Split polygons at dateline #580

sindizzy · 2022-01-05T21:51:35Z

Prerequisites

I have written a descriptive pull-request title
I have verified that there are no overlapping pull-requests open
I have verified that I am following the existing coding patterns and practice as demonstrated in the repository.

I have provided test coverage for my change (where applicable)

Description

Presents a technique for converting coordinates to a polygon. With the coordinates all in one hemisphere it is a
fairly simple process. However, when the geometry crosses the dateline it becomes a more challenging problem.
This routine attempts to do both. That is to say, it converts coordinates to a poly geometry regardless if the
geometry traverses the dateline or not. If the geometry does cross the dateline then it is split up into multiple
parts. This benefits the user in two ways: first, any GIS will be able to display the geometry in a proper way regardless of map projection and two, the GIS will be able to perform geoprocessing functions on the geometry. Note that feature classes do
not support features with multiple parts that share the same edge. Thus, it is left up to the user on how to
commit the geometries to a feature class. Suggestions are to save the multiple geometries as individual features
or to slightly alter the edges of each geometry so as to save all parts as a single multi-part feature. For the
latter option, the algorithm provides an input to trim edges of parts so that they can be used in a single feature.

The algorithm arose from attempting to convert worldwide airspaces (which can be found in any online aviation db) to polygons. Some of the coordinates result in polygons traversing the dateline and the result is ugly. That is because the GIS is not inherently aware that the polygons cross the dateline. This is what a standard conversion will look like:

The goal is to split those special polygons so that the airspaces are properly represented:

The PR includes a sample input with three processing options. If the item crosses the dateline then you will always get back multiple geometries no matter what options you use. The difference is what the dev intends to do with the geometries. The sample input will produce the results discussed below. For easy viewing the map projection has been set to Plate Carrée with the central meridian at 180 degrees (the dateline). The PR is also heavy on the debugs and I can remove those once the NTS devs can take a look at the initial proposal.

If the dev intends to add them as a single part per feature then this is what you will get with the standard options:

If the dev intends to add them as multiple parts per feature then its a bit more complicated. Through research and observation it was noted that existing datasets representing worldwide polygons already deal with the multi-part issue by trimming one of the polygons ever so slightly such that the two parts do not mathematically share an edge. From my research this trim gap appears to be connected to the feature class tolerance. The trim gap is a variable that the user must provide but should be made as small as possible. For demonstration purposes the trim gap has been made quite large and this is what it would look like with advanced options:

Abel

Added a routine to split coordinates into geometry that is split by the dateline. Added to NetTopologySuite.Samples.Technique.SplitPolygonAtDateline

Enhanced logic to determine geometry that crosses the dateline.

Added trim gap functionality such that if a polygon is split by the dateline then a part either on the western or eastern hemisphere can have a trim gap. This means it will not exactly touch the dateline.

Added more checks and debugs. Added comments and links for feature class tolerance.

added another test case

small tweaks

CLAassistant · 2022-01-05T21:51:40Z

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
2 out of 3 committers have signed the CLA.

✅ DGuidi
✅ sindizzy
❌ Perez

Perez seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

Tweak for zero trim gap

Small tweaks

Small tweaks Added checks at start of routine Modified block that closes a list of coords to create a polygon

DGuidi · 2022-01-11T07:56:24Z

src/NetTopologySuite/NetTopologySuite.csproj

@@ -31,7 +31,6 @@
  </ItemGroup>

  <ItemGroup Condition=" '$(EnableApiCompat)' == 'true' ">
-    <PackageReference Include="Microsoft.DotNet.ApiCompat" Version="6.0.0-beta.21159.11" PrivateAssets="All" />


I see a potential problem here, because EnableApiCompat == True but the reference to the package was removed.
Honestly I don't even know how this package is used, but I think it should be restored

I wrote an article on my issue here #577
I didn't know how to fix it. It was suggested I remove a package and recompile. That seemed to work.

