Need help producing scraps.kml without white blobs
Hello there, I "own" a fairly large dataset of a maze cave. Producing PDF output works well, but when producing kml using export map -projection plan -output Scraps.kml I get large white blobs where there should be transparency. See attachment. First assumption was that Therion needs some help by marking certain walls "outline in". But that didn't help. Then I tried redrawing the walls of the affected passages to get long lines of walls rather than concatenated short sections of walls, assuming that would help Therion to better understand. Didn't help. Anyone had this problem so far? If there is a way to post-process the kml file to remove the white areas, that would also help. Or any other hint to systematically track this down. Kind Regards, Michael
Hi Michael, it looks to me like it is some kind of inner/outer ring problem. Are you able to isolate and post some minimal sample, where this problem happens? It would help a lot. Thanks, S. On Mon, 19 Apr 2021 at 16:26, Michael Ross <knollesimage@t-online.de> wrote:
Hello there,
I "own" a fairly large dataset of a maze cave. Producing PDF output works well, but when producing kml using
export map -projection plan -output Scraps.kml
I get large white blobs where there should be transparency. See attachment.
First assumption was that Therion needs some help by marking certain walls "outline in". But that didn't help.
Then I tried redrawing the walls of the affected passages to get long lines of walls rather than concatenated short sections of walls, assuming that would help Therion to better understand. Didn't help.
Anyone had this problem so far?
If there is a way to post-process the kml file to remove the white areas, that would also help.
Or any other hint to systematically track this down.
Kind Regards, Michael
_______________________________________________ Therion mailing list Therion@speleo.sk https://mailman.speleo.sk/listinfo/therion
Hi Michael, I have found one issue with the current version of Google Earth - it was ignoring (not filling) polygons, where the last point was not equal to the first point. It is fixed in18eca78 <https://github.com/therion/therion/commit/18eca78edd9dab377a0aa585a61a273d81024afe>, but I am not sure, it if will fix your issues. S. On Mon, 19 Apr 2021 at 16:44, Stacho Mudrak <lists@group-s.sk> wrote:
Hi Michael,
it looks to me like it is some kind of inner/outer ring problem. Are you able to isolate and post some minimal sample, where this problem happens? It would help a lot.
Thanks, S.
On Mon, 19 Apr 2021 at 16:26, Michael Ross <knollesimage@t-online.de> wrote:
Hello there,
I "own" a fairly large dataset of a maze cave. Producing PDF output works well, but when producing kml using
export map -projection plan -output Scraps.kml
I get large white blobs where there should be transparency. See attachment.
First assumption was that Therion needs some help by marking certain walls "outline in". But that didn't help.
Then I tried redrawing the walls of the affected passages to get long lines of walls rather than concatenated short sections of walls, assuming that would help Therion to better understand. Didn't help.
Anyone had this problem so far?
If there is a way to post-process the kml file to remove the white areas, that would also help.
Or any other hint to systematically track this down.
Kind Regards, Michael
_______________________________________________ Therion mailing list Therion@speleo.sk https://mailman.speleo.sk/listinfo/therion
Dear Stacho, Bruce, Martin, Thanks for your replies. The low hanging fruits have been picked already long time ago: Yellow ticks checked, outline parms checked, passage fills are displayed correctly in PDF. Perhaps it displays differently in GoogleEarth compared to other GIS? (Have not yet used OpenGIS or ArcGIS). I think, Stacho is closest to a potential cause. It appears to me that the code path for kml generation is different from the codepath for PDF generation and follows different heuristics. For example, kml generation seems to be agnostic of "outline" parameters and direction of yellow sticks. I produced subsections by selecting small maps for the output, the behaviour is the same, i.e., no scalability / interdependency problem. I attach one of the scraps and the resulting map in kml. I cannot easily extract a data sample for replay by you, but I'm happy to provide the complete data to anyone who is willing to dive into this problem. As already said, a post-processing hint (replacing colours in the kml) would also help. Cheers, Michael. Von: Therion [mailto:therion-bounces@speleo.sk] Im Auftrag von Stacho Mudrak Gesendet: Montag, 19. April 2021 19:25 Cc: List for Therion users <therion@speleo.sk> Betreff: Re: [Therion] Need help producing scraps.kml without white blobs Hi Michael, I have found one issue with the current version of Google Earth - it was ignoring (not filling) polygons, where the last point was not equal to the first point. It is fixed in <https://github.com/therion/therion/commit/18eca78edd9dab377a0aa585a61a273d81024afe> 18eca78, but I am not sure, it if will fix your issues. S. On Mon, 19 Apr 2021 at 16:44, Stacho Mudrak <lists@group-s.sk <mailto:lists@group-s.sk> > wrote: Hi Michael, it looks to me like it is some kind of inner/outer ring problem. Are you able to isolate and post some minimal sample, where this problem happens? It would help a lot. Thanks, S. On Mon, 19 Apr 2021 at 16:26, Michael Ross <knollesimage@t-online.de <mailto:knollesimage@t-online.de> > wrote: Hello there, I "own" a fairly large dataset of a maze cave. Producing PDF output works well, but when producing kml using export map -projection plan -output Scraps.kml I get large white blobs where there should be transparency. See attachment. First assumption was that Therion needs some help by marking certain walls "outline in". But that didn't help. Then I tried redrawing the walls of the affected passages to get long lines of walls rather than concatenated short sections of walls, assuming that would help Therion to better understand. Didn't help. Anyone had this problem so far? If there is a way to post-process the kml file to remove the white areas, that would also help. Or any other hint to systematically track this down. Kind Regards, Michael _______________________________________________ Therion mailing list Therion@speleo.sk <mailto:Therion@speleo.sk> https://mailman.speleo.sk/listinfo/therion
Hi Michael,
but I'm happy to provide the complete data to anyone who is willing to dive into this problem.
