Open-Harbor/osrm-backend
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 3 Wiki Activity
8,070 Commits 27 Branches 246 Tags 163 MiB
a67c4bf84d
Commit Graph

1432 Commits

Author SHA1 Message Date
Daniel Patterson
a67c4bf84d
Calculating durations is unavoidable due to tie-breaking minimums, but we can avoid accumulating distances if they're not requested. 2018-10-30 15:41:06 -07:00
Daniel Patterson
498259b220
Replace dynamic distance calculation for table plugin with pre-calculated distances on shortcuts, avoiding unpacking cost. 
Adds approx 10% to total data size.  Speeds up large table requests by 2 orders of magnitude.

Co-authored-by: Kajari Ghosh <ghoshkaj@gmail.com>
 2018-10-30 15:41:06 -07:00
Daniel Patterson
954121634f
Merge branch 'master' into ghoshkaj_mmaperize 2018-10-29 09:52:50 -07:00
Daniel Patterson
594a45e7e0
Re-use msinttypes bundled with RapidJSON when compiling on Windows. (#5249) 2018-10-29 09:51:56 -07:00
Daniel Patterson
b7e7d32361
Expose new --mmap switch (mmap_memory: true in NodeJS), and run test suite in this mode, as well as shared memory mode. 2018-10-26 23:54:00 -07:00
Daniel Patterson
2f9cb44368
mmap tarfiles directly when mmapping is enabled, instead of copying data into separate mmapped block 
Co-authored-by: Kajari Ghosh <ghoshkaj@gmail.com>
 2018-10-26 23:53:59 -07:00
Daniel Patterson
d80318f8ea
Match serialized bit-packing for vector<bool> to match in-memory layout for vector_view<bool> so that data can be directly mmapped. 2018-10-26 23:53:50 -07:00
Kajari Ghosh
5597415f28
Revert "Improve speed of Map Matching" (#5196) 
* Revert "Update changelog"

This reverts commit 9b779c704f.

* Revert "Fix formating"

This reverts commit 5bd7d04fe3.

* Revert "Fix bug in computation of distance offset for phantom node"

This reverts commit 0f78f7b2cc.

* Revert "Adjust text cases for flightly different matching due to rounding"

This reverts commit 8473be69d2.

* Revert "Round network distance to deci-meter to retain previous behavior"

This reverts commit c0124f7d77.

* Revert "Preserve heap state in map matching"

This reverts commit b630b4e32a.

* Revert "Use distance functions from many to many"

This reverts commit 89fabc1b9c.

* Revert "Use FCC algorithm for map matching distance calculation"

This reverts commit a649a8a5cf.
 2018-09-06 12:05:28 -04:00
Daniel Patterson
0971f06193 Add option to node bindings to return result as a pre-generated JSON string (this avoids a lot of overhead, and moves JSON string rendering out of the main event loop). 2018-09-05 15:09:13 -07:00
Daniel Patterson
85515f063a Render floating point numbers to string using Grisu2 algorithmt instead of stdlib to speed up JSON generation. 2018-09-05 14:20:47 -07:00
Daniel Patterson
69d7825542 Increase allowed shared memory regions to 512 from ~120 2018-09-05 11:48:02 -07:00
Patrick Niklaus
0f78f7b2cc Fix bug in computation of distance offset for phantom node 2018-08-02 12:19:24 +02:00
Patrick Niklaus
b630b4e32a Preserve heap state in map matching 2018-08-02 12:19:24 +02:00
Patrick Niklaus
89fabc1b9c Use distance functions from many to many 2018-08-02 12:19:24 +02:00
Patrick Niklaus
a649a8a5cf Use FCC algorithm for map matching distance calculation 2018-08-02 12:19:24 +02:00
Patrick Niklaus
c459530cb6 Address PR comments 2018-05-08 15:50:25 +00:00
Kajari Ghosh
2a15e6dec8 Add support for annotations=distances in MLD 
This commit brings feature parity with CH for the `table` pluging.
 2018-05-08 15:50:25 +00:00
Michael Krasnyk
cacb162520 Review and rebase updates 2018-04-28 08:59:18 +02:00
Michael Krasnyk
9695fa7941 Remove weight and duration from customizer::EdgeBasedGraphEdgeData 2018-04-28 08:59:18 +02:00
Michael Krasnyk
fd9bebbfa7 Add EBG node durations 2018-04-28 08:59:18 +02:00
Michael Krasnyk
db18e8669f Always read .osrm.enw file in updater 2018-04-28 08:59:18 +02:00
Michael Krasnyk
9b4a4fdd82 Add customizer::MultiLevelGraph 2018-04-28 08:59:18 +02:00
Michael Krasnyk
8adbfe06ed Throw an exception if a shared region is not found 2018-04-28 08:59:18 +02:00
Huyen Chau Nguyen
4e68f3a7e1 change haversine computation to cheap ruler in distance tables 2018-04-27 05:21:08 +02:00
Daniel Patterson
a6cf2eee7e Use faster method for calculating distances. 
fix cmakelist
 2018-04-27 05:21:08 +02:00
Kajari Ghosh
89f6e2d55b
Parse table annotations param correctly (#5050) 
* fix incorrect parameter parsing for node osrm and add tests

* fix boost spirit grammar parsing for annotations

* return NotImplemented when distance annotation is requested for MLD in table plugin

* update docs
 2018-04-24 11:05:35 -04:00
Michael Krasnyk
c628ecbf24 Update tile unit tests 2018-04-24 15:40:57 +02:00
Kajari Ghosh
14860b62e9
Unpack paths and return total distance in matrix plugin for CH (#4990) 2018-04-20 18:18:55 -04:00
Michael Krasnyk
b2aeb47630 Remove dummy edges before inplace permutation 2018-04-19 12:18:48 +02:00
karenzshea
e3b831364f empty list of shmem regions if none found 2018-04-13 12:07:10 +02:00
karenzshea
730d2b5ef2 log err instead of throwing when no shmem regions found 2018-04-13 12:07:10 +02:00
Michael Krasnyk
16abee1022 Avoid using signed integers for edge IDs 2018-04-12 10:03:37 +02:00
Michael Krasnyk
17cd1575f6 Use base-256 encoding for files larger 68G 
Reference:
http://lists.busybox.net/pipermail/busybox/2011-May/075596.html
 2018-04-10 21:12:39 +02:00
Patrick Niklaus
473ebfcbf6 Fix debug printing for coordinates 2018-04-10 18:08:18 +02:00
Michael Krasnyk
c4bf450fc6 Restore storage constructors with one arguments 2018-04-09 15:08:00 +02:00
Patrick Niklaus
bf2b45120a Use ArrayStorage for boundary nodes to optimize MLD 
For the MLD algorithm we can partition the NodeID range into boundary
and non-boundary nodes. Since there are only we boundary nodes we can
use the ArrayStorage for those yielding much faster query times.
 2018-04-09 15:08:00 +02:00
Patrick Niklaus
7edf0f218c Better statistics for osrm-partition and osrm-custimize 2018-04-09 13:39:39 +02:00
Patrick Niklaus
600ca06166 Simplify toEdges and make it more robust against accidental memory allocations 2018-04-09 13:03:28 +02:00
Michael Krasnyk
8d8042ebae Use ranges with fixed types 2018-04-09 11:05:01 +02:00
Kajari Ghosh
be123cd72f Use ranges in datafacade instead of vectors 
Range type must use immutable references due to a
regression in `boost::any_range`.
References:
https://svn.boost.org/trac10/ticket/10493
https://stackoverflow.com/questions/42427395/boostany-range-with-optimization-level-o2-causes-crash/42427662
 2018-04-09 11:05:01 +02:00
Patrick Niklaus
29d1b34140 Fix formating 2018-04-07 00:22:29 +02:00
Patrick Niklaus
a915542916 Add option to only update the metric dependent data 2018-04-07 00:22:29 +02:00
Patrick Niklaus
fea07f343b Load data in two separate data regions 2018-04-07 00:22:29 +02:00
Patrick Niklaus
c7daa521ad Address PR comments 2018-04-05 15:12:47 +02:00
Patrick Niklaus
4610fd9ff1 Fix formating 2018-04-05 15:12:47 +02:00
Patrick Niklaus
0e8b8b4901 Add nodejs test for dataset name 2018-04-05 15:12:47 +02:00
Patrick Niklaus
2c80f76004 Add support for naming the dataset 2018-04-05 15:12:47 +02:00
Patrick Niklaus
666ce46d36 Refactor shared memory storage to allow for multiple named datasets 2018-04-05 15:12:47 +02:00
Michael Krasnyk
bc120776f0 Don't apply unimplemented SH and PH conditions 2018-04-05 13:31:48 +02:00
Patrick Niklaus
51fe0dd5a0 Add missing check of a tar writing function 2018-04-05 00:54:50 +02:00