Re: [OSM-dev] time osm2pgsql

2009-05-15 Thread Lennard
Sam Mor wrote:
 and now after 7 hours still the same:
 Going over pending relations
 node cache: stored: 104740975(31.71%), storage efficiency: 99.89%, hit rate: 
 31.36%

Yes, it can take a long time. Did you tune your postgresql according to 
http://wiki.openstreetmap.org/wiki/Mapnik/PostGIS#Tuning_the_database ?

An extra line in osm2pgsql, where it tells you it's doing the ANALYZE, 
could be handy as well. When I first started using osm2pgsql, I've been 
wondering too what it was doing, and if it might be hanging.

-- 
Lennard

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] time osm2pgsql

2009-05-15 Thread Frederik Ramm
Hi,

Lennard wrote:
 and now after 7 hours still the same:
 Going over pending relations
 node cache: stored: 104740975(31.71%), storage efficiency: 99.89%, hit rate: 
 31.36%
 
 An extra line in osm2pgsql, where it tells you it's doing the ANALYZE, 
 could be handy as well. When I first started using osm2pgsql, I've been 
 wondering too what it was doing, and if it might be hanging.

Sure it's ANALYZE and not CREATE INDEX (which I'd expect to run for at 
least a day on a full planet)?

ps auxw|grep postgis

will tell you what the database thinks it's doing.

Bye
Frederik


___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] Can SQLite3 handle OSM 150G data file?

2009-05-15 Thread Tomas Kolda
Did you try my version of import? I have no responses, so maybe not :). 
I did not tried complete planet, but indexes was created quite fast. 
30seconds on 150MB osm file.

I think that you should make a chart (xml size-time to create index) 
and than you can see complexity of creating indexes in sqlite based on 
OSM size. I think that it will be O(n log(n)), but maybe not.

If I will have time, I can try to convert planet using my tool.

T

Kelly Jones napsal(a):
 Thanks to everyone who replied.

 OK, loading all the nodes for OSM doesn't take much time, but INDEXing
 the fields takes forever (days).

 The first time I tried this, I pre-created the indexes before loading
 the data. That's why the load took so long.

 This time, I loaded the data first (fairly quick) and then created the
 indexes. It's now been several days and the indexes are still being
 created.

 Thoughts?

   


___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] time osm2pgsql

2009-05-15 Thread Ldp
Frederik Ramm wrote:

 Sure it's ANALYZE and not CREATE INDEX (which I'd expect to run for at 
 least a day on a full planet)?

No, I'm not sure. That's why some additional osm2pgsql output would be 
nice to have.

-- 
Lennard

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


[OSM-dev] Time Confusion

2009-05-15 Thread Frederik Ramm
Hi,

I notice that there is a mismatch between timestamps given in the 
planet file and timestamps as reported by the API. The planet file 
contains node #2 like this:

   node id=2 lat=50.1359444 lon=8.3013034 
timestamp=2009-04-14T14:42:58Z
  version=8 changeset=524633 user=woodpeck uid=5164

But the API returns

   node id=2 lat=50.1359444 lon=8.3013034 version=8 
changeset=524633 user=woodpeck uid=5164 visible=true 
timestamp=2009-04-14T15:42:58Z

Who is right, and why the difference?

Bye
Frederik

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] Time Confusion

2009-05-15 Thread Tom Hughes
Frederik Ramm wrote:

 I notice that there is a mismatch between timestamps given in the 
 planet file and timestamps as reported by the API. The planet file 
 contains node #2 like this:
 
node id=2 lat=50.1359444 lon=8.3013034 
 timestamp=2009-04-14T14:42:58Z
   version=8 changeset=524633 user=woodpeck uid=5164
 
 But the API returns
 
node id=2 lat=50.1359444 lon=8.3013034 version=8 
 changeset=524633 user=woodpeck uid=5164 visible=true 
 timestamp=2009-04-14T15:42:58Z
 
 Who is right, and why the difference?

The API is right.

