Commit graph

1409 commits

Author SHA1 Message Date
artemp
1c3dbdd3ab postgis.input + use mapnik-geometry 2015-02-19 16:05:09 +01:00
artemp
9fed6fd571 shape.input - fix multi-polygon handling 2015-02-19 14:32:45 +01:00
artemp
33d45db318 c++ style - better syntax 2015-02-19 10:20:17 +01:00
Blake Thompson
5180168491 Merge branch 'master' into release/image_data_any 2015-02-18 23:03:51 -06:00
artemp
b49b8042c1 shape.input/agg - add polygon3 support 2015-02-18 18:12:47 +01:00
artemp
133f7d04ef agg_renderer - support multi_line_strings 2015-02-18 14:26:11 +01:00
artemp
8e34077aef more cleanups 2015-02-18 13:51:20 +01:00
artemp
604d7b08af use reserve to pre-allocate line_strings + cleanups 2015-02-18 13:48:05 +01:00
artemp
2e4e1f71b8 adapt to use mapnik::new_geometry::geometry (shape.input only) 2015-02-18 13:21:16 +01:00
Dane Springmeyer
a2dabb370d fix build+link of oracle plugin 2015-02-16 23:09:16 -08:00
Blake Thompson
b2c1c86d99 Merge branch 'master' into release/image_data_Any
Conflicts:
	benchmark/test_polygon_clipping.cpp
2015-02-16 14:37:01 -06:00
artemp
bb57ebd258 implement get_geometry_type() for 'large json' strategy
@springmeyer - we should talk re: validity of this approach. I know the baclground but still.
2015-02-12 15:50:13 +01:00
artemp
e8f397c41c instantiate geojson grammar using chr_iterator_type = char const* - via #2700 2015-02-12 11:51:11 +01:00
Dane Springmeyer
89a58c5879 Merge branch 'master' of github.com:mapnik/mapnik into release/image_data_any 2015-02-11 10:01:42 -08:00
Dane Springmeyer
7d2b34d654 finish fixing windows build - closes #2698 2015-02-10 19:02:32 -08:00
Dane Springmeyer
8608ea3308 only enable geojson memory mapped files if SHAPE_MEMORY_MAPPED_FILE defined (TODO: rename this define) - refs #2698 2015-02-10 18:50:27 -08:00
Dane Springmeyer
41ed56df14 Merge branch 'master' of github.com:mapnik/mapnik into vertex_adapter 2015-02-09 18:16:57 -08:00
Dane Springmeyer
578f8a7c29 report proj4 string in extra metadata for ogr plugin - refs #2650 2015-02-09 15:46:00 -08:00
artemp
e19fdad3a6 implement vertex interface in vertex_adapter and make mapnik::geometry_type immutable 2015-02-06 16:45:51 +01:00
Even Rouault
7c44b2412e gdal.input: take into account nodata_tolerance in optimized code path
Solves failures of https://travis-ci.org/mapnik/mapnik/jobs/49647112
on tiff-nodata-edge-rgba and tiff-nodata-tolerance tests
2015-02-06 00:41:57 +01:00
Even Rouault
efdca26f2d gdal.input: optimize nodata handling in RGB case
When the data type is Byte, instead of reading a first time
the red band to deduce the value of the alpha component of
the output image, read the RGB bands first.
2015-02-05 20:38:27 +01:00
Even Rouault
073bad21c0 gdal.input: use GDALDataset::RasterIO() to read RGB sources
Enabled when the R,G,B channels are bands 1, 2 and 3
And also read the alpha channel at the same time, if it is band 4
2015-02-05 20:33:52 +01:00
Blake Thompson
ffb34544e6 Merge branch 'master' into release/image_data_any 2015-02-04 15:46:54 -06:00
Blake Thompson
2b8bd59d82 A large set of updates:
* Added new gray data types adding those to the variants and updating all the code necessary for them
* Added basic SSE to the image compare method, (only for RGBA) must be enabled with the -DSSE_MATH flag this is not yet put into the build process in any location.
* Fixed the resulting image for some TIFF visual tests, most likely they were incorrect due to fixes in TIFF reader
* Added some MAPNIK_DECL where necessary to grid rendering.
* Added support for more data types in GDAL plugin with grayscale images.
* Added views for all the new gray data types
* Updated python bindings for new gray data types.

