osm-oscar / oscar-web Goto Github PK
View Code? Open in Web Editor NEWA simple search engine for OpenStreetMap data
Home Page: https://www.oscar-web.de
License: Other
A simple search engine for OpenStreetMap data
Home Page: https://www.oscar-web.de
License: Other
Currently markers are drawn. But sometimes it is better to draw the geometry instead of the markers.
We should add an option to do that. Maybe add this to the query as well (as configuration option) so that users can customize their search shortcuts
Currently items that are not in a leaf node of the dag are not draw. This is unfortunately not so easy to fix since there may be many items that are not in a leaf node. Consider an island with few cities. The cities are leaf nodes but all the other stuff on the island is not in a leaf node.
This solves numerous problems.
Cppcms has no commits since April 2016. We may need to change our framework. Crow seems to be a promising alternative. Need to look into websockets as well.
The background of the visualization of the dag should have a opacity less than 1
Currently there is none. Hence most files are large, with bad code quality, no way to easily see what's in a file. This should be fixed.
Currently updating the map is not very asynchronous. It takes way too long to compute a new view. Unfortunately this not going to be easy to turn into an async function. This may only be possible after a larger refactor of the whole code.
On chromium zooming skips some levels. Firefox is ok.
Currently server and client need to have matching configs to ensure correct processing of requests. There should be some way to tell the server that it should generate smaller requests. Maybe a capabilities api would be good aswell.
Query: barfußpark
Zoom to Hallwangen -> Württemberg cluster with 12 hits stays on the map although items are visible. This should not be the case
Tracker bug. There are likely many things that need to be done to support this (if this is useful at all)
Currently there is no way to load more results into the result list. There should be some way to do this. Though there is the problem of changing result list since the result list depends on the map view. This might create strange paginations.
Some part of the website produces out of bounds coordinates in the region of 500 lat/lon
@tourism:hotel "new york"
List all hotels in New York
but @shop:clothes "Москва" will return 812 results in Russia but also this node in Kazachstan too, something is inaccurate here.
Inaccurate docs?
A human can only process a small number of things at once. For oscar-web the things are the clusters on the map. As a consequence there should be no more than about 5 clusters on the map.
The "Show children" link should only be visible if no children have been loaded yet
There seems to be something wrong with ItemIndexPrivateNative though the good looks correct.
Maybe we should calculate the center of clusters by using cluster centers of their children. This should improve zoom-consistency since the parent cluster center may be closer to the cluster center of the children.
Clicking on an Item in the results view does not open the detailed information. Tested with FF46
If one clicks on "OnePath" then the whole dag ist refreshed and any translation is discarded. Instead we should translate/scale the new dag such that the old node is at the same coordinates
And it should also work if only accept is pressed without finish before
If a large item is highlighted and the map viewport changes considerably it may and does happen that to view returns to a cluster view hence removing the item. That is something that needs to be avoided.
The question is how? Somehow we need know if the use just browses the map or wants to inspect a single element.
We currently do not have any kind of routing engine. This might be a nice addition. As a first step it should be possible to do normal routing but at a later stage we may want to do advanced routing taking heights, surrounding and so on into account.
Currently the map and result list updates are quite slow. About 1 msec per item. This is way to slow. This also make the website unusable on older hardware.
Add a new display "mode" that allows one to inspect multiple elements at once.
In this mode we want to keep inspected elements on the map.
Maybe we want to disable clustering.
Broken query: % $item:53357456
Use sigma-geosistemas/Leaflet.awesome-markers for markers?
Clicking on an object on the map does not highlight the item in the result list and vice versa. This seems to be a problem on Firefox.
Currently it may be the case that in a given cell items are drawn but a cluster icon of a region spanning that cell is drawn as well. This is the case if the cell is part of a region without children and one with children. This should be fixed in favor of the cluster icon. In order to do this the client needs to understand the concept of cells.
For regions with many cells that overlap other regions it might be useful to have cluster markers for cells. Consider the region A = {1, 2,3}, B = {1}, C={3}. For regions B and C get cluster markers. Hence we can't add a cluster marker for region A. But there are still items in cell 2.
Replace current usage of taginfo with our own version to retrieve top-k tags in our database
The counts of cluster icons that result from clustering of regions are wrong. It's not a good idea to just sum them together
Sometimes map objects and the result list are not in sync. Clicking on a map object then results in an exception since the result list entry does not exist. Need to investigate when this happens.
The query Aalen Döner has a lot of results in NRW in oscardev. But Aalen is in BW.
On oscardev the following query results in 100% cpu usage:
@ amenity:post_office vaihingen
Link: http://oscardev.fmi.uni-stuttgart.de/?q=%40amenity%3Apost_office%20vaihingen
You have to zoom in to Stuttgart. Then the loading animation appears and never goes away.
There are currently many possibilities for denial of service attacks by issuing queries that take very long to compute. We need some way to first filter these queries and at a later stage stop processing queries if they take too long.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.