Before the move to 0.6 the database was in UK local time, but it is now 
held in UTC no matter what time of year the edit was made.

The database has 2009-04-14 15:42:58 as the timestamp for that edit 
which is what the API is returning (as a Zulu time) so the API seems to 
be correct.

My guess is that the planet dumper is still assuming the DB is in UK 
local time and is converting to UTC for output.

Tom

-- 
Tom Hughes (t...@compton.nu)
http://www.compton.nu/

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] Time Confusion

2009-05-15 Thread Matt Amos
On Fri, May 15, 2009 at 4:51 PM, Tom Hughes t...@compton.nu wrote:
 My guess is that the planet dumper is still assuming the DB is in UK
 local time and is converting to UTC for output.

d'oh. fixed now.

cheers,

matt

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


[OSM-dev] dev servers down

2009-05-15 Thread Stefano Salvador
Hi,

It seems that the last commit on dev server (api06 and new06) has corrupted 
database.yml :-(

could someone fix this ?

Bye,

Stefano

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] dev servers down

2009-05-15 Thread Thomas Wood
2009/5/15 Stefano Salvador stefano.salva...@gmail.com:
 Hi,

 It seems that the last commit on dev server (api06 and new06) has corrupted
 database.yml :-(

 could someone fix this ?

 Bye,

 Stefano

Now fixed, thanks for the report, I need to fix the cron job to handle
conflicts nicely, I thought I'd nailed it, but obviously not...

-- 
Regards,
Thomas Wood
(Edgemaster)

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


[josm-dev] WMS plugin 'overlap' patch

2009-05-15 Thread Radomír Černoch
Hi all!

in the Czech republic we have a problem with one WMS server, which
tends to wipe the WMS tiles near its borders.

To solve this I've created a patch allowing WMS plugin in JOSM to
download tiles, whose sides do not align perfectly. Instead WMS can
add a certain percentage of overlap between neighbouring tiles so that
the missing information in one tile is 'overlayed' by the neigbouring
one (if server supports alpha channel). This feature can be disabled
(which is the default) and the amount of overlap can be set in
preferences.

The patch is included in the attachment as I do not have access to the
SVN. Could I ask for including it into upstream?

Yours sincerely,
Radomir Cernoch

PS: Can I also ask for access to JOSM SVN? What should I do for it? I
have developed another plugin, which could be included.

-- 
Radomir Cernoch
+44 750 708 8293 / +420 607 282 031
Email, Jabber: radomir.cern...@gmail.com
Index: src/wmsplugin/WMSLayer.java
===
--- src/wmsplugin/WMSLayer.java	(revision 15029)
+++ src/wmsplugin/WMSLayer.java	(working copy)
@@ -83,6 +83,7 @@
 executor = Executors.newFixedThreadPool(3);
 }
 
+@Override
 public void destroy() {
 try { 
 executor.shutdown();  
@@ -175,6 +176,7 @@
 for(int x = bminx; xbmaxx; ++x)
 for(int y = bminy; ybmaxy; ++y){
 GeorefImage img = images[modulo(x,dax)][modulo(y,day)];
+g.drawRect(x, y, dax, bminy);
 if(!img.paint(g, mv, dx, dy)  !img.downloadingStarted){
 img.downloadingStarted = true;
 img.image = null;
Index: src/wmsplugin/Grabber.java
===
--- src/wmsplugin/Grabber.java	(revision 15029)
+++ src/wmsplugin/Grabber.java	(working copy)
@@ -11,6 +11,7 @@
 import java.awt.Color;
 import java.awt.Font;
 import javax.swing.JOptionPane;
+import org.openstreetmap.josm.data.coor.LatLon;
 
 abstract public class Grabber implements Runnable {
 protected Bounds b;
@@ -22,7 +23,25 @@
 
