aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
27 hoursimplement OsmAnd/MbTiles in python interfaceosmand-sqliteDaniel Schadt
27 hoursrun CI on non-master branchDaniel Schadt
27 hoursupdate docsDaniel Schadt
27 hoursimplement mb-tiles output formatDaniel Schadt
2 daysreplace space with =Daniel Schadt
to make it consistent with the other options
2 daysadd osmand sqlite output format (lib & cli)Daniel Schadt
2025-12-14lazily load tile in blit_nonzeroHEADmasterDaniel Schadt
In the benchmarks, this gives a quite good speed increase, but in real-life workloads, the speedup is not as stark (from ~66s to ~61s here). What is maybe a neat secondary effect here is that we don't end up with empty tiles, as we only allocate (in tile_mut) when there is a nonzero pixel to blit. This decreases the number of saved tiles, especially on larger zooms.
2025-11-29update dependenciesDaniel Schadt
2025-11-29fix license expressionv0.1.0Daniel Schadt
2025-11-29slight readme adjustmentsDaniel Schadt
2025-11-29add some more documentationDaniel Schadt
2025-11-29actually install python on ciDaniel Schadt
2025-11-29install python on ciDaniel Schadt
2025-11-29enable woodpecker ciDaniel Schadt
2025-11-29Merge branch 'tests'Daniel Schadt
2025-11-29more tests for heatmap rendererDaniel Schadt
2025-11-28apply cubic scaling to heat countDaniel Schadt
2025-11-27add preliminary test for heatmap_prepareDaniel Schadt
2025-11-27small speedup for colorize_tileDaniel Schadt
This brings 1. a color lookup table so we avoid re-computing the gradient every time 2. better iterating over the pixels, so we don't need an explicit get_pixel_mut
2025-11-26make merge_heat_counter more efficientDaniel Schadt
Accessing the underlying buffers directly and iterating over them is more efficient (and gives the compiler more options to optimize) than the dance with get_pixel_mut. On my local tests with 119 tracks, this shaves off ~30 seconds (from 90 seconds down to 60 seconds).
2025-10-15update PyO3 versionDaniel Schadt
Should support Python 3.14 now
2025-08-26add first testsDaniel Schadt
2025-08-26properly clamp Coordinates::to_mercatorDaniel Schadt
2025-08-25Merge branch 'py'Daniel Schadt
2025-08-12add type hintsDaniel Schadt
It seems like the "simple" way with just the pyi file doesn't work, so we need to add the py.typed manually. Also, enjoy this path: hittekaart/hittekaart-py/hittekaart_py/hittekaart_py.pyi
2025-08-02string repr for Track and StorageDaniel Schadt
2025-08-02custom min_zoom/max_zoom/threadsDaniel Schadt
This replaces the set_threads function, and works even better because we don't rely on a global thread pool!
2025-07-29implement other renderers in python moduleDaniel Schadt
2025-07-29more docsDaniel Schadt
They are actually visible as Python docstrings, so they are useful!
2025-07-28add some docstringsDaniel Schadt
They are not really visible but useful documentation anyway.
2025-07-28expose function to set thread countDaniel Schadt
2025-07-28implement sqlite output in hittekaart-pyDaniel Schadt
We cannot use a Box<dyn Storage> anymore, because the Sqlite connection is not thread-safe. Therefore, we use a normal enum and open the connection late.
2025-07-09first working Python interfaceDaniel Schadt
2025-07-08add public constructor for coordinatesDaniel Schadt
2025-07-08fix wrong doc linksDaniel Schadt
2025-06-26split crate into core and cliDaniel Schadt
2023-10-14update dependenciesDaniel Schadt
2023-05-13add more benchmarksDaniel Schadt
2023-05-13fix benchmarkDaniel Schadt
2023-05-13update dependenciesDaniel Schadt
2023-03-25update READMEDaniel Schadt
2023-03-25enable LTODaniel Schadt
2023-03-13implement "proper" tile hunter modeDaniel Schadt
Now with fixed zoom level for the hunting.
2023-03-11add reference to small PNG craftingDaniel Schadt
2023-03-11first implementation of tile hunter modeDaniel Schadt
This is a very basic idea, the downside is that you basically cannot "fix" the tilehunt overlay to a specific zoom level. It might be nice to get a second renderer that renders the tile hunt for a fixed zoom, and then adjusts the output tiles accordingly (then it also takes more space again though).
2023-03-11abstract away tile rendering logicDaniel Schadt
This is in prepration for the tilehunt mode, where we want to render tiles differently.
2023-03-11update dependenciesDaniel Schadt
2023-01-18add some more docstringsDaniel Schadt
2023-01-18add support for SQLite outputDaniel Schadt
2023-01-17factor out saving logicDaniel Schadt
Since we want to support SQLite at some point, it makes sense to have the exact storage method abstracted away.