My doubt is about pushing the package removal, I suggest to re-add the reference and then investigate the problem in #577

This has to go away. If it does not work in your environment, you have to fix your environment.
If you want to turn it off temporarily, change EnableApiCompat to false (see 1st PropertyGroup)

Yes but I don't know how to go about fixing my environment. I'm not even sure where to start. I did a google search and not much came up. I'm just unfamiliar with ApiCompat and how it works.

@sindizzy looks that - thanks to @airbreather - we have a fix for this!

we have a fix for this

Not sure about that one, since the NuGet.config file does actually exist in this project... the question is what's causing it to get ignored? (edit: assuming that is, in fact, what's causing the issues).

It's possible that we could try requiring the released .NET 6 SDK, since I think I remember reading something that the API compat feature was going to be more built-in in that version of the SDK.

DGuidi · 2022-01-11T08:00:51Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+    public class SplitPolygonAtDateline
+    {
+        [STAThread]
+        public static void main()


maybe this code can be moved to a unit test (along with other sample code, possibly?) I know that other "Technique" items uses this parrern (inherited from JTS code) but basically this means that this code isn't tested.

I haven't done any unit testing in any of my projects so am a complete rookie at that. As you mentioned I just followed the pattern of testing in other routines in the Technique area.

I will help you with that

@sindizzy added the unit test code, hope to not have done a mess.
This is just a sample, please feel free to add your own asserts and other checks.

DGuidi · 2022-01-11T08:03:06Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+            // ========================================================================================
+            // inputs
+            // ========================================================================================
+            string wktPrj = "GEOGCS['WGS 84',DATUM['WGS_1984',SPHEROID['WGS 84',6378137,298.257223563,AUTHORITY['EPSG','7030']],AUTHORITY['EPSG','6326']]," +


From what I understand, the projection string is used just ot check if the data is in a GEOGCS format. To me this check is useless, because we are not checking the geometry but a separate parameters. Maybe a check of the geometry values (inside the valid range of values) can be slower but useful?

I suppose I can do that but say if someone has coordinates in UTM that go from 50 to 170 meters. That will pass the geographic check as we are only checking values and not units. My thought was that I didn't want users to just hand over their coordinates hoping it would work with any projection. The easiest way, for me at least, was to force them to provide a projection string.

I see your point, but maybe it's something related to the documentation of the method/api. my2cents

Absolutely. I strived to add quite a bit of a description buttttttt there will be many that don't read it and just dive in. :)

DGuidi · 2022-01-11T08:04:39Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+        /// </remarks>
+        public static Geometry ToPolyExOp(List<Coordinate> inpCoords, string inpProj4Wkt, OgcGeometryType outType, double outTrimGap, double outDensifyResolution, bool isAttFixOutInvPolygons)
+        {
+            Debug.Print("$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$");


these "debug.print" should be removed from the main methods.

Yes, thank you. Per my main write-up "The PR is also heavy on the debugs and I can remove those once the NTS devs can take a look at the initial proposal."

DGuidi · 2022-01-11T08:05:12Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+            // ###################################################################################################################
+            if (inpCoords == null)
+            {
+                throw new Exception("Null input coordinates.");


please use "ArgomentNullException" instead of generic "Exception"

DGuidi · 2022-01-11T08:05:57Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+            {
+                case OgcGeometryType.LineString:
+                    if (inpCoords.Count < 2)
+                        throw new Exception("Input list has fewer than 2 coordinates. To create a polyline, the minimum is 2 coordinates.");


ArgumentOutOfRange (of maybe InvalidOperationException?) here

I think I agree with your assessment here and will change my code. However, on my work related code I like to tell the user exactly what is wrong. At times I've encountered the ArgumentOutOfRange exception and wondered "ok what is the range then?".

you can leave the message. I see your point and ArgumentException can be used too.
I must admit that, once you don't use the base Exception class, the proper exception to use can be arguable , so feel free to use the exception class you want. Often I use InvalidOperationException, but not sure I can be taken as an example :)

Sure thing. I am learning on the go.