Ref #2681
2015-02-04 15:41:58 -06:00
artemp
eec661d263 geojson plugin
* use parse_geojson(Iterator start, Iterator end) signiture
* use Iterator = char const*
* remove unneeded I/O in 'initialise_index'
* use memory mapped file (currently #if/def)
2015-02-03 13:37:49 +01:00
Dane Springmeyer
71e7faed0d Merge branch 'master' of github.com:mapnik/mapnik into release/image_data_any
Conflicts:
	tests/python_tests/image_test.py
	tests/python_tests/image_tiff_test.py
	tests/visual_tests/images/tiff-opaque-edge-raster2-600-400-1.0-agg-reference.png
	tests/visual_tests/images/tiff-opaque-edge-raster2-600-400-2.0-agg-reference.png
2015-02-03 00:05:15 -08:00
Dane Springmeyer
192f62ecbf turn off default debug prints 2015-02-02 22:57:16 -08:00
Dane Springmeyer
5551cae0be cache GDALDataset object for re-use 2015-02-02 22:50:17 -08:00
artemp
89e516b493 unicode string grammar via boost/libs/spirit/example/qi/json/json/parser/grammar.hpp
update json based grammars
2015-02-02 11:22:54 +01:00
artemp
e3c96666c4 json - make feature_collection_grammar "on-feature" callback based 2015-01-29 12:33:18 +01:00
Dane Springmeyer
4b3f69e851 #2651 is fixed, remove workaround in postgis plugin 2015-01-28 18:18:09 -08:00
artemp
0d62e0f83c link to Oracle client library (libclntsh) to avoid missing symbols - ref #2659 2015-01-28 13:07:03 +01:00
Dane Springmeyer
1753d0ab47 require boost >= 1.56 for geojson/topojson plugins 2015-01-27 22:20:03 -08:00
Dane Springmeyer
d94f5a0cee get pgraster tests passing again + update expected images - refs #2639 2015-01-26 22:14:38 -08:00
Dane Springmeyer
48982c0619 Merge branch 'master' of github.com:mapnik/mapnik into release/image_data_any
Conflicts:
	tests/python_tests/image_tiff_test.py
2015-01-26 20:09:00 -08:00
artemp
2a50372d25 use correct grammar instantiation 2015-01-26 12:51:34 +01:00
artemp
a1c562c6fc cleanup 2015-01-26 12:51:10 +01:00
artemp
cc47bfd7fa extract attributes schema + remove std::cerr 2015-01-26 11:09:06 +01:00
artemp
2a3af1eb40 move boost.geometry <-- mapnik::box2d<double> into separate header for better modularity and code re-use 2015-01-26 10:37:09 +01:00
artemp
1263bc9c2e merge large geojson plug-in into geojson.input 2015-01-26 10:19:18 +01:00
artemp
44905ffc43 geojson.input - use adapted box2d<double> as box_type + make spatial index compatible with large_geojson 2015-01-26 09:42:06 +01:00
artemp
d45b9db86e remove unused include directives 2015-01-26 09:41:20 +01:00
Blake Thompson
31a0caf2e4 Merge branch 'master' into release/image_data_any 2015-01-25 14:09:29 -06:00
artemp
7f0e77d172 Merge branch 'master' into large-geojson 2015-01-23 18:12:33 +01:00
Dane Springmeyer
ec111c8a27 add extra metadata to lyr desc + use for srid/keyfield in postgis - closes #2650 2015-01-22 18:44:56 -08:00
Dane Springmeyer
faddf1a615 fix compile of rasterlite plugin 2015-01-22 15:59:01 -08:00
Dane Springmeyer
4a1f4a9b5e Use geometry_table_ rather than table_ for SRID lookup
- this will result in a faster query in the case that
   the user provides an explicit geometry_table option
2015-01-22 15:55:30 -08:00
Blake Thompson
51172c8fdf Move image_data.hpp to image.hpp and renamed most everything from image_data to image alone. This might lead to the need to clean up some variables that are currently named image through out the code at some time, but I think in the long term is much better as image is a better name for the base class. 2015-01-22 11:39:37 -06:00
artemp
58f770bf85 support BOOST_VERSION < 105600 2015-01-22 16:34:28 +01:00
artemp
e42d0deb61 correct types usage 2015-01-22 16:23:44 +01:00
artemp
2b82157127 cleanup 2015-01-22 16:21:01 +01:00
artemp
669d369e30 Merge branch 'master' into large-geojson 2015-01-22 16:12:14 +01:00
Blake Thompson
e01ce5b7d6 Changed image_data_gray* to image_gray* Ref #2633. 2015-01-21 21:08:04 -06:00
Blake Thompson
22a384ef33 Moved image_data_rgba8 to image_rgba8. Ref #2633 2015-01-21 20:31:02 -06:00
Blake Thompson
0f388ed68f Changed the name of image_data_any to image_any. Moved header file for image_data_any to image_any. Ref #2633 2015-01-21 19:40:12 -06:00
Dane Springmeyer
d6175adb1a Update README.md 2015-01-21 14:28:12 -08:00
artemp
9ba25cefe4 large_geojson - use 'C' style I/O for performance 2015-01-20 11:30:27 +01:00
artemp
85cacfe378 use vector<char> as a buffer 2015-01-20 10:36:34 +01:00
artemp
bfe1d63ea0 debug : print total extent 2015-01-19 22:06:24 +01:00
artemp
1f693b8fa2 larger_geojson - store {offset, size} in spatial index and remove multi_pass requirement
TODO: try std::fread instead of ifstream, use std::vector<uint8_t> instead of std::string
2015-01-19 18:10:59 +01:00
artemp
f50a34d1b8 use boost::spirit::iterator_policies::first_owner in multi_pass 2015-01-16 17:54:04 +01:00
artemp
da02dc40c2 sort index array by offsets 2015-01-16 17:19:59 +01:00
artemp
0c280b5066 reduce verbosity 2015-01-16 16:41:38 +01:00
artemp
788fd8c80b large_geojson.input - initial implementation 2015-01-16 16:22:46 +01:00
Blake Thompson
4184f75011 Moved premultiply and demultiply out of image_32 and other parts of the code. The image_data object is now responsible for keeping track of its own premultiplied_alpha status. Created a new utility method in image_util to preform premultiplication.
Added visitor pattern to several different methods as well to prepare for image_data_any including compositing.

Ref #2633
2015-01-14 12:42:30 -06:00
Dane Springmeyer
248c8a79c2 enable rasterlite error again, accidentally disabled in 62dbfea 2015-01-09 16:13:57 -08:00
Dane Springmeyer
62dbfeaa28 fix rasterlite plugin + add visual test - closes #2630 2015-01-09 16:13:04 -08:00
Dane Springmeyer
ac04512bb5 fix rasterlite linking when deps are static libs 2015-01-09 15:37:45 -08:00
artemp
5361d21beb move mapnik::noncopyable to mapnik::util::noncopyable where it belongs 2015-01-07 13:11:09 +01:00
artemp
18554ec0b1 remove static_visitor usage and rely on automatic result type deduction
(NOTE: expression_evaluator requires  ```using result_type = T1;``` )

Conflicts:
	src/image.cpp
2015-01-07 11:39:06 +01:00
artemp
ce291a553b I don't see enough performance gain to justify this - reverting for now
Revert "shape.input - sqeeze a bit more performance by reinterpret casting bytes array to point {double,double}"

This reverts commit 2f8426a019.
2014-12-18 18:44:04 +01:00
artemp
2f8426a019 shape.input - sqeeze a bit more performance by reinterpret casting bytes array to point {double,double}
avoiding one std::vector memory allocation - NOTE: this works because coordinates are stored in Double: Signed 64-bit IEEE double-precision floating point number (8 bytes) with little endian byte order - ref #2151
2014-12-18 11:57:52 +01:00
artemp
4f93b7c879 shape.input - avoid expensive hit_test and instead use relatively cheap orientation test to distinguish
between inner and outter rings (Polygon/NultiPolygon) -- ref #2151  https://github.com/mapbox/unpacker/issues/324

```txt
A polygon consists of one or more rings. A ring is a connected sequence of four or more
points that form a closed, non-self-intersecting loop. A polygon may contain multiple
outer rings. The order of vertices or orientation for a ring indicates which side of the ring
is the interior of the polygon. The neighborhood to the right of an observer walking along
the ring in vertex order is the neighborhood inside the polygon. Vertices of rings defining
holes in polygons are in a counterclockwise direction. Vertices for a single, ringed
polygon are, therefore, always in clockwise order. The rings of a polygon are referred to
as its parts.
```
2014-12-18 11:23:38 +01:00
artemp
6130b7f40d set nodata for single band raster (got missed in spaghetti code) 2014-12-17 14:23:42 +01:00
artemp
a59d634ffc nodata value can be floating point 2014-12-17 14:22:30 +01:00
artemp
5adbd85439 explicit std::move 2014-12-17 11:05:13 +01:00
artemp
1dfad4b10b raster.input - attempt getting bounding box from image_reader 2014-12-10 12:54:35 +01:00
Dane Springmeyer
cf4f775855 Merge branch 'master' of github.com:mapnik/mapnik into image_data_any 2014-12-09 10:31:19 -05:00
Dane Springmeyer
5504692a43 remove duplicate include 2014-12-09 10:30:55 -05:00
Rafa de la Torre
af4c2620aa Ignore overviews with 0 scale in pgraster
refs #2551

Postgis raster_columns view is returning NULL values for raster
overviews with large scale factors. That issue in postgis is described
in http://trac.osgeo.org/postgis/ticket/3006

This causes two main problems:

  - The first overview with scale = NULL is wrongly chosen for rendering
    always

  - The messed up scaling factor causes the render symbolizer to spent
    an insane amount of CPU and memory to render a messed up tiles.

The patch in postgis is expected to be released with the new version, a
few months from now.

Conflicts:
	plugins/input/pgraster/pgraster_datasource.cpp
2014-12-04 12:29:16 -05:00
Rafa de la Torre
6cb8b6e14b Ignore overviews with 0 scale in pgraster
refs #2551

Postgis raster_columns view is returning NULL values for raster
overviews with large scale factors. That issue in postgis is described
in http://trac.osgeo.org/postgis/ticket/3006

This causes two main problems:

  - The first overview with scale = NULL is wrongly chosen for rendering
    always

  - The messed up scaling factor causes the render symbolizer to spent
    an insane amount of CPU and memory to render a messed up tiles.

The patch in postgis is expected to be released with the new version, a
few months from now.

Conflicts:
	plugins/input/pgraster/pgraster_datasource.cpp
2014-12-04 11:09:44 -05:00
artemp
38da080f82 make concrete image_data type names more expressive and explicit about color channels/depth:
image_data_32 -> image_data_rgba8
image_data_8  -> image_data_gray8
image_data_16 -> image_data_gray16
image_data_float32  -> image_data_gray32f

NOTE: currently image_data_gray16::pixel_type = std::int16_t (signed 16 bit int) to match GDAL
should we support unsigned types?
2014-12-04 11:02:42 +01:00
artemp
de7a268333 read single band as int16 then agg_scale and colorize in gray16 color space, output rgb8 (aka image_data_32)
NOTE: proof-of-concept needed re-factoring/duplicate code removal
2014-12-03 18:19:44 +01:00
artemp
6fc8002dec use new image_reader API 2014-12-03 10:58:05 +01:00
Dane Springmeyer
46215c0f35 fix compile of pgraster plugin 2014-12-03 00:35:34 -05:00
artemp
6fec43bf9e gdal - read single band as image_data_float32
(scaling/compositing FIXME)
2014-11-28 12:51:23 +01:00
artemp
bf9c99e3d3 initial image_data_any implentation (work-in-progress) 2014-11-26 12:22:36 +01:00
artemp
33c1d21b7f use four letter abriviations to determine datatypes (https://www.sqlite.org/datatype3.html) 2014-11-25 11:09:42 +01:00
Artem Pavlenko
db4b74e8bb Merge pull request #2580 from StevenLooman/master
Tests and fix for SQLite datasource missing column
2014-11-24 12:51:55 +01:00
artemp
ee9481cb38 feature_collection_grammar : allow passing start_id at parse time - #2582 2014-11-24 12:26:11 +01:00
Dane Springmeyer
62d5cf7f16 fix error message in geojson plugin 2014-11-21 19:57:24 -08:00
Steven Looman
07a07c6268 Default to string type when column type could not be determined. Refs #2579 2014-11-20 18:48:31 +01:00
artemp
333965b588 update copyright year 2014-11-20 15:25:50 +01:00
Dane Springmeyer
437d486bbd fix building with PLUGIN_LINKING=static 2014-11-11 14:42:53 -08:00
Dane Springmeyer
4a2735fe91 c++ style: use reinterpret_cast rather than c style casts 2014-11-10 16:53:58 -08:00
Robert Coup
76d5bdae8c Merge pull request #2556 from stephend/master
Throw datasource_exception on GDAL RasterIO() failures
2014-11-05 18:21:13 +13:00
Dane Springmeyer
a65987af31 geojson plugin: fix compile against boost 1.57 2014-11-04 11:41:23 -05:00
Stephen Davis
824c99f565 throw datasource_exception on GDAL RasterIO failures 2014-11-03 11:24:38 +13:00
Dane Springmeyer
f34ac45b86 remove unneeded osm files - refs #2547 2014-10-27 08:56:58 -07:00
Jérémy Lal
87f9481621 Add missing argument to mapnik::raster call 2014-10-24 14:58:16 +02:00