Scrollable Mapview on tmtest

[Jim]

My first major Summer 2020 TM/METAL project has been to implement one of the features I've always wanted most myself: a version of Mapview that automatically loads all relevant TM data that's visible as the map is panned and zoomed.  I'm calling it "Scrollable Mapview."  I've made a ton of progress in the last several days and it's ready for some testing and feedback.

http://tmtest.teresco.org/user/mapview.php?v

Add a &u=username QS parameter if you also would like it to reflect your own clinched travels.

I'm most interested in hearing what you think so far and to report any problems you notice.  Please first check to see if it's something I already have noted with an Issue in the Web repository.  You can report problems by commenting on existing or creating new Issues there or by replying in this thread.

I'd also welcome general testing of existing Mapview functionality on tmtest and any other use there to make sure my work on the new Mapview functionality hasn't broken anything else.  It's pretty much a whole different codebase than the existing Mapview, so I'm confident I haven't broken much.  I'd like to fold this version back into the production site before continuing much further, which I hope will involve using the new codebase to replicate/improve existing Mapview functionality as well.

[Jim]

Note that the tmtest server doesn't switch databases to minimize disruptions during site updates, so if you notice problems with data not loading, try back in a few minutes as the DB update might be in progress.

[Jim]

Also, I'd like to call a bit of attention to the fact that waypoint markers are not shown in Mapview by default, but in the existing production version can be shown under the "Show Map Options" of the dropdown menu on the upper right.  The Scrollable Mapview implementation does not include waypoints, so can't show the markers.  If you care, please weigh in at https://github.com/TravelMapping/Web/issues/425 .

[mapcat]

https://drive.google.com/open?id=1X2p-NExZVjamjOi9GxGRDjnJZiELf9_D

The concurrency of US 20 & NY 146 shows as unclinched, even though I have 100% of US 20 according to the stats on the right.

This is happening anywhere a clinched route is concurrent with a route I haven't entered into my .list.

[michih]

- Select user in start box?

- Highlight clinched routes

- Clinched and overall mileage for loaded segments only?

[Jim]

https://drive.google.com/open?id=1X2p-NExZVjamjOi9GxGRDjnJZiELf9_D

The concurrency of US 20 & NY 146 shows as unclinched, even though I have 100% of US 20 according to the stats on the right.

This is happening anywhere a clinched route is concurrent with a route I haven't entered into my .list.

Thanks - quick fix is in.  It fixed the couple spots I checked.  Let me know if you see it cropping up elsewhere.

[mapcat]

Thanks - quick fix is in.  It fixed the couple spots I checked.  Let me know if you see it cropping up elsewhere.

Looks good now.

[yakra]

Gotta wonder how that happened, as siteupdate takes care of marking concurrencies as clinched, and all that info makes it into the DB.
I couldn't satisfy my curiosity on GitHub, as the code's not there yet. Which makes sense, with it all still undergoing heavy revision.

[michih]

I couldn't satisfy my curiosity on GitHub, as the code's not there yet. Which makes sense, with it all still undergoing heavy revision.

It is there: https://github.com/TravelMapping/Web/commit/5294e738d9792cb71ec926bc5ef36c30c653563a

[yakra]

Found it:
https://github.com/TravelMapping/Web/tree/scrollmap

[Jim]

Gotta wonder how that happened, as siteupdate takes care of marking concurrencies as clinched, and all that info makes it into the DB.
I couldn't satisfy my curiosity on GitHub, as the code's not there yet. Which makes sense, with it all still undergoing heavy revision.

As you found, the "scrolltest" branch has the latest.  I was indexing into an array to find out if the segment was clinched, but that array's indices had a different meaning.  Since no convenient array exists, I added a field to the connections so it was readily available when needed.

[Jim]

There's a new version up on tmtest that has stats for both the overall route and the visible portion of the route in the routes table.  Looking for everyone's impressions of this change.

[si404]

It's unsurprisingly slow-loading, but looks excellent.

Thanks for this.

[mapcat]

I like the table enhancements. Unfortunately I can still break it by panning too quickly.

https://drive.google.com/file/d/1SmZTwLbO9Tk7Y-a2gfRH0HLii2AR9Jjm/view?usp=sharing

I loaded an image centered on Pittsburgh, and the route table was correct. I zoomed out and panned west to Los Angeles without waiting for the route table to update. I didn't count how many times I clicked and dragged, but it must have been about 12-15. I zoomed in to the view shown in the image, and watched the route table slowly update with routes progressively farther west. It regenerated 10 times before the segments finally appeared, but never managed to show the routes actually in LA (in the image you can see that it quit after getting to Arizona). I left it alone while I uploaded the screenshot and it's now showing routes in Maryland and NYC, even though I never panned east after I loaded the Pittsburgh map (although I initially got to Pittsburgh by panning west from the mapview accessed by the link in the first post, centered at Siena).

Could you delay updating the map and route table by 1-2 seconds, to avoid loading any information while the user is still repositioning the map? Is that what you meant by "debouncing" in https://github.com/TravelMapping/Web/issues/423?

[Jim]

Yes, @mapcat, I believe the problem you describe will be addressed by debouncing.  I'm also looking into ways to speed the queries themselves, so more frequent reloads will not be so much of a problem.