I think I can use something like:

throw new ArgumentOutOfRangeException(nameof(count), count, "Must be non-negative.");

DGuidi · 2022-01-11T08:07:02Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+                        throw new Exception("Input list has fewer than 3 coordinates. To create a polygon, the minimum is 3 coordinates.");
+                    break;
+                default:
+                    throw new Exception("Unsupported output geometry type.");


NotSupportedException if the operation can not be implemented, or NotImplementedException if you plan to extend the code to support these operations in the future

got it. thanks.

DGuidi · 2022-01-11T08:09:23Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+        /// https://desktop.arcgis.com/en/arcmap/10.3/manage-data/editing-fundamentals/creating-and-editing-multipart-polygons.htm
+        /// https://desktop.arcgis.com/en/arcmap/10.3/guide-books/map-projections/what-happens-to-features-at-180-dateline-.htm
+        /// </remarks>
+        public static Geometry ToPolyExOp(List<Coordinate> inpCoords, string inpProj4Wkt, OgcGeometryType outType, double outTrimGap, double outDensifyResolution, bool isAttFixOutInvPolygons)


IList can be used, I think

How different is using ListOf vs IList? I'm curious as to the pros and cons.
Thanks.

Using IList, the caller can use as input any implementation if the IList interface, like arrays.

The only code that needs to be changed is the Exists checks at lines 197-198, then you can get rid of all the ToList and use directly the Coordinate array from the Geometry object itself.

I'd prefer inpCoordinates to be of type CoordinateSequence. Maybe this whole function should be encapsulated in a class implementing IEntireCoordinateSequenceFilter invoked by Geometry.Apply(filter)

Or have an overload of the function taking a Geometry as an argument. In that case IGeometryFilter would be an option.

Ok let me take a look at the NTS codebase and see how CoordinateSequence is used.

So how would the Geometry as an input work?

The assumption that any geometry can be broken down to a coordinate list and then the user selects what kind of output comes from those coordinates? -OR-

The assumption that if I provide a Polygon then I want output as polygon. If I provide a polyline then I want polyline for output? What if the user inputs other non-supported geometry types?

What makes the most sense?

DGuidi · 2022-01-11T08:11:40Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+                    Debug.Print("cutterPgon isValid = {0}", valOp2.IsValid);
+
+                    // perform an Erase to split the polygon with a polygon. this consists of an Intersection and then a Difference.
+                    //var intx = pPolygonShifted.Boundary.Intersection(cutterPgon);


Old commented code that can be removed?

yep. will remove after the review is done. I'm an over-debugger I have to confess. I wanted to double check every output so have tons of debugs. :)

DGuidi · 2022-01-11T08:12:22Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+                    Debug.Print("resGeoms were fixed");
+
+                    // method 2
+                    //var pm = new PrecisionModel(10000000000.0);


Old code that can be removed? or it can be handled using a specific methofd parameter?

well no. the routine to fix invalid geometries resulted from some research and eventually this article: #124
From that discussion it was noted that Buffer does not always work. So I added the second method in case I wanted to add as an option later. I could probably make the input an Enum rather than a Boolean. So I would have

public enum attFixOutInvPolygons: int { FixNone= 0, FixViaBuffer=1, FixViaPrecision=2, }

I like the enum idea

Im adding it as my next commit.

DGuidi · 2022-01-11T08:14:44Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+        /// https://desktop.arcgis.com/en/arcmap/10.3/manage-data/editing-fundamentals/creating-and-editing-multipart-polygons.htm
+        /// https://desktop.arcgis.com/en/arcmap/10.3/guide-books/map-projections/what-happens-to-features-at-180-dateline-.htm
+        /// </remarks>
+        public static Geometry ToPolyExOp(List<Coordinate> inpCoords, string inpProj4Wkt, OgcGeometryType outType, double outTrimGap, double outDensifyResolution, bool isAttFixOutInvPolygons)


"out" parameters (all the parameters following "inpProj4Wkt") can be moved to a single "parameter class", just for the sake of the readability

Question, is there a number of parameters where you move to a parameter class? Obviously if you have 30 params you don't want those all in the function call. If I have 3 or 4, I typically leave it in the function call as-is. Looking at the NTS core code there are various calls with 5 even 6 params right in the function call. So I'm wondering if this change is needed? I'm open to it.

Simple a suggestion, just because there are several "out" parameters, that maybe can be grouped for the sake of the readability.
Totally unrelated, see something like JsonConvert.Deserialize(jsonString, jsonDeserializationParamsClass).

Ok I will take a look. Thanks.

DGuidi · 2022-01-11T08:16:30Z

I just added some "formal" comments about the code, I don't know so much about the algorithm but I think you can add few unit test methods to show how to use the class to obtain different kind of results.
thanks for the addition anyway.

euangordon · 2022-01-13T15:35:57Z

@sindizzy This looks fantastic and fingers crossed it works well.

One question: We have many polygons crossing the 180 meridian. With this change will we be able to intersect with those polygons?

I will try and test this out myself, just wondering if you know already

sindizzy · 2022-01-14T01:29:09Z

I believe with the polygons split into two parts you can use them like any other geometry.

DGuidi · 2022-01-14T07:27:12Z

This looks fantastic and fingers crossed it works well

@euangordon can you suggest some more unit tests?

sindizzy · 2022-01-16T19:46:22Z

I just added some "formal" comments about the code, I don't know so much about the algorithm but I think you can add few unit test methods to show how to use the class to obtain different kind of results. thanks for the addition anyway.

Thanks I will spend some time this week to clean up the code and fix the commit authoring.

As for the tests, I added 3 tests on how to use the routine. I followed other routines in the same area ie test\NetTopologySuite.Samples.Console\Technique.

DGuidi · 2022-01-17T10:54:08Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+        /// A Geometry class that contains the result of converting the coordinates into a Polyline or Polygon. The result could be multiple
+        /// geometries.
+        /// </returns>
+        private static Geometry ToPolyExOp(List<Coordinate> coordsInp, string proj4WktInp)


should be made public, I suppose

with the minimum changes required to the original class, just to allow the unit test to be compiled

sindizzy · 2022-02-04T02:56:44Z

Please have a look at https://gist.github.com/FObermaier/ba25a5b6ccc98cb86efbbd485d17bc01

What am I looking for?

sindizzy · 2022-02-08T02:19:27Z

Please have a look at https://gist.github.com/FObermaier/ba25a5b6ccc98cb86efbbd485d17bc01

Is that the final routine?

FObermaier · 2022-02-08T08:49:45Z

@sindizzy, that is how I would do it, does it work for you?

sindizzy · 2022-02-08T15:59:04Z

@sindizzy, that is how I would do it, does it work for you?

Let me test it. Can I give you some feedback later in the week?

sindizzy · 2022-02-13T17:42:44Z

Ok I have taken a look and here are my comments:
The shifting of the geometry by 360 and splitting is easy. It is the detecting when it crosses the dateline that's the hard part. I work in the aviation industry and more often than not I encounter coordinates in the standard -180 to +180 even when the geometry crosses the dateline. So my proposal is something similar to what I have in my routine:
if the coordinates are <-180 or >180 then you know it crossed the dateline. we are done.
if not then we assume the coords are in the normal range -180 to +180. now test the sign of the coordinates. if they are all one sign then assume they do not cross the dateline. if we have a mix of positive and negative coords then potentially they have crossed the dateline. here is where it gets tricky and we have to make a simple guesstimate. that is to say if the coordinates are closer to the dateline then we assume they cross the dateline. if the coordinates are closer to the prime meridian then they do not cross the dateline. in this last part, the guesstimate could also be forced by the user to say that the coordinates DO cross the dateline. for example a linestring with one point at -10, 0 and another at +10, 0 but the user knows that this doesn't wrap towards the prime meridian. he knows it wraps towards the dateline. its an extreme case.

sindizzy · 2022-02-13T17:45:44Z

Also on line 87 I am not certain I understand the need to union the results. if we have a multi part polygon and we split it along the dateline then the output may be something we don't expect after the union. I would have to give this some more thought.

sindizzy · 2022-02-13T17:51:31Z

Another thing I would definitely like to see added is the trim gap. As I noted in my routine, after some research, it is apparent to me that one major GIS vendor takes land masses the cross the dateline (like Fiji) and creates a single feature with multiple parts.

However the rule for single feature with multi-part polygons is that they cannot share an edge. So by inspection these land masses have one part where a small gap is introduced so as to have it spatially not share an edge with the other part. What I have seen is one part has the edge at 180.000000 and the other has and edge at -179.99999999. Just a small enough gap so that the two parts can be a single feature.

As an option I have that in my routine and would use it quite often.

FObermaier · 2022-02-14T11:50:26Z

@sindizzy the example posted was not meant to replace your solution, it was meant to demonstrate how to work with CoordinateSequences and I[Entire]CoordinateSequenceFilters. I have next to no knowledge about dateline crossing caveats, everything I have to work with is planar or far enough from the dateline.

sindizzy · 2022-02-15T15:16:24Z

Ahhhh gotcha. Ok let me use that as a base for my logic.

Based on suggestion from @ FObermaier at https://gist.github.com/FObermaier/ba25a5b6ccc98cb86efbbd485d17bc01

sindizzy · 2022-02-21T03:14:10Z

@FObermaier Ok based on your suggestion I added a DatelineUtility class to my PR. I think I understand your preferred workflow now.

I did change the main filter to
private class IsCrossesDatelineFilter : IGeometryFilter

This is because if a user passes a multi-part polygon then each polygon part could exist wholly in one hemisphere. If we use the CoordinateSequence then that could give us a false positive. Before I continue with this code I wanted to see if I am on the right track here. The filter cycles through each part in the geometry and then through the Coordinate array. I don't use the CoordinateSequence so wondering if that will pose a problem?

Edit: I just thought about it and I think I will add another filter class for CoordinateSequence and that will be called from the main IsCrossesDatelineFilter. I think that's the best way forward.

Added CoordinateSequence filter to the DatelineUtility. It is called from the main Geometry filter in case the main geometry has multiple parts like a multi-part polygon.

sindizzy · 2022-02-21T04:29:34Z

Alright, I added a CoordinateSequence filter that is called from the Geometry filter. This should account for multi-part geometries as the input. If we just do a CoordinateSequence then the coordinate sweep could jump from one part to another and we could get a false positive.

FObermaier · 2022-01-31T08:39:15Z

test/NetTopologySuite.Samples.Console/Technique/InvalidGeomFixMethod.cs

+    /// <summary>
+    /// Method to use to fix invalid geometry.
+    /// </summary>
+    public enum InvalidGeomFixMethod : int


We don't need this. As NTS v2.4 we have NetTopologySuite.Geometries.Utilities.GeometryFixer to fix invalid geometries

FObermaier · 2022-01-31T10:14:59Z

test/NetTopologySuite.Samples.Console/Technique/SplitPolygonAtDateline.cs

+                var valOp = new IsValidOp(resGeoms);
+                if (!valOp.IsValid)
+                {
+                    if (outParams.InvGeomFixMethod == InvalidGeomFixMethod.FixViaBuffer)


This block should read:

resGeoms = NetTopologySuite.Geometries.Utilities.GeometryFixer.Fix(resGeoms);

test/NetTopologySuite.Samples.Console/Technique/DatelineUtility.cs

FObermaier · 2022-02-22T08:18:09Z

You don't need to split up into seperate filters. IEntireCoordinateSequenceFilter runs on all CoordinateSequences of a geometry's components, as long as IsDone is false.

Note: Only Point and LineString geometry classes have CoordinateSequence members. So if you have a Polygon the IEntireCoordinateSequenceFilter will be applied to the shell and holes (which are LinearRings (a special form of a LineString)) individually. Same for all the Multi... geometries and GeometryCollections.

Undo the two prior filters and setup only the IEntireCoordinateSequenceFilter filter. This filter will cycle through all parts in a geometry as long as Done=false. Our logic sets Done=true when any part crosses the dateline so this should work with any type of input geometry.

Updated comments.

sindizzy · 2022-02-23T02:03:30Z

@FObermaier Ahhhhh I think I got it now. Thanks for taking time to teach me how this works. I went back to only the IEntireCoordinateSequenceFilter and did a commit. I think this is what you were talking about so when you get a chance please take a look.

I see you updated a new routine so I took that and tweaked it just a little bit. Added my logic and the geographic CRS list. I will need to go through and study some parts more in detail but the baseline is there.

sindizzy · 2022-02-24T06:02:50Z

I noticed you updated your routine so I started over with DatelineUtility2. Added my logic and the geographic CRS list. I will need to study it further to understand certain parts.

airbreather · 2022-02-25T13:28:14Z

So, I haven't really been following this closely at all, but I do want to at least point out that the term "dateline" probably shouldn't show up in the code. I don't mind referring to it that way informally, but the term "antimeridian" is a more accurate term to use when referring to the thing that causes problems for NTS, since the International Date Line deviates from the true meridian in several places in order to avoid chopping countries in half.

sindizzy · 2022-02-26T17:42:31Z

Yes I agree on the definition of the international dateline. However, I would like to point out that most users, even major GIS companies, refer to the antemeridian as the "dateline". When I search for articles and info on this challenge it's always using the "dateline" moniker. I'm good either way though.

euangordon · 2022-07-28T08:44:44Z

Hi All, what is the status of this? Is there any chance it will get merged in?

sindizzy · 2022-07-28T13:51:10Z

I was working on it and I lost steam. I'll pick it up again soon.

euangordon · 2022-07-28T14:35:48Z

I was working on it and I lost steam. I'll pick it up again soon.

Thanks @sindizzy we did try and use your fork but didn't manage to get it working. Hopeing that once it is in the release build we will manage to get it working.

As a workaround, I am working on splitting shapes that cross the dateline, but it painful.

FObermaier · 2022-07-28T14:46:23Z

Hi All, what is the status of this? Is there any chance it will get merged in?

In the meantime you can try https://gist.github.com/FObermaier/ba25a5b6ccc98cb86efbbd485d17bc01

sindizzy · 2022-07-28T15:19:55Z

I was working on it and I lost steam. I'll pick it up again soon.

Thanks @sindizzy we did try and use your fork but didn't manage to get it working. Hopeing that once it is in the release build we will manage to get it working.

As a workaround, I am working on splitting shapes that cross the dateline, but it painful.

Yes I know how painful it is. You can try the algorithm here #496 That's where the discussion started.

sindizzy · 2024-04-28T16:36:24Z

I had this as a side project and am going to see if I can get back to refactoring it. I've spent the last couple of years transitioning from VB to C#. I've also upgraded my own algorithm slightly.

Perez added 6 commits December 20, 2021 11:41

SplitPolygonAtDateline initial

09b00e4

Added a routine to split coordinates into geometry that is split by the dateline. Added to NetTopologySuite.Samples.Technique.SplitPolygonAtDateline

Enhanced logic

38050fa

Enhanced logic to determine geometry that crosses the dateline.