 Grabber(Bounds b, Projection proj,
 double pixelPerDegree, GeorefImage image, MapView mv, WMSLayer layer) {
-this.b = b;
+
+
+if (b.min != null  b.max != null  WMSPlugin.doOverlap) {
+double latCent = (b.min.lat() + b.max.lat()) / 2;
+double lonCent = (b.min.lon() + b.max.lon()) / 2;
+
+double latSize =  b.max.lat() - b.min.lat();
+double lonSize =  b.max.lon() - b.min.lon();
+
+double latCoef = (100.0 + WMSPlugin.overlapLat) / 100.0 / 2.0;
+double lonCoef = (100.0 + WMSPlugin.overlapLon) / 100.0 / 2.0;
+
+this.b = new Bounds( new LatLon(latCent - latCoef * latSize,
+lonCent - lonCoef * lonSize),
+ new LatLon(latCent + latCoef * latSize,
+lonCent + lonCoef * lonSize));
+} else
+this.b = b;
+
 this.proj = proj;
 this.pixelPerDegree = pixelPerDegree;
 this.image = image;
Index: src/wmsplugin/WMSPlugin.java
===
--- src/wmsplugin/WMSPlugin.java	(revision 15029)
+++ src/wmsplugin/WMSPlugin.java	(working copy)
@@ -37,6 +37,10 @@
 static ArrayListWMSInfo wmsList = new ArrayListWMSInfo();
 static TreeMapString,String wmsListDefault = new TreeMapString,String();
 
+static boolean doOverlap = false;
+static int overlapLat = 4;
+static int overlapLon = 14;
+
 // remember state of menu item to restore on changed preferences
 static private boolean menuEnabled = false;
 
@@ -76,6 +80,22 @@
 MapString,String prefs = Main.pref.getAllPrefix(wmsplugin.url.);
 
 TreeSetString keys = new TreeSetString(prefs.keySet());
+   
+// Here we load the settings for overlap checkbox and spinboxes.
+
+try {
+doOverlap = Boolean.valueOf(prefs.get(wmsplugin.url.overlap));
+} catch (Exception e) {} // If sth fails, we drop to default settings.
+
+try {
+overlapLat = Integer.valueOf(prefs.get(wmsplugin.url.overlapLat));
+} catch (Exception e) {} // If sth fails, we drop to default settings.
+
+try {
+overlapLon = Integer.valueOf(prefs.get(wmsplugin.url.overlapLon));
+} catch (Exception e) {} // If sth fails, we drop to default settings.
+
+// And then the names+urls of WMS servers
 int prefid = 0;
 String name = null;
 String url = null;
Index: src/wmsplugin/GeorefImage.java
===
--- src/wmsplugin/GeorefImage.java	(revision 15029)
+++ src/wmsplugin/GeorefImage.java	

Re: [josm-dev] WMS plugin 'overlap' patch

2009-05-15 Thread Dirk Stöcker

On Fri, 15 May 2009, Radomír Černoch wrote:


in the Czech republic we have a problem with one WMS server, which
tends to wipe the WMS tiles near its borders.

To solve this I've created a patch allowing WMS plugin in JOSM to
download tiles, whose sides do not align perfectly. Instead WMS can
add a certain percentage of overlap between neighbouring tiles so that
the missing information in one tile is 'overlayed' by the neigbouring
one (if server supports alpha channel). This feature can be disabled
(which is the default) and the amount of overlap can be set in
preferences.

The patch is included in the attachment as I do not have access to the
SVN. Could I ask for including it into upstream?


Please make a Trac ticket in JOSM-Trac for it. I'm currently a bit behind 
in checking bugs and patches and stuff in mailinglists tends to beeing 
forgotten.



PS: Can I also ask for access to JOSM SVN? What should I do for it? I
have developed another plugin, which could be included.


This is not JOSM SVN, but normal OSM SVN. You need to ask Tim to get an 
account. Say that you want to handle JOSM plugins. Otherwise attaching 
patches to the JOSM Trac also works.


To get JOSM SVN access is a bit more complicated. You must make yourself 
so valuable, that I (or Frederik) decide to give you SVN access. :-)


Ciao
--
http://www.dstoecker.eu/ (PGP key available)___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev