8bdfc67f98
Rename for clarity.
2012-10-01 21:58:19 -07:00
d27a70f0c0
Refactor d3_geo_type.
2012-10-01 21:50:18 -07:00
bdc7c3f9e6
Add tests for d3.geo.bounds.
2012-10-01 20:39:38 -07:00
95c88ebd06
Enforce built-in precision for outline, too.
2012-10-01 18:20:14 -07:00
fb449fca26
Don't use function.apply; it's slower.
2012-09-29 15:54:07 -07:00
4d49ef72b3
More precision setter down.
2012-09-29 15:52:37 -07:00
a0ef36d5d5
Rename p to projection for clarity.
2012-09-29 14:11:12 -07:00
85041c616d
Replace projection.ring with projection.polygon.
...
This is now forwards-compatible should desire complex clipping.
2012-09-29 13:55:30 -07:00
e3b58c4550
Merge pull request #826 from jasondavies/projection
...
Only attempt to resample if distance < δ.
2012-09-28 08:53:49 -07:00
8d326607ff
Increase resampling min. distance threshold.
...
It seems reasonable to expect that the perpendicular distance will be
less than half the distance between two points, especially if they are
already close together.
2012-09-28 09:15:06 +01:00
71df2b80e6
Only attempt to resample if distance < δ.
...
This improves performance in the case where our input data is already of
a high enough resolution, even when subject to distortion. It’s unlikely
that the Douglas–Peucker estimate (perpendicular distance) will be
greater than δ if the two input points are already closer than δ. We
could probably even increase the threshold a little more e.g. 2 * δ
seems reasonable.
2012-09-28 08:57:26 +01:00
67c4bf8b3f
Increase default precision.
2012-09-27 15:29:36 -07:00
2616f03311
Fix a typo.
2012-09-27 15:21:51 -07:00
c2fdedd348
Adaptive using Douglas–Peucker estimate.
...
Use deviation from straight line rather than distance between points.
2012-09-27 15:14:39 -07:00
5d02d49877
Bah, fix another bug. Need tests, stat!
2012-09-27 12:39:23 -07:00
96d2f9adb0
Fix bug in previous commit.
2012-09-27 12:13:01 -07:00
815c76f79f
Setting path.context(null) restores default.
2012-09-27 12:11:20 -07:00
06cf988c5a
Expose raw projections.
2012-09-26 21:19:56 -07:00
99c444ddaf
path(null) should return null.
...
This frequently happens when a feature is fully clipped.
2012-09-26 21:01:42 -07:00
f36a6acdf3
Fix for precision(0).
2012-09-26 17:33:55 -07:00
e499edf533
Merge branch 'projection' of https://github.com/jasondavies/d3 into projection
2012-09-26 17:25:55 -07:00
650ebcffab
Fix for garticule.outline.
2012-09-26 12:23:32 -07:00
e40787c165
Expose d3.geo.azimuthalEqualArea.raw.
2012-09-26 11:51:49 -07:00
83af6c4c0b
Resampling for closePath, too.
2012-09-26 11:50:20 -07:00
3b480a2549
Expose d3.geo.projectionMutator.
2012-09-26 11:34:39 -07:00
f7e6dc93dc
Simplify.
2012-09-26 11:25:14 -07:00
04b85ef364
Make resampling precision configurable.
2012-09-26 11:25:14 -07:00
a91954bbac
Simplify.
2012-09-26 18:30:36 +01:00
f7174ce1a1
Make resampling precision configurable.
2012-09-26 17:29:44 +01:00
9ac8ad3c7d
Merge pull request #823 from jasondavies/projection
...
Interpolate polygon cuts correctly, and resample.
2012-09-26 07:58:33 -07:00
7af2884394
Improve interpolation at poles.
2012-09-26 14:57:55 +01:00
9c7e3a8d9a
Increase resampling recursion depth.
...
Otherwise Albers shows some graticule artefacts near the North Pole for
rotate([0, 0, 45]). It may be worth investigating what happens to k near
the poles, in case we need to override it.
2012-09-26 14:47:21 +01:00
e8b3ac10bc
Restore graticule precision.
...
This is not configurable for now, since resampling should do the rest.
The distortion of parallels is quite noticeable without this,
particularly nearer the poles (since parallels are not all great
circles), so we need *some* precision to begin with here.
2012-09-26 14:30:40 +01:00
9399591f9a
Simplify.
2012-09-26 14:16:03 +01:00
c7284d7dc3
Fix accuracy issue when interpolating near poles.
2012-09-26 14:14:13 +01:00
45d4b31461
Adaptive resampling.
...
Can still be optimised further by caching trig. results.
2012-09-26 13:35:51 +01:00
302576bbcd
Initial version of resampling.
...
This non-adaptively uses d3.geo.greatArc at the moment. This needs to be
rewritten to be adaptive, more efficient, and handle interpolating at
poles e.g. 〈-180°, -90°〉↦〈180°, -90°〉 should interpolate
longitudinally.
2012-09-26 12:18:04 +01:00
6b16dfa4cd
Checkpoint adaptive resampling.
2012-09-26 12:07:00 +01:00
b8a47f1c8c
Minor optimisation.
2012-09-26 10:01:32 +01:00
48dacb3599
Minor optimisation.
2012-09-26 09:59:49 +01:00
cc123a96be
Fix bug at rotate([62, -174, 0]).
...
Includes a minor optimisation.
2012-09-26 09:51:47 +01:00
e431bc5962
Another minor optimisation.
2012-09-26 09:29:01 +01:00
57ee53a232
Minor optimisation.
2012-09-26 09:23:58 +01:00
69b36c117f
Interpolate polygon cuts correctly.
2012-09-26 09:14:48 +01:00
e4ead4e008
Tiny optimization.
2012-09-25 17:15:11 -07:00
62fbc0c210
Add some TODO comments.
2012-09-26 00:35:02 +01:00
a59b718cfc
Better polygon cutting.
...
One more thing left to do: interpolate along the cut line when joining
endpoints. Ideally, the distance between interpolated points should be
less than 180°. This should probably also apply to points interpolated
at ±90° latitudes.
2012-09-26 00:28:27 +01:00
b12a5352d8
Remove unnecessary modulo.
2012-09-25 15:46:29 -07:00
448c3d085f
Extract static function.
2012-09-25 15:36:07 -07:00
1b7c291179
Merge branch 'projection' of https://github.com/jasondavies/d3 into projection
2012-09-25 15:29:29 -07:00
Mike Bostock
Jason Davies