Re: [GRASS-user] Re: v.breach addon
M S pisze: Kubuntu 9.10 64bit, using GRASS 6.5. I ran v.breach on a stream network extracted with r.stream.extract. It runs through the program, and puts up an error dialogue at v.parallel. Attached is the log file. It looks like a negative number is being passed to v.parallel. I saw in the documentation this might be a problematic module in the script for non-metric coordinate system. Below is the portion of the output box from the module. Note the error 11 lines down. --- Building topology for vector map breach_pt_segm_2l_r_tmp_5636_0_v_breach_sh... Registering primitives... 100 primitives registered 259 vertices registered Building areas... 0 areas built 0 isles built Attaching islands... Attaching centroids... Number of nodes: 100 ERROR: value -0.1 out of range for parameter distance Number of primitives: 100 Number of points: 0 Number of lines: 100 Number of boundaries: 0 Number of centroids: 0 Legal range: 0-1 Description: Creates parallel line to input vector lines. Keywords: vector, geometry --- Should I look in the script for this portion, and check if it is a metric issue? or does this seem like a different problem? A different one. v.parallel used to accept negative offsets to distinguish left/right. Now (after v.parallel2 replaced it) it doesn't accept negative offsets anymore. Maciek 2009/12/14 Maciej Sieczka msiec...@sieczka.org: Luís Ferreira pisze: I used v.breach with both SQLite and dbf, with ETRS89 metric negative coordinates, from 20m to 2m resolution and worked fine. System: Ubuntu 9.10 Karmic Koala, 64 bits, GRASS 6.4SVN. But I remember to see some warnings like yours. Are these warnings critical? Visually, the interpolated DEMs with v.surf.rst are coherent. I'll let you know next time I take a hobbyist look at v.breach. But I really don't know when it will be. Good luck. Maciek On Mon, 2009-12-14 at 04:57 +0100, Maciej Sieczka wrote: @Luís: I applied most your of patch. Thanks! Are you sure that's all errors? I keep on getting warnings like: WARNING: Unable to get point on line: cat = 1 offset = 2.831711 (line length = 0) at v.segment calls in the script, and a crash at v.in.ascii. -- Maciej Sieczka http://www.sieczka.org ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user -- Maciej Sieczka http://www.sieczka.org ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
Re: [GRASS-user] Re: v.breach addon
Yes, it worked with sqlite db connection. Luís On Mon, 2009-12-14 at 15:52 -0500, M S wrote: Luis: I noticed you mentioned you had run it with sqlite and dbf database drivers. Mine is mapped to sqlite currently. I see some v.to.db commands in the script, but do I understand correctly that you didnt have issues using sqlite as backend? ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
[GRASS-user] Re: v.breach addon
I used v.breach with both SQLite and dbf, with ETRS89 metric negative coordinates, from 20m to 2m resolution and worked fine. System: Ubuntu 9.10 Karmic Koala, 64 bits, GRASS 6.4SVN. But I remember to see some warnings like yours. Are these warnings critical? Visually, the interpolated DEMs with v.surf.rst are coherent. Luís On Mon, 2009-12-14 at 04:57 +0100, Maciej Sieczka wrote: @Luís: I applied most your of patch. Thanks! Are you sure that's all errors? I keep on getting warnings like: WARNING: Unable to get point on line: cat = 1 offset = 2.831711 (line length = 0) at v.segment calls in the script, and a crash at v.in.ascii. ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
[GRASS-user] Re: v.breach addon
Luís Ferreira pisze: I used v.breach with both SQLite and dbf, with ETRS89 metric negative coordinates, from 20m to 2m resolution and worked fine. System: Ubuntu 9.10 Karmic Koala, 64 bits, GRASS 6.4SVN. But I remember to see some warnings like yours. Are these warnings critical? Visually, the interpolated DEMs with v.surf.rst are coherent. I'll let you know next time I take a hobbyist look at v.breach. But I really don't know when it will be. Good luck. Maciek On Mon, 2009-12-14 at 04:57 +0100, Maciej Sieczka wrote: @Luís: I applied most your of patch. Thanks! Are you sure that's all errors? I keep on getting warnings like: WARNING: Unable to get point on line: cat = 1 offset = 2.831711 (line length = 0) at v.segment calls in the script, and a crash at v.in.ascii. -- Maciej Sieczka http://www.sieczka.org ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
