Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Refactored location sharing timeline items to use the mapserver configured by the wellknown file #7852

Open
wants to merge 6 commits into
base: develop
Choose a base branch
from

Conversation

artkoenig
Copy link
Contributor

@artkoenig artkoenig commented Dec 27, 2022

Type of change

  • Feature
  • Bugfix
  • Technical
  • Other :

Content

The map server configuration (style.json), provided by the wellknown file, should be used for all use-cases, showing locations on a map.

Motivation and context

The map configuration (style.json), provided by the wellknown is not used for displaying the location in the timeline. Currently the Maptiler static map API is used for this, which is loading map data from the other source. This is problematic, due to copyright and/or network limitations, which may prohibit the use of map sources other than specified in the wellknown file.

This PR also removes the static attribution string, hard-coded in the app. The attribution should be defined in the style.json file to match the used data source.

Tested devices

  • Physical
  • Emulator
  • OS version(s):
    API 21, API 33

Signed-off-by: Artjom König artjom.koenig.ext@bwi.de

@artkoenig artkoenig changed the title Refactored location sharing to use the mapserver provided by the wellknown file Refactored location sharing timeline items to use the mapserver configured by the wellknown file Dec 27, 2022
Copy link
Contributor

@mnaturel mnaturel left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for raising this PR! I have added some comments I think we should address.
Also, when testing the build in "Matrix.org", I could not test the changes since the location icon was no more visible in the composer. It seems to confirm one of my comments about hiding the feature based on the "Well Known" server config.

In fact, Matrix.org does not have a map tiler config so we should have a fallback in a certain way.
Could you please test again using Matrix.org?

Also, I don't know if you have done it but could you perform a Leak Canary analysis when scrolling the timeline to see if there is any leak? It can be done by enabling the developer mode in Advanced settings. Then, on the home screen there will be a small gear icon on top right corner. Pressing it will open the debug screen with a "Memory leaks" entry.

holder.staticMapView.getMapAsync { mapbox ->
mapbox.setStyle(safeLocationUiData.locationUrl)
safeLocationUiData.locationData?.let {
mapbox.zoomToLocation(it, animate = false)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here I think the zoom level may not be correct. What we want in the timeline is INITIAL_MAP_ZOOM_IN_TIMELINE. And zoomToLocation is not using this value but another one. We may have to change the extension method to pass the required zoom level.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

})
.transform(imageCornerTransformation)
.into(holder.staticMapImageView)
holder.staticMapView.addOnDidFailLoadingMapListener {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Some clean up about the map should be added in the unbind(holder: H) method. All listeners that can be removed should be removed.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

private val rawService: RawService,
locationSharingConfig: LocationSharingConfig,
) {
private val keyParam = "?key=${locationSharingConfig.mapTilerKey}"
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if we should also remove the LocationSharingConfig class and the MapTiler API key.
@onurays What do you think? I am asking you since it appears you have added this key in the project and you may have more context about it.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

After some tests, it seems a fallback is still necessary on Matrix.org for example.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I added the fallback implementation to the maptiler API, it is configurable via LocationSharingConfig

@Johennes
Copy link
Contributor

Johennes commented Jan 2, 2023

My understanding was that we previously decided not to use the dynamic maps because of poor performance in the timeline. Have we falsified this concern?

@artkoenig
Copy link
Contributor Author

@Johennes to favor performance over functional correctness is IMO problematic. The new solution is slower and more memory consuming (here some tests results for scrolling a timeline full with location events):

Old:
old_performance

New:
new_performance

And also network performance (for one location event in the timeline):
Old:
old_network

New:
new_network

But

  • The Web-Client and the iOS-Client already use the dynamic maps in the timeline - for the sake of consistency, the Android-App should have the same behavior
  • The solution has to respect the customer's mapstyle settings. As I stated above - in the enterprise setup, there are copyright- and network access limitations (due to security considerations) which prevent using other than specified map servers
  • Even with the worst case scenario (timeline is full with location events), the performance impact is not so bad that you could not use the solution

Copy link
Contributor

@mnaturel mnaturel left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the updates. I have added some new remarks mostly about code styling/naming I think we should handle for clarity.

@@ -1,5 +1,6 @@
/*
* Copyright (c) 2021 New Vector Ltd
* Copyright (c) 2022 BWI GmbH
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why are there these changes of copyrights? I am not entirely sure is is allowed to be changed to something else than New Vector Ltd.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It looks like @artkoenig has added here the copyright notice to show that copyright goes to BWI GmbH.
It is fine to have the joint copyright here.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not an expert in this area but I know there has been irritation about this topic recently. So just to double check that we're doing the correct thing here: Has this particular question been confirmed with legal?

width = mapWidth
height = mapHeight

holder.staticMapView.apply {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As a general practice in the project, we try to avoid using apply as much as possible as it leads sometimes to strange bugs since the scope of this is changed inside.

