5 routing data and lookups.dat

Important note is, that Brouter does not expose
the tagging for waterway signs.

I have fixed fish_pass treatment, but even without it
the past should be forbidden, so I guess it was mapped incorrectly.

> --
> You received this message because you are subscribed to the Google
> Groups "OSM Android bikerouting" group.
> To unsubscribe from this group and stop receiving emails from it, send

> an email to osm-android-biker...@googlegroups.com
> <mailto:osm-android-biker...@googlegroups.com>.

[P Wat]

Poutnik - Thank you for your ongoing interest and effort.
I had a look at Sunbury Lock - The OSM has data missing.  The small weir, which also contains the fish pass (if I remember correctly) is not even shown as a weir.  For the moent I've added a note to OSM.
I'll look at your other comments, and also review another waterway known to me.
PW

> an email to osm-android-bikerouting+unsub...@googlegroups.com
> <mailto:osm-android-bikerouting+unsubscribe@googlegroups.com>.

[Poutnik]

I have just tried the routing the Thames river
for portable boats along the last cca 105 km to the sea.

For portables, I have set weirs and boat=no as formally pass-able, with
one-time penalty for carrying/towing around.

I have not noticed the waterway NEVER encountered any weir.
( See data link or download CSV file for the route data.

For big boats, when the routing avoid weirs,
it always does so because of boat=no or motorboat=no.

If there is no such restriction,
it happily crosses over weirs,
as it is not aware of them at all.

Poutnik

> > an email to osm-android-biker...@googlegroups.com
> > <mailto:osm-android-biker...@googlegroups.com>.

> > For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.

>
> --
> You received this message because you are subscribed to the Google
> Groups "OSM Android bikerouting" group.
> To unsubscribe from this group and stop receiving emails from it, send

> an email to osm-android-biker...@googlegroups.com
> <mailto:osm-android-biker...@googlegroups.com>.

[P Wat]

Thanks Poutnik, Sounds hopeful.
I'll run my own check
Can you confirm which is the correct version of the file to use?
PW

>     > an email to osm-android-bikerouting+unsub...@googlegroups.com
>     > <mailto:osm-android-bikerouting+unsubscribe@googlegroups.com>.

>     > For more options, visit https://groups.google.com/d/optout
>     <https://groups.google.com/d/optout>.
>
> --
> You received this message because you are subscribed to the Google
> Groups "OSM Android bikerouting" group.
> To unsubscribe from this group and stop receiving emails from it, send

> an email to osm-android-bikerouting+unsub...@googlegroups.com
> <mailto:osm-android-bikerouting+unsubscribe@googlegroups.com>.

[poutnik]

Correct version is always at the same location, being updated over.
Currently V1.0.3, giving higher priority to river.

6. října 2016 16:20:02 CEST, 'P Wat' via OSM Android bikerouting <osm-android...@googlegroups.com> napsal: