include index of MultiLineString geometries in nearest-point-on-line

[andrewharvey]

Please fill in this template.

• Use a meaningful title for the pull request. Include the name of the package modified.
• Have read How To Contribute.
• Run npm test at the sub modules where changes have occurred.
• Run npm run lint to ensure code style at the turf module level.
  this gave a lot of unreleated errors

[mfedderly]

It looks like a lot of the test output files got modified in whitespace-only ways. Can you run prettier against them to get the formatting sorted out and lower the line count of changes overall? The pre-commit hooks should have done it for you but I guess there's some issue there that I can look into.

[twelch]

Yes I would guess there is an issue, the last PR I merged had a lot of whitespace-only changes in test output.

[mfedderly]

Aha. We disabled prettier on the test files intentionally.

Disabled here: https://github.com/Turfjs/turf/pull/2519/files#diff-b640b344ee7f3f03d2a443795a5d0708ef50e2e6e34214109ab2aad13ad6ba98R13-R19

Suggested change here: #2670

@andrewharvey running prettier won't actually change the test files until we do something with 2670.

[andrewharvey]

Would you like me to try and manually fix this to keep the existing style but still adding the new property? Or would you be okay to accept the style changes and then leave the prettier fix to later?

[mfedderly]

Yeah I'm inclined to just merge this without worrying about whitespace since that change may take a little while to come to a consensus on. If I thought it would be quicker it might have been worth waiting so that we can more easily see what the test fixture changes were from this PR.

[mfedderly]

Just for posterity, the actual changes this introduces in the output that aren't whitespace are just adding multiFeatureIndex entries into the properties object as expected.

diff --git a/packages/turf-nearest-point-on-line/test/out/line-northern-latitude-#344.geojson b/packages/turf-nearest-point-on-line/test/out/line-northern-latitude-#344.geojson
index 66e3da1cc..64a5af130 100644
--- a/packages/turf-nearest-point-on-line/test/out/line-northern-latitude-#344.geojson
+++ b/packages/turf-nearest-point-on-line/test/out/line-northern-latitude-#344.geojson
@@ -40,6 +40,7 @@
       "type": "Feature",
       "properties": {
         "dist": 5.964157,
+        "multiFeatureIndex": 0,
         "location": 19.828788,
         "index": 1,
         "marker-color": "#F0F"
diff --git a/packages/turf-nearest-point-on-line/test/out/line1.geojson b/packages/turf-nearest-point-on-line/test/out/line1.geojson
index f327e2e25..634477fa6 100644
--- a/packages/turf-nearest-point-on-line/test/out/line1.geojson
+++ b/packages/turf-nearest-point-on-line/test/out/line1.geojson
@@ -40,6 +40,7 @@
       "type": "Feature",
       "properties": {
         "dist": 2.555507,
+        "multiFeatureIndex": 0,
         "location": 22.141834,
         "index": 1,
         "marker-color": "#F0F"
diff --git a/packages/turf-nearest-point-on-line/test/out/multiLine1.geojson b/packages/turf-nearest-point-on-line/test/out/multiLine1.geojson
index 1574904d8..b6765f1fe 100644
--- a/packages/turf-nearest-point-on-line/test/out/multiLine1.geojson
+++ b/packages/turf-nearest-point-on-line/test/out/multiLine1.geojson
@@ -69,6 +69,7 @@
       "properties": {
         "dist": 114.725451,
         "index": 21,
+        "multiFeatureIndex": 1,
         "location": 9479.011715,
         "marker-color": "#F0F"
       },
diff --git a/packages/turf-nearest-point-on-line/test/out/multiLine2.geojson b/packages/turf-nearest-point-on-line/test/out/multiLine2.geojson
index b20e34e7f..6c55987cc 100644
--- a/packages/turf-nearest-point-on-line/test/out/multiLine2.geojson
+++ b/packages/turf-nearest-point-on-line/test/out/multiLine2.geojson
@@ -50,6 +50,7 @@
       "properties": {
         "dist": 390.942725,
         "index": 0,
+        "multiFeatureIndex": 1,
         "location": 1656.139708,
         "marker-color": "#F0F"
       },
diff --git a/packages/turf-nearest-point-on-line/test/out/multiLine3.geojson b/packages/turf-nearest-point-on-line/test/out/multiLine3.geojson
index 3b7ea40b1..67be313c2 100644
--- a/packages/turf-nearest-point-on-line/test/out/multiLine3.geojson
+++ b/packages/turf-nearest-point-on-line/test/out/multiLine3.geojson
@@ -58,6 +58,7 @@
       "type": "Feature",
       "properties": {
         "dist": 120.886021,
+        "multiFeatureIndex": 0,
         "location": 214.548785,
         "index": 0,
         "marker-color": "#F0F"
diff --git a/packages/turf-nearest-point-on-line/test/out/route1.geojson b/packages/turf-nearest-point-on-line/test/out/route1.geojson
index 737c9a49a..9ca24d947 100644
--- a/packages/turf-nearest-point-on-line/test/out/route1.geojson
+++ b/packages/turf-nearest-point-on-line/test/out/route1.geojson
@@ -4793,6 +4793,7 @@
       "type": "Feature",
       "properties": {
         "dist": 7.876559,
+        "multiFeatureIndex": 0,
         "location": 183.461611,
         "index": 3104,
         "marker-color": "#F0F"
diff --git a/packages/turf-nearest-point-on-line/test/out/route2.geojson b/packages/turf-nearest-point-on-line/test/out/route2.geojson
index 3103b9857..d17968e69 100644
--- a/packages/turf-nearest-point-on-line/test/out/route2.geojson
+++ b/packages/turf-nearest-point-on-line/test/out/route2.geojson
@@ -3797,6 +3797,7 @@
       "type": "Feature",
       "properties": {
         "dist": 19.22748,
+        "multiFeatureIndex": 0,
         "location": 303.703041,
         "index": 1185,
         "marker-color": "#F0F"

[mfedderly]

should we add the featureIndex too since we have it?

[andrewharvey]

This method is having an identity crisis, it's called nearestPointOn**Line** (singular) implying it only accepts one line, however the first argument is named lines (plural) implying it accepts multiple lines, the lines argument type is defined to accept Geometry|Feature<LineString|MultiLineString> which implies only a single line Feature is accepted, albeit you can have a MultiLineString geometry within that single Feature, however it may also accept multiple lines through a GeometryCollection type of Geometry.

Checking the implementation it just passes lines through to flattenEach which accepts FeatureCollection | Feature | Geometry thereby giving rise to the featureIndex.

We should decide that either:

1. the function is to only accept a single line feature and therefore doesn't need to deal with featureIndex or,
2. the function can accept multiple features and therefore also include the featureIndex in the result.

While officially supporting FeatureCollection sounds like a good idea, it gets complicated when that FeatureCollection contains multiple geometry types, and defining how the method behaves in those situations is messy. On the other hand we already support GeometryCollection which has the same issue. 🤷

I think for now we go ahead and just include the multiFeatureIndex and leave the decision on featureIndex to another discussion/issue/PR.

[mfedderly]

Fair points, thanks for the detailed writeup!

[andrewharvey]

just rebased to resolve the merge conflict