when I want to index a batch of points(about 240 million),some points got id like 18011072989319135231, when I do intersect ,I got overflowed id number like -435671084390416385. I think there is some problems in converting number from python to C lib.
And I also check the poi id is in the range of uint64. I check the spatial index lib, I found they use uint64 as the id types.

Hi everybody,

I started using rtree some weeks ago and find it very useful, thanks for providing a rtree in python!

I run into a problem recently that might affect other people as well. I am accessing a postgis database with approximately 5 mio rows. I can build a rtree without problems as long as I keep the number of objects below 1 million. The file sizes increase up to 300MB for the data and 800kb for the index.

When I try to store more than approximately 1 mio entries I get this error:

Traceback (most recent call last):
  File "/home/christoph/dev/my_project/trees.py", line 116, in create_from_db
    tree = FastRtree(TREE_FILENAME, generator_function(), property=tree_property)
  File "/home/christoph/.virtualenvs/my_project/lib/python2.7/site-packages/rtree/index.py", line 235, in __init__
    self.handle = self._create_idx_from_stream(stream)
  File "/home/christoph/.virtualenvs/my_project/lib/python2.7/site-packages/rtree/index.py", line 691, in _create_idx_from_stream
    return core.rt.Index_CreateWithStream(self.properties.handle, next)
  File "/home/christoph/.virtualenvs/my_project/lib/python2.7/site-packages/rtree/core.py", line 25, in check_void
    raise RTreeError(msg)
RTreeError: Error in "Index_CreateWithStream": Tools::BufferedFileWriter::open: Cannot open file.

Does someone have a clue about what is happening? File creation is definitely not a problem when I create a smaller tree.

From PyPI, the latest release is 0.8.2, but on GitHub the latest tagged release is 0.8.1. Shouldn't there be a 0.8.2 tag for that release?

Our license is not being included in the source tree because of this.

Now that #18 is closed, Rtree works on Python 3.4 and this warrants a new release. Will tag and upload to PyPI after a little more testing with GeoPandas.

cc @hobu

I am getting the following error when trying to create a new disk based index:

@sgillies is going to come up with some json-like key/value pairing for this so we can easily see the property sets. Maybe even set them?

This is more of a question rather than a issue. Pardon me if this is a wrong place to ask. For my thesis, I want to modify the rtree. For that I need references to root or internal nodes so that I can traverse the tree. I debugged the code and couldn't find a tree construction. Can some one help me?

I am new to libspatialindex python package.

From original RTree Paper.

The CustomStorage needs to register a flush() callback to keep up with changes to IStorageManager for libspatialindex 1.8.0+

import rtree
import random
import time

def createDefaultIndex():
idx = rtree.index.Rtree(u"defaultIndex")

make boxes....

boxes = []
for boxIndex in range(0,100):
xmin = random.random()
ymin = random.random()
xmax = xmin + random.random()
ymax = ymin + random.random()
boxes.append((xmin,ymin,xmax,ymax))

for boxIndex, box in enumerate(boxes):
pickledObject = "makethisbig " * 100000
idx.insert(boxIndex, box, obj=pickledObject)

del idx

def readIndexDefault():
indexPath = u"defaultIndex"
#idx = rtree.index.Rtree(indexPath)
idx = rtree.index.Index(indexPath)
boundingCoords = [0,0,60,60]
print "Consuming Index"
startTime = time.time()
#hits = list(idx.intersection(boundingCoords, objects=True)) # memory leak
#hits = list(idx.intersection(boundingCoords, objects="raw")) # no memory leak
hits = list(idx.intersection(boundingCoords, objects=False)) # no memory leak
#hits = list(idx.intersection(boundingCoords))
endTime = time.time()
print "Consuming time: ", endTime - startTime
print "Hits: ", len(hits)
time.sleep(10)

print "Done Consuming - erasing memory"
idx = None
hits=None
for counter in range(0,10):
time.sleep(1)

what we have learned:

if objects == true, we have memory leak. hits, idx each have

the size of the dat file and those are released. however, there is

a 3rd memory allocation for the size of the dat file, mysterious, that

is not released.

if object != true, then there is no memory leak.

if object == "raw" then no memory leak

if name == "main":
createDefaultIndex()
readIndexDefault()

It looks like when doing stream loading for a 3D index the tree is created fine but the IDs aren't preserved for some reason. This issue does not exist when doing tree.insert inside a loop.

I tried this with a generator function as per the documentation, as well as passing the bounds as a tuple instead of a list. I verified that in index/_create_idx_from_stream/py_next_item it was getting the correct ID into p_id[0]. I'm not sure what else to check, any ideas would be greatly appreciated.