Certainly I would be happy to help dissecting a survey to work out what's going wrong. So do share if you are comfortable doing so. One thing I immediately notice with your survey is that you seem to expect a scrap to have two outlines, which is not permitted. Near the southern end - where I see a white overlap in Google Earth due to the underlying passage - you end one passage, then start a new one on the far side of the pitch, without starting a new scrap. This happens at the junction in the north as well. If you were hoping to have two separate outlines there, then you need separate scraps. If not, then you are going to get a weird overlap anyway with the passage above and below both showing at the same time. In the middle of the scrap, your "-outline in" oxbow also is incomplete (the part that appears as a huge chamber in KML). Close off the open passage end with a wall, set to "-outline in -visibility off". I don't know if that will fix it, but it might. Further north, you start a passage inside the other passage, which causes another overlap. That may be correct if the passage does indeed pop out part way inside the outer wall of the other passage. But if you want to avoid that overlap, join the scrap walls to each other so that the passages do not overlap. The northern-most junction is a mess of overlaps, because you have ended the passage with it open at each side passage, and none of them extend through the junction in between. This is not unique to KML. If you render your cave with a background colour in PDF, you will see the same problem. One passage needs to "own" the junction. Meaning it must have an outline that extends fully around the junction (or you can use horrendous joins and have passages each owning parts of the junction). Use invisible borders with "-outline out" if you need to, in order to extend one of the passages through the junction so that it owns the passage. This particular situation is covered very nicely in Footleg's tutorial: http://wscc.darkgem.com/footleg/therion/Therion%20Tutorial%2015thMar2016.pdf Lesson 9 Does this help fix the problems you are seeing? Cheers, Tarquin
Hi Michael As Tarquin mentioned, I have found that stopping the walls on both sides of the passage in a single scrap will upset kml generation. Maybe pictures are clearer. In the scenarios pictured below you need to create separate scraps each side of the gap. So in your example there should be at least three plan scraps. Personally I prefer much shorter scraps, I find 250 m to be very long, but other than making debugging a little harder, excessive scrap length/size should not be a cause of your problems. Bruce From: Therion <therion-bounces@speleo.sk> On Behalf Of Michael Ross Sent: Tuesday, 20 April 2021 22:58 To: 'List for Therion users' <therion@speleo.sk> Subject: Re: [Therion] Need help producing scraps.kml without white blobs Dear Stacho, Bruce, Martin, Thanks for your replies. The low hanging fruits have been picked already long time ago: Yellow ticks checked, outline parms checked, passage fills are displayed correctly in PDF. Perhaps it displays differently in GoogleEarth compared to other GIS? (Have not yet used OpenGIS or ArcGIS). I think, Stacho is closest to a potential cause. It appears to me that the code path for kml generation is different from the codepath for PDF generation and follows different heuristics. For example, kml generation seems to be agnostic of "outline" parameters and direction of yellow sticks. I produced subsections by selecting small maps for the output, the behaviour is the same, i.e., no scalability / interdependency problem. I attach one of the scraps and the resulting map in kml. I cannot easily extract a data sample for replay by you, but I'm happy to provide the complete data to anyone who is willing to dive into this problem. As already said, a post-processing hint (replacing colours in the kml) would also help. Cheers, Michael. Von: Therion [mailto:therion-bounces@speleo.sk] Im Auftrag von Stacho Mudrak Gesendet: Montag, 19. April 2021 19:25 Cc: List for Therion users <therion@speleo.sk <mailto:therion@speleo.sk> > Betreff: Re: [Therion] Need help producing scraps.kml without white blobs Hi Michael, I have found one issue with the current version of Google Earth - it was ignoring (not filling) polygons, where the last point was not equal to the first point. It is fixed in <https://github.com/therion/therion/commit/18eca78edd9dab377a0aa585a61a273d81024afe> 18eca78, but I am not sure, it if will fix your issues. S. On Mon, 19 Apr 2021 at 16:44, Stacho Mudrak <lists@group-s.sk <mailto:lists@group-s.sk> > wrote: Hi Michael, it looks to me like it is some kind of inner/outer ring problem. Are you able to isolate and post some minimal sample, where this problem happens? It would help a lot. Thanks, S. On Mon, 19 Apr 2021 at 16:26, Michael Ross <knollesimage@t-online.de <mailto:knollesimage@t-online.de> > wrote: Hello there, I "own" a fairly large dataset of a maze cave. Producing PDF output works well, but when producing kml using export map -projection plan -output Scraps.kml I get large white blobs where there should be transparency. See attachment. First assumption was that Therion needs some help by marking certain walls "outline in". But that didn't help. Then I tried redrawing the walls of the affected passages to get long lines of walls rather than concatenated short sections of walls, assuming that would help Therion to better understand. Didn't help. Anyone had this problem so far? If there is a way to post-process the kml file to remove the white areas, that would also help. Or any other hint to systematically track this down. Kind Regards, Michael _______________________________________________ Therion mailing list Therion@speleo.sk <mailto:Therion@speleo.sk> https://mailman.speleo.sk/listinfo/therion
I have found that stopping the walls on both sides of the passage in a single scrap will upset kml generation.
Absolutely. I have reproduced the problem on a small scale like this. When you have a scrap containing two disconnected sections of "-outline out", then it gets confused when you use "-outline in" and treats it as a new "outline out" layered on top, instead of punching out the hole. [attached outline.png] [attached whiteblob.jpg] If you put the disconnected segment of outline into its own scrap, then the white blob disappears, and you get the correct rendering. [attached twoscraps.png] [attached correctrendering.jpg] This mistake is made a lot of times in the codebase, and seems to be the main cause of the blobs. [attached exampleproblem.png] When rendering the PDFs you also get a tonne of warnings, and these are trying to help you avoid the problem. The warnings are in the compiler output, and look like this: [Warning: scrap outline intersects itself in scrap WGE17-20@E5-E23Survey.Wassergang.fulab] Most of the time, these are an attempt to turn a corner too hard (where two lines join, try turning the last drag handle towards the next line's direction), or a Bézier curve point with the drag handles in the wrong direction, or two walls that overlap very slightly instead of being joined to the same point. There are quite a few hints randomly scattered on the Therion wiki about this. It's usually a less important warning though. You also get a *really* useful debugging output when generating a .lox file, where it tells you this exact problem: C:\Program Files (x86)\Therion\therion.exe: warning -- 24B1Scrap@21NordSurvey.Alter_Teil.fulab -- multiple scrap outer outlines not supported yet *This is the cause of your problems.* Fix that, and you fix the white blobs. I am also seeing "invalid scrap outline" warnings. This usually means that you have a station without a -name attribute.
21. 4. 2021 v 11:41, Tarquin Wilton-Jones via Therion <therion@speleo.sk>:
I have found that stopping the walls on both sides of the passage in a single scrap will upset kml generation.
Absolutely. I have reproduced the problem on a small scale like this.
This should solve your problem. Anyway, if you will try to imagine such situation as missing wall in passage is it is not real. There should be some kind off wall to circumscribe passage (emty space) from bedrock. M.
Me puedes reenviar el ultimo mensaje, se ha cruzado con otro el email. El 21/04/2021 a las 12:04, Eulogio Pardo Iguzquiza escribió:
21. 4. 2021 v 11:41, Tarquin Wilton-Jones via Therion <therion@speleo.sk <mailto:therion@speleo.sk>>:
I have found that stopping the walls on both sides of the passage in a single scrap will upset kml generation.
Absolutely. I have reproduced the problem on a small scale like this.
This should solve your problem. Anyway, if you will try to imagine such situation as missing wall in passage is it is not real. There should be some kind off wall to circumscribe passage (emty space) from bedrock.
M.