@@ -18,4 +18,5 @@ package im.vector.app.features.location

data class LocationSharingConfig(
val mapTilerKey: String,
val isMapTilerFallbackEnabled: Boolean
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could you add a trailing comma here?

fun MapboxMap?.zoomToLocation(
locationData: LocationData,
animate: Boolean = true,
zoomLevel: Double = INITIAL_MAP_ZOOM_IN_PREVIEW) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think there is a missing breaking line after the last parameter and also a missing trailing comma.

@@ -332,7 +332,7 @@ class LiveLocationMapViewFragment :
.find { it.matrixItem.id == userId }
?.locationData
?.let { locationData ->
mapboxMap?.get()?.zoomToLocation(locationData, preserveCurrentZoomLevel = false)
mapboxMap?.get()?.zoomToLocation(locationData, true)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It think we should name the argument when assigning the value to make it clear about the purpose of the boolean..

Suggested change
mapboxMap?.get()?.zoomToLocation(locationData, true)
mapboxMap?.get()?.zoomToLocation(locationData, animate = true)

@@ -136,7 +135,7 @@ internal class AttachmentTypeSelectorViewModelTest {
return AttachmentTypeSelectorViewModel(
initialState,
vectorFeatures = fakeVectorFeatures,
vectorPreferences = fakeVectorPreferences.instance,
vectorPreferences = fakeVectorPreferences.instance
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Trailing comma should be kept accross the project.

@@ -0,0 +1,56 @@
/*
* Copyright (c) 2022 BWI GmbH
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Same comment as before about copyright of new files. I guess it should be New Vector Ltd.


@Test
fun `given enabled fallback to maptiler API, when map configuration is not set, then the fallback url should be returned`() {
val wellknownService = FakeWellknownService()
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it possible to separate by a new line the different Given, When, Then blocks of code? This is a pattern we try to follow to ease readability of tests and also separate well what is part of each block.

fun `given enabled fallback to maptiler API, when map configuration is not set, then the fallback url should be returned`() {
val wellknownService = FakeWellknownService()
wellknownService.givenMissingMapConfiguration()
val urlMapProvider = UrlMapProvider(wellknownService.instance, LocationSharingConfig("", true))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we should use named parameter in these tests as well so that it is clear of the boolean purpose.

Suggested change
val urlMapProvider = UrlMapProvider(wellknownService.instance, LocationSharingConfig("", true))
val urlMapProvider = UrlMapProvider(wellknownService.instance, LocationSharingConfig("", isMapTilerFallbackEnabled = true))

@@ -27,6 +27,7 @@ import im.vector.app.core.platform.EmptyViewEvents
import im.vector.app.core.platform.VectorViewModel
import im.vector.app.core.platform.VectorViewModelAction
import im.vector.app.features.VectorFeatures
import im.vector.app.features.location.UrlMapProvider
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This import should be removed.

@Johennes
Copy link
Contributor

Johennes commented Jan 3, 2023

Thanks for compiling the performance data @artkoenig. The CPU activity does look quite a bit inflated, sadly. Seems like a factor of 5 in the peaks. 😞

I think it's reasonable to allow forks to use the vector maps with style.json if they can accept the performance overhead. I don't think we should force this onto all FOSS users though, especially since with the default style.json there seem to be little to no differences between the static and the vector maps.

What seems better to me is to add support for both variants to the well-known file and let clients pick (which coincidentally is what we already had planned here).

WDYT?

@bmarty
Copy link
Member

bmarty commented Jan 3, 2023

Thanks for the PR. If I am correct it is addressing this comment?

@artkoenig
Copy link
Contributor Author

@bmarty actually not

@Johennes after an internal discussion, we decided to implement this internally due to time constraints. I will adjust the implementation to keep the differences to FOSS as minimal as possible to avoid merge conflicts. We will wait until you provide an official solution for configuring the static map API via the well-known. Feel free to close the PR or postpone it for later use.

@Johennes
Copy link
Contributor

Johennes commented Jan 5, 2023

Here's a screen recording of scrolling through a room with multiple location tiles on a Nokia 5.4 with a build from this PR:

vector.mp4

And here is a recording of the same action on the current store version of the app:

static.mp4

Unfortunately, the scrolling experience is quite poor with the vector maps compared to the static images.

@bmarty bmarty added the Z-Community-PR Issue is solved by a community member's PR label Jan 10, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Z-Community-PR Issue is solved by a community member's PR
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

5 participants