Re: [GRASS-user] Re: v.breach addon
Kubuntu 9.10 64bit, using GRASS 6.5. I ran v.breach on a stream network extracted with r.stream.extract. It runs through the program, and puts up an error dialogue at v.parallel. Attached is the log file. It looks like a negative number is being passed to v.parallel. I saw in the documentation this might be a problematic module in the script for non-metric coordinate system. Below is the portion of the output box from the module. Note the error 11 lines down. --- Building topology for vector map breach_pt_segm_2l_r_tmp_5636_0_v_breach_sh... Registering primitives... 100 primitives registered 259 vertices registered Building areas... 0 areas built 0 isles built Attaching islands... Attaching centroids... Number of nodes: 100 ERROR: value -0.1 out of range for parameter distance Number of primitives: 100 Number of points: 0 Number of lines: 100 Number of boundaries: 0 Number of centroids: 0 Legal range: 0-1 Description: Creates parallel line to input vector lines. Keywords: vector, geometry --- Should I look in the script for this portion, and check if it is a metric issue? or does this seem like a different problem? Thanks, Mark 2009/12/14 Maciej Sieczka msiec...@sieczka.org: Luís Ferreira pisze: I used v.breach with both SQLite and dbf, with ETRS89 metric negative coordinates, from 20m to 2m resolution and worked fine. System: Ubuntu 9.10 Karmic Koala, 64 bits, GRASS 6.4SVN. But I remember to see some warnings like yours. Are these warnings critical? Visually, the interpolated DEMs with v.surf.rst are coherent. I'll let you know next time I take a hobbyist look at v.breach. But I really don't know when it will be. Good luck. Maciek On Mon, 2009-12-14 at 04:57 +0100, Maciej Sieczka wrote: @Luís: I applied most your of patch. Thanks! Are you sure that's all errors? I keep on getting warnings like: WARNING: Unable to get point on line: cat = 1 offset = 2.831711 (line length = 0) at v.segment calls in the script, and a crash at v.in.ascii. -- Maciej Sieczka http://www.sieczka.org ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user v.breach.log Description: Binary data ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
Re: [GRASS-user] Re: v.breach addon
Luis: I noticed you mentioned you had run it with sqlite and dbf database drivers. Mine is mapped to sqlite currently. I see some v.to.db commands in the script, but do I understand correctly that you didnt have issues using sqlite as backend? Thanks, Mark 2009/12/14 Luís Ferreira lferreira7...@gmail.com: I used v.breach with both SQLite and dbf, with ETRS89 metric negative coordinates, from 20m to 2m resolution and worked fine. System: Ubuntu 9.10 Karmic Koala, 64 bits, GRASS 6.4SVN. But I remember to see some warnings like yours. Are these warnings critical? Visually, the interpolated DEMs with v.surf.rst are coherent. Luís On Mon, 2009-12-14 at 04:57 +0100, Maciej Sieczka wrote: @Luís: I applied most your of patch. Thanks! Are you sure that's all errors? I keep on getting warnings like: WARNING: Unable to get point on line: cat = 1 offset = 2.831711 (line length = 0) at v.segment calls in the script, and a crash at v.in.ascii. ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
[GRASS-user] Re: v.breach addon
Luís Ferreira pisze: On Mon, 2009-12-07 at 08:11 +0100, Maciej Sieczka wrote: Luís Ferreira pisze: I had to change some strings to make v.breach work on GRASS 6.4svn revision 39873, Ubuntu/Debian 9.10. Can you send a diff (diff original_file your_file), as attachment? Here is it :) The problem with v.breach is down to v.parallel interface change in v.parallel2 - option `distance' does not accept negative offsets anymore, and introduces a new `side' option to distinguish left/right, instead: $ v.parallel in=lines out=lines_par dist=-1 ERROR: value -1 out of range for parameter distance Legal range: 0-1 @Devs: It would be cool if v.parallel could be changed to accept negative offsets, for compatibility within 6.x line. @Luís: I applied most your of patch. Thanks! Are you sure that's all errors? I keep on getting warnings like: WARNING: Unable to get point on line: cat = 1 offset = 2.831711 (line length = 0) at v.segment calls in the script, and a crash at v.in.ascii. Best, Maciek -- Maciej Sieczka http://www.sieczka.org ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
[GRASS-user] Re: v.breach addon
Aftermath... I had to change some strings to make v.breach work on GRASS 6.4svn revision 39873, Ubuntu/Debian 9.10. Luís Ferreira - line 310 v.parallel input=${OUTP}_grid_addcat_bnd_patch_break_ln_2l_tmp_${UNQ} output=${OUTP}_segm_2l_r_tmp_${UNQ} distance=0.1 --o /dev/null and line 312 v.parallel input=${OUTP}_grid_addcat_bnd_patch_break_ln_2l_tmp_${UNQ} output=${OUTP}_segm_2l_l_tmp_${UNQ} distance=-0.1 --o /dev/null change to v.parallel input=${OUTP}_grid_addcat_bnd_patch_break_ln_2l_tmp_${UNQ} output=${OUTP}_segm_2l_r_tmp_${UNQ} side=right distance=0.1 --o /dev/null v.parallel input=${OUTP}_grid_addcat_bnd_patch_break_ln_2l_tmp_${UNQ} output=${OUTP}_segm_2l_l_tmp_${UNQ} side=left distance=0.1 --o /dev/null -- line 388 v.db.addcol map=${OUTP} layer=2 'columns=x double, y double, z double, z_breach double' /dev/null change to v.db.addcol map=${OUTP} layer=2 'columns=x double precision, y double precision, z double precision, z_breach double precision' /dev/null --- line 474 db.execute input=$TMP.${PROG}.sql database=$dbs driver=$drv /dev/null change to db.execute input=$TMP.${PROG}.sql database=${database} driver= ${driver} /dev/null --- change $tbl to $table -- v.clean tool=prune defaulting to treshold=0.0 -- line 464 change to table=`v.db.connect -g ${OUTP} | awk '{print $2}'` database=`v.db.connect -g ${OUTP} | awk '{print $4}'` driver=`v.db.connect -g ${OUTP} | awk '{print $5}'` make explicit field separator for awk awk -F | On Wed, 2009-12-02 at 21:35 +0100, Maciej Sieczka wrote: Luís Ferreira pisze: I want to make a hydrological corrected DEM, using v.surf.rst for interpolation/aproximation and v.breach to introduce aditional elevation points at streams. My spatial reference system is EPSG:3763 ETRS89 / Portugal TM06, (metrical). As required in v.breach document: - each input vect line have a unique category in layer 1; - input lines not cross themselves; - input lines not constitute loops; - the input vector lines not stand out of the input raster DEM; - the input lines direction is from upstream to downstream. After creating a DEM using v.surf.rst I want to correct the false sinks in narrow valleys and sparse points areas. Using v.breach I have as result two output vectors for streams lines and points but without entities or attributes. I tried with both PostgreSQL and Sqlite database connections. For PostgreSQL connection: DBMI-Postgres driver error: Cannot select: SELECT cat,lcat,z,z_breach FROM topografia.vbreachpoints_2 ERROR: column z does not exist LINE 1: SELECT cat,lcat,z,z_breach FROM topografia.vbreachpoints_2 ^ Unable to open select cursor: 'SELECT cat,lcat,z,z_breach FROM topografia.vbreachpoints_2' Copy table failed Table vbreach_2 does not exist in database host=localhost,dbname=grass_pp_muda_pttm06 The table vbreach_2 is now part of vector map vbreach and may be deleted or overwritten by GRASS modules DBMI-Postgres driver error: Cannot create index: create unique index vbreach_2_cat on vbreach_2 ( cat ) ERROR: relation vbreach_2 does not exist Cannot create index DBMI-Postgres driver error: Cannot grant on table: grant select on vbreach_2 to public ERROR: relation vbreach_2 does not exist Cannot grant privileges on table vbreach_2 Select privileges were granted on the table Any ideas? If wanted I'll send the files needed for replication. Luís, Glad to see someone trying to use my stuff :). What GRASS version do you