You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One of the NL-Net tasks is to speed up data loading
Introduction
Currently, MapComplete loads all data from Overpass (or from the OSM-AP if zoomed in). For
some themes, it would be nice to see globally how much data is available and to be able to load
the data quickly. While Overpass is an amazing service and allows extremely flexible queries, it
often takes a long time to load the data – especially on a global level.
A service like MapComplete doesn’t need this flexibility: once a thematic visualization is created, it
is well-known what features should be loaded. A service could then cache all these features and
keep these layers up-to-date with the minutely diffs. This also ties in with the internal handling of
data: within MapComplete, the ‘FeaturePipeline’ is a complicated piece of code which merges the
data from different sources (overpass, fallback-overpass, cache, external sources, the actual
OSM-API). It contains several bugs, and issues, for example: it can’t handle deleted points nor can
and emits many duplicate features in some cases.
Tasks
Find and adapt a vector tile generating server, with a data layer for every mapcomplete-
layer
Refactor the feature pipeline and feature sources to remove unneeded code
Fix caching behaviour and behaviour of deleted items
As this issue is mostly for project management, no discussions are allowed here. Please use #581 for technical discussions
The text was updated successfully, but these errors were encountered:
One of the NL-Net tasks is to speed up data loading
Introduction
Currently, MapComplete loads all data from Overpass (or from the OSM-AP if zoomed in). For
some themes, it would be nice to see globally how much data is available and to be able to load
the data quickly. While Overpass is an amazing service and allows extremely flexible queries, it
often takes a long time to load the data – especially on a global level.
A service like MapComplete doesn’t need this flexibility: once a thematic visualization is created, it
is well-known what features should be loaded. A service could then cache all these features and
keep these layers up-to-date with the minutely diffs. This also ties in with the internal handling of
data: within MapComplete, the ‘FeaturePipeline’ is a complicated piece of code which merges the
data from different sources (overpass, fallback-overpass, cache, external sources, the actual
OSM-API). It contains several bugs, and issues, for example: it can’t handle deleted points nor can
and emits many duplicate features in some cases.
Tasks
layer
As this issue is mostly for project management, no discussions are allowed here. Please use #581 for technical discussions
The text was updated successfully, but these errors were encountered: