35d8ebca22
use emplace where possible
2019-04-10 22:34:01 -04:00
9ec714063a
many performance tweaks
2019-04-10 22:34:01 -04:00
2cedda703c
Fixed up iterative methods to use struct
2019-04-10 22:34:01 -04:00
d814710575
Added emplace
2019-04-10 22:34:01 -04:00
60d07287c9
made KeyInfoEntry into a struct
2019-04-10 22:34:01 -04:00
1a862c24a6
Reserve memory for cache
2019-04-10 22:34:01 -04:00
50d9e1ef1c
numerical expected results rather than regression
2019-04-10 22:34:01 -04:00
99aa09ce54
Merged in jlblancoc/gtsam/index-type-for-factors (pull request #409 )
...
Type for Factor indices, dual to "Key"
2019-04-08 23:07:02 +00:00
4fb718a943
prefer auto in range for loops
2019-04-09 00:29:31 +02:00
60b2c3b157
Merged in fix/SubgraphPreconditioner (pull request #407 )
...
Fix/SubgraphPreconditioner
2019-04-08 21:21:36 +00:00
285ebd7dbd
Type for Factor indices, dual to "Key"
...
This avoids a couple of confusing uses of KeySet to refer to lists of
Factors, and makes code more readable where formerly using size_t to
index factors.
2019-04-08 09:08:43 +02:00
1304d26e81
exit()-> throw ( Closes #427 )
2019-04-07 23:59:53 +02:00
d012922043
Cleanup, made solve faster (eliminating copy) and no longer in-place
2019-04-07 13:15:05 -04:00
1c2646000b
Added tests for matrix/vector conversion
2019-04-07 13:15:05 -04:00
42284355f4
Some cleanup, renaming
2019-04-07 13:15:05 -04:00
798e730d87
Fixed issue where number of extra "augmentation" edges asked for are larger than the number of remaining edges after taking out the spanning tree.
2019-04-07 13:15:05 -04:00
18d26d12af
Added some xml test files
2019-04-07 13:15:05 -04:00
6b637bda9e
Cleanup
2019-04-07 13:15:05 -04:00
334c85a298
Using keys not indices
2019-04-07 13:15:05 -04:00
a7826ab417
Added comments, removed cruft
2019-04-07 13:15:05 -04:00
9581e4939b
Made it so X can be variable dimension as long as you know dimension of tested value at compile time.
2019-04-07 13:15:05 -04:00
3737474d1e
Deprecated all but three constructors.
2019-04-04 09:23:02 -04:00
140c666c41
Moved DSFMap to gtsam
2019-04-04 09:22:22 -04:00
6d938ce5cc
Replaced initialize calls with C++11 delegating constructors
2019-04-04 01:17:28 -04:00
485175e2f8
Fixed major bug: splitting off a subgraph preconditioner only worked if keys were numbered 0...n-1, because we used DSFVector to implement Kruskal. Now it'll be a bit slower but will work for any keys. Note this only affected two constructors.
2019-04-04 01:04:18 -04:00
89ebed53cc
fixed comments
2019-04-03 23:42:47 -04:00
3e10adb178
Cleaned up tests, testing size() as well.
2019-04-03 23:42:36 -04:00
3126979ad5
Fixed memory issue (passing temporary to optional reference)
2019-04-03 22:45:49 -04:00
5a8363a775
Removed Ordering again -> templated vector method simply works
2019-04-03 20:17:18 -04:00
c450222ff1
test on ordering
2019-04-03 20:16:37 -04:00
ecaf415d1e
Better tests on backSubstituteTranspose
2019-04-03 18:45:16 -04:00
59df91d295
Added optional ordering argument when converting to Matrix/Vector
2019-04-03 18:44:18 -04:00
1365a04b09
Added default noise model argument in two crucial factors.
2019-03-25 23:01:05 -04:00
1e944fb86a
now initialization is aware of rotation noise model
2019-03-19 18:44:09 -04:00
e2cf42773a
Switched to struct with static methods as apparently global methods in namespaces are not wrapped.
2019-03-19 17:04:31 -04:00
37eba50932
Modernized, documented
2019-03-19 15:47:01 -04:00
5af2256277
Added missing clone for MATLAB wrapper
2019-03-19 10:30:30 -04:00
90f5dae3d0
Merge remote-tracking branch 'origin/develop' into feature/optimization_params
2019-03-19 10:21:26 -04:00
205803a0ea
Better optimization parameter wrapping, plus python test
2019-03-19 00:11:45 -04:00
88ac6de4af
Wrapped parse3DFactors
2019-03-14 01:25:06 -04:00
a47c52cb5e
Split parsing and moved to dataset.*
2019-03-14 00:58:22 -04:00
d8ee79fb8f
Working parseG2o3D
2019-03-14 00:27:02 -04:00
0ca3f9d199
Use c+11 initializer lists
2019-03-13 23:22:37 -04:00
6723b481a6
fix missing Eigen in Cython wrapper
2019-02-17 01:02:17 +01:00
40d67f49da
Merged in jlblancoc/gtsam/more-cmake-modernizations (pull request #379 )
...
A lot of cmake modernizations as well as cleanup some cmake issues/docs.
close issue #425
Approved-by: Chris Beall <chrisbeall@gmail.com>
2019-02-15 21:24:09 +00:00
7625c21777
Various fixes to cmake exported targets
...
List of changes:
* -I boost is no longer required
(Since the use of Boost::xxx imported targets)
* fix missing Boost deps in imported gtsam by
searching for Boost inside GTSAMConfig.cmake
* Including the dirs for Eigen/MKL/SuiteSparse/Metis
into exported targets public interface.
* Fix missing cmake changes in wrap/*
* Split build flags into private/public, not to
expose to users flags that may be invasive.
* Removed now useless include_dirs in "extra cmake"
* Update cmake/example_project
* Make cppunitlite to find boost headers via Boost::boost
* Update README / INSTALL to reflect the updated minimum CMake >= 3.0
2019-02-15 22:04:04 +01:00
37ed46cee1
Added underscore to case conflicts until wrap can handle this.
2019-02-11 10:44:24 -05:00
c0686f1f09
Fixed findExampleDataFile issue. Note the wrapped version is not the one that will be available in MATLAB. Still have to test whether we can use that or not.
2019-02-11 10:43:48 -05:00
b81f845f16
Merged in jlblancoc/gtsam (pull request #372 )
...
port to target-based CMake compile options and definitions
Approved-by: Chris Beall <chrisbeall@gmail.com>
2019-01-30 04:05:50 +00:00
5acd89f634
Non-include of <iostream> caused compilation errors related to std::cout usage
2019-01-29 00:28:54 -08:00
Frank Dellaert
José Luis Blanco-Claraco
jlblancoc
Frank Dellaert
lucacarlone
Mike Sheffler