use? What OS? What was your command line, exactly? Maybe the culprit is the DB backend. I have never used the script with anything else than DBF. Try to switch to DBF in GRASS (db.connect). Or a GRASS version too new. I think I used it only against 6.3. There *might* have been changes to underlying GRASS modules. I've written the script, and used it the last time, about 2 years ago. It did the job fine then. I don't have a personal interest in developing it at the moment. I could provide you a commercial support if you like. That would include testing the script against your data, in a GRASS setup similar to yours, and fixing the tool if necessary. Please contact me in priv if you are interested. I don't charge much :). Or, if there's a GRASS hacker willing to do it for free - please go ahead, anybody. Best, Maciek v.breach_changed Description: application/shellscript
[GRASS-user] Re: v.breach addon
Luís Ferreira pisze: Aftermath... I had to change some strings to make v.breach work on GRASS 6.4svn revision 39873, Ubuntu/Debian 9.10. Can you send a diff (diff original_file your_file), as attachment? Maciek -- Maciej Sieczka http://www.sieczka.org ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
[GRASS-user] Re: v.breach addon
Luís Ferreira pisze: I want to make a hydrological corrected DEM, using v.surf.rst for interpolation/aproximation and v.breach to introduce aditional elevation points at streams. My spatial reference system is EPSG:3763 ETRS89 / Portugal TM06, (metrical). As required in v.breach document: - each input vect line have a unique category in layer 1; - input lines not cross themselves; - input lines not constitute loops; - the input vector lines not stand out of the input raster DEM; - the input lines direction is from upstream to downstream. After creating a DEM using v.surf.rst I want to correct the false sinks in narrow valleys and sparse points areas. Using v.breach I have as result two output vectors for streams lines and points but without entities or attributes. I tried with both PostgreSQL and Sqlite database connections. For PostgreSQL connection: DBMI-Postgres driver error: Cannot select: SELECT cat,lcat,z,z_breach FROM topografia.vbreachpoints_2 ERROR: column z does not exist LINE 1: SELECT cat,lcat,z,z_breach FROM topografia.vbreachpoints_2 ^ Unable to open select cursor: 'SELECT cat,lcat,z,z_breach FROM topografia.vbreachpoints_2' Copy table failed Table vbreach_2 does not exist in database host=localhost,dbname=grass_pp_muda_pttm06 The table vbreach_2 is now part of vector map vbreach and may be deleted or overwritten by GRASS modules DBMI-Postgres driver error: Cannot create index: create unique index vbreach_2_cat on vbreach_2 ( cat ) ERROR: relation vbreach_2 does not exist Cannot create index DBMI-Postgres driver error: Cannot grant on table: grant select on vbreach_2 to public ERROR: relation vbreach_2 does not exist Cannot grant privileges on table vbreach_2 Select privileges were granted on the table Any ideas? If wanted I'll send the files needed for replication. Luís, Glad to see someone trying to use my stuff :). What GRASS version do you use? What OS? What was your command line, exactly? Maybe the culprit is the DB backend. I have never used the script with anything else than DBF. Try to switch to DBF in GRASS (db.connect). Or a GRASS version too new. I think I used it only against 6.3. There *might* have been changes to underlying GRASS modules. I've written the script, and used it the last time, about 2 years ago. It did the job fine then. I don't have a personal interest in developing it at the moment. I could provide you a commercial support if you like. That would include testing the script against your data, in a GRASS setup similar to yours, and fixing the tool if necessary. Please contact me in priv if you are interested. I don't charge much :). Or, if there's a GRASS hacker willing to do it for free - please go ahead, anybody. Best, Maciek -- Maciej Sieczka http://www.sieczka.org ___ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user