Added trim gap functionality

87083ec

Added trim gap functionality such that if a polygon is split by the dateline then a part either on the western or eastern hemisphere can have a trim gap. This means it will not exactly touch the dateline.

Updates to SplitPolygonAtDateline

a498e70

Added more checks and debugs. Added comments and links for feature class tolerance.

updates for SplitPolygonAtDateline

4127578

added another test case

SplitPolygonAtDateline

a06df7d

small tweaks

This was referenced Jan 5, 2022

Polygon split at Dateline Code DotSpatial/DotSpatial#1407

Open

CLA issue #581

Open

Perez and others added 3 commits January 7, 2022 16:58

SplitPolygonAtDateline

e405257

Tweak for zero trim gap

SplitPolygonAtDateline tweaks

a820eba

Small tweaks

SplitPolygonAtDateline

2c7852a

Small tweaks Added checks at start of routine Modified block that closes a list of coords to create a polygon

DGuidi reviewed Jan 11, 2022

View reviewed changes

DGuidi reviewed Jan 17, 2022

View reviewed changes

added a sample unit test for the code

ed8b536

with the minimum changes required to the original class, just to allow the unit test to be compiled

DatelineUtility baseline

d8405fa

Based on suggestion from @ FObermaier at https://gist.github.com/FObermaier/ba25a5b6ccc98cb86efbbd485d17bc01

DatelineUtility: Added CoordinateSequence filter

7f13e18

Added CoordinateSequence filter to the DatelineUtility. It is called from the main Geometry filter in case the main geometry has multiple parts like a multi-part polygon.

FObermaier reviewed Feb 21, 2022

View reviewed changes

sindizzy added 2 commits February 22, 2022 19:56

DatelineUtility tweaks

eea0fcf

Updated comments.

DatelineUtility2 baseline

21e747e

I see you updated a new routine so I took that and tweaked it just a little bit. Added my logic and the geographic CRS list. I will need to go through and study some parts more in detail but the baseline is there.

This was referenced Mar 9, 2022

polygon.Contains doesn't work if your polygon crosses the international dateline. #532

Open

Add orthographic projection class NetTopologySuite/ProjNet4GeoAPI#101

Merged

Split polygons at dateline #580

Are you sure you want to change the base?

Split polygons at dateline #580

Conversation

sindizzy commented Jan 5, 2022 • edited

Prerequisites

Description

CLAassistant commented Jan 5, 2022 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

airbreather Mar 14, 2022 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

DGuidi Jan 17, 2022 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

sindizzy Jan 21, 2022 • edited

Choose a reason for hiding this comment

DGuidi commented Jan 11, 2022

euangordon commented Jan 13, 2022

sindizzy commented Jan 14, 2022 • edited

DGuidi commented Jan 14, 2022

sindizzy commented Jan 16, 2022

Choose a reason for hiding this comment

sindizzy commented Feb 4, 2022

sindizzy commented Feb 8, 2022

FObermaier commented Feb 8, 2022

sindizzy commented Feb 8, 2022

sindizzy commented Feb 13, 2022

sindizzy commented Feb 13, 2022

sindizzy commented Feb 13, 2022

FObermaier commented Feb 14, 2022

sindizzy commented Feb 15, 2022

sindizzy commented Feb 21, 2022 • edited

sindizzy commented Feb 21, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

FObermaier commented Feb 22, 2022

sindizzy commented Feb 23, 2022

sindizzy commented Feb 24, 2022

airbreather commented Feb 25, 2022

sindizzy commented Feb 26, 2022

euangordon commented Jul 28, 2022

sindizzy commented Jul 28, 2022

sindizzy commented Jan 5, 2022 •

edited

CLAassistant commented Jan 5, 2022 •

edited

airbreather Mar 14, 2022 •

edited

DGuidi Jan 17, 2022 •

edited

sindizzy Jan 21, 2022 •

edited

sindizzy commented Jan 14, 2022 •

edited

sindizzy commented Feb 21, 2022 •

edited