Here's an example which was demonstrating the issue using rtree from Toblerity/rtree/master and libspatialindex from the latest packages in the Ubuntu 14.04 repo:

import numpy as np                                                                  
import rtree                                                                        

if __name__ == '__main__':                                                          
    properties = rtree.index.Property()                                             
    properties.dimension = 3                                                        

    points_count  = 100                                                             
    points_random = np.random.random((points_count,3,3))                            
    points_bounds = np.column_stack((points_random.min(axis=1),                     
                                     points_random.max(axis=1)))                    

    stacked = zip(np.arange(points_count),                                          
                  points_bounds,                                                    
                  [None] * points_count)                                              

    tree = rtree.index.Index(stacked,                                               
                             properties = properties)                               

    # this should intersect every box, as our random boxes are all inside [0,0,0,1,1,1]                   
    tid = list(tree.intersection([-1,-1,-1,2,2,2]))                                 

The correct answer for tid would be np.arange(points_count). The actual output of this is:

In [6]: len(tid)
Out[6]: 100

In [7]: tid[0:10]
Out[7]: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]

In [8]: (np.array(tid) == 0).all()
Out[8]: True

Thanks for providing RTree for Python!

I was wondering if rtree can be used in a multithreaded context?

Importing both arcpy and rtree causes the arcpy.ArcSDESQLExecute function to break. No other functions seem affected (that I have come across). Update it seems any functions connection to SDE are broken.

This occurs for the Python 2.7 packages from http://www.lfd.uci.edu/~gohlke/pythonlibs/ for both x32 and x64. Using ArcGIS 10.1 on Windows 7.

I know arcpy modifies some system paths etc. but does rtree have any globals that could cause issues with arcpy?

import arcpy
import rtree
sde_conn = arcpy.ArcSDESQLExecute(r"C:\Temp\conn.sde")
Traceback (most recent call last):
File "", line 1, in
File "C:\Program Files (x86)\ArcGIS\Desktop10.1\arcpy\arcpy\arcobjects\mixins.py", line 893, in init
_BaseArcObject.init(self, *gp_fixargs((server, instance, database, user, password), True))
File "C:\Program Files (x86)\ArcGIS\Desktop10.1\arcpy\arcpy\arcobjects_base.py", line 47, in init
for arg in args))
RuntimeError: Object: Could not open SDE workspace C:\Temp\conn.sde

hi,
i'm trying to use rtree in python3. i'm using the version installed by:

pip3 install rtree

on my ubuntu 15.04.

creating a (disk stored) index, i got an error about wrong type in set_dat_extension. had a look at the code and noticed that, at least in my version, there is:

    v = value.encode('utf-8')
    return core.rt.IndexProperty_SetFileNameExtensionDat(self.handle, value)

so, you encode the string into bytes but after you still pass the original string to the c function. i fixed this, here as well as in set_idx_extension, passing v instead of value. i also modified get_idx_extension and get_dat_extension adding a .decode('utf-8') in order to return strings and not bytes (without this step there were other errors about adding bytes to strings...). that solved the problem.
i guess you'll probably want to reflect the fix on your sources.

and thank you for the great job done so far!

Scenario:

• loading about 15 million polygons into rtree which are spread over about 60 shapefiles.
• I'm using the generator style insert
• about 2 million polygons in, inserts slow dramatically then segmentation faults.
• I'm main guess is that I am running out of memory
• Am I abusing rtree? Is there a way to flush to disk every so often?

Thanks

I am using rtree to find the nearest city by lat,lon. But on Nginx server after few requests I am getting following error.

  File "/home/ubuntu/core/greedy/utilities.py", line 88, in current_location
    city = GeoLocation.CityByLatLong(location[0], location[1], idx)

  File "/home/ubuntu/core/greedy/apps/geolocations/models.py", line 13, in CityByLatLong
    near_list = idx.nearest((latitude, longitude), 1, objects=False)

  File "/usr/local/lib/python2.7/dist-packages/rtree/index.py", line 567, in nearest
    p_num_results)

ArgumentError: argument 5: <type 'exceptions.TypeError'>: expected LP_LP_c_longlong instance instead of pointer to LP_c_longlong

PS: On apache server, it is working perfectly well.