Thanks, Tarquin and Martin, for the many hints. I need to digest those first and then apply them, this will take some time. Cheers, Michael.
Hi Michael Unless Stacho's magic fixes your problem, I would expect that it is most likely that the yellow tick marks do not point inside of the passages in all cases, or -outline in/out is not set correctly. Your example image is exactly the sort of thing that first time kml users produce. It is very common. There should not be any problem with many short wall lines, except that with many, there is more opportunity to get a yellow tick or -outline wrong. A couple of months ago I encountered a case where ends of two adjacent scraps were one above the other, and one of them was well rendered, and the other as in your example. Turned out that both scraps were drawn correctly and simply moving one of the end lines a bit fixed the problem. Have never encountered that before as far as I know, and I would expect that this is unlikely to be your problem. Have a look at https://therion.speleo.sk/wiki/tips#avoiding_invisible_or_misshapen_scraps_i n_loch_and_google_earth It is a bit outdated now, as I think the behaviours are different now, but it gives you an idea of what to look out for. When troubleshooting this sort of thing, also produce pdf outputs with a passage foreground colour. This can point to problems with gaps between scraps and incorrect yellow ticks (interior of passage), but will not identify -outline errors (because the pdf generator seems to reliably know better than the user!) Bruce -----Original Message----- From: Therion <therion-bounces@speleo.sk> On Behalf Of Michael Ross Sent: Tuesday, 20 April 2021 02:20 To: therion@speleo.sk Subject: [Therion] Need help producing scraps.kml without white blobs Hello there, I "own" a fairly large dataset of a maze cave. Producing PDF output works well, but when producing kml using export map -projection plan -output Scraps.kml I get large white blobs where there should be transparency. See attachment. First assumption was that Therion needs some help by marking certain walls "outline in". But that didn't help. Then I tried redrawing the walls of the affected passages to get long lines of walls rather than concatenated short sections of walls, assuming that would help Therion to better understand. Didn't help. Anyone had this problem so far? If there is a way to post-process the kml file to remove the white areas, that would also help. Or any other hint to systematically track this down. Kind Regards, Michael
Yellow tick indicate “the air”. M. Odesláno z iPhonu 19. 4. 2021 v 21:52, Bruce Mutton <bruce@tomo.co.nz>:
Hi Michael Unless Stacho's magic fixes your problem, I would expect that it is most likely that the yellow tick marks do not point inside of the passages in all cases, or -outline in/out is not set correctly. Your example image is exactly the sort of thing that first time kml users produce. It is very common. There should not be any problem with many short wall lines, except that with many, there is more opportunity to get a yellow tick or -outline wrong.
A couple of months ago I encountered a case where ends of two adjacent scraps were one above the other, and one of them was well rendered, and the other as in your example. Turned out that both scraps were drawn correctly and simply moving one of the end lines a bit fixed the problem. Have never encountered that before as far as I know, and I would expect that this is unlikely to be your problem.
Have a look at https://therion.speleo.sk/wiki/tips#avoiding_invisible_or_misshapen_scraps_i n_loch_and_google_earth It is a bit outdated now, as I think the behaviours are different now, but it gives you an idea of what to look out for.
When troubleshooting this sort of thing, also produce pdf outputs with a passage foreground colour. This can point to problems with gaps between scraps and incorrect yellow ticks (interior of passage), but will not identify -outline errors (because the pdf generator seems to reliably know better than the user!)
Bruce
-----Original Message----- From: Therion <therion-bounces@speleo.sk> On Behalf Of Michael Ross Sent: Tuesday, 20 April 2021 02:20 To: therion@speleo.sk Subject: [Therion] Need help producing scraps.kml without white blobs
Hello there,
I "own" a fairly large dataset of a maze cave. Producing PDF output works well, but when producing kml using
export map -projection plan -output Scraps.kml
I get large white blobs where there should be transparency. See attachment.
First assumption was that Therion needs some help by marking certain walls "outline in". But that didn't help.
Then I tried redrawing the walls of the affected passages to get long lines of walls rather than concatenated short sections of walls, assuming that would help Therion to better understand. Didn't help.
Anyone had this problem so far?
If there is a way to post-process the kml file to remove the white areas, that would also help.
Or any other hint to systematically track this down.
Kind Regards, Michael
_______________________________________________ Therion mailing list Therion@speleo.sk https://mailman.speleo.sk/listinfo/therion
participants (7)
-
Bruce Mutton -
Evaristo Quiroga Raimundez -
Martin Sluka -
Martin Sluka -
Michael Ross -
Stacho Mudrak -
Tarquin Wilton-Jones