I'm using travis-ci to automatically build and test my project, which uses rtree.

travis-ci encounters an error when installing rtree, or rather libspatialindex ( see
https://travis-ci.org/goulu/Goulib/builds/24608247 ) :

rtree
    Traceback (most recent call last):
      File "<string>", line 17, in <module>
      File "/home/travis/virtualenv/python2.7.6/build/rtree/setup.py", line 5, in <module>
        import rtree
      File "rtree/__init__.py", line 1, in <module>
        from index import Rtree
      File "rtree/index.py", line 2, in <module>
        import core
      File "rtree/core.py", line 85, in <module>
        rt = ctypes.CDLL(lib_name)
      File "/opt/python/2.7.6/lib/python2.7/ctypes/__init__.py", line 365, in __init__
        self._handle = _dlopen(self._name, mode)
    OSError: libspatialindex_c.so: cannot open shared object file: No such file or directory
    Complete output from command python setup.py egg_info:
    Traceback (most recent call last):
  File "<string>", line 17, in <module>
  File "/home/travis/virtualenv/python2.7.6/build/rtree/setup.py", line 5, in <module>
    import rtree
  File "rtree/__init__.py", line 1, in <module>
    from index import Rtree
  File "rtree/index.py", line 2, in <module>
    import core
  File "rtree/core.py", line 85, in <module>
    rt = ctypes.CDLL(lib_name)
  File "/opt/python/2.7.6/lib/python2.7/ctypes/__init__.py", line 365, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: libspatialindex_c.so: cannot open shared object file: No such file or directory
----------------------------------------
Cleaning up...
Command python setup.py egg_info failed with error code 1 in /home/travis/virtualenv/python2.7.6/build/rtree
Storing debug log for failure in /home/travis/.pip/pip.log
The command "pip install --allow-all-external -r requirements.txt" failed and exited with 1 during install.

any idea why ? thanks !

hallo,
i have the following test code:

from rtree import index.Index

def test():
    for i in range(10):
        yield (i,(i,i,i,i),{'i':i})

idx=Index('/tmp/idxtest',test())
for i in idx.nearest((0,0,0,0),20,objects=True):
    print( i.id, i.bounds, i.object )

and the output is:

0 {'i': 0} [0.0, 0.0, 0.0, 0.0]
0 {'i': 1} [1.0, 1.0, 1.0, 1.0]
0 {'i': 2} [2.0, 2.0, 2.0, 2.0]
0 {'i': 3} [3.0, 3.0, 3.0, 3.0]
0 {'i': 4} [4.0, 4.0, 4.0, 4.0]
0 {'i': 5} [5.0, 5.0, 5.0, 5.0]
0 {'i': 6} [6.0, 6.0, 6.0, 6.0]
0 {'i': 7} [7.0, 7.0, 7.0, 7.0]
0 {'i': 8} [8.0, 8.0, 8.0, 8.0]
0 {'i': 9} [9.0, 9.0, 9.0, 9.0]

... all ids are 0!
i have verified the ids arrive unchanged at least until here, but what happens after is a little too hard to track for me.

Hi,

I'm a new user of the R-tree package. I'm reading the documentation and I notice that stream loading is much faster in creating a R-tree than individual inserts. I have a bout 21 million bboxes I need to insert. If I insert individually using index.Index('rtree'), I see the the index/dat file created. But when I try
index.Index(filename='rtree', stream=foo()), the rtree get initialized but it not written to file. I'm not sure if this usage mode is supported?

Is there some sample code somewhere I can look at on how to create a file-based rtree w/ stream loading?

Thanks.

There is an error during installation of the pip 0.8.2 version. Here is part of the error message:

  File "/tmp/pip_build_root/Rtree/setup.py", line 4, in <module>

    import rtree

  File "rtree/__init__.py", line 1, in <module>

    from .index import Rtree

  File "rtree/index.py", line 6, in <module>

    from . import core

  File "rtree/core.py", line 110, in <module>

    rt.Error_GetLastErrorNum.restype = ctypes.c_int

  File "/usr/lib/python2.7/ctypes/__init__.py", line 378, in __getattr__

    func = self.__getitem__(name)

  File "/usr/lib/python2.7/ctypes/__init__.py", line 383, in __getitem__

    func = self._FuncPtr((name_or_ordinal, self))

AttributeError: /usr/bin/python: undefined symbol: Error_GetLastErrorNum

While tracking down some verson compatibility issues I noted in a Gentoo ebuild I developed a couple of years ago that RTree cannot work on python version 3.0+. Has this been fixed? Also related to this, would you like RTree to be added to the Gentoo standard distribution?

Thanks and best regards,

John 9EBo) David --

I've installed spatialindex/1.8.5 from source and I'm trying to compile rtree 0.8.2 against it. Running

python setup.py install 

tries to import rtree and fails in ctype

Traceback (most recent call last):
  File "setup.py", line 4, in <module>
    import rtree
  File "/glade/scratch/ddvento/build/Rtree-0.8.2/rtree/__init__.py", line 1, in <module>
    from .index import Rtree
  File "/glade/scratch/ddvento/build/Rtree-0.8.2/rtree/index.py", line 6, in <module>
    from . import core
  File "/glade/scratch/ddvento/build/Rtree-0.8.2/rtree/core.py", line 110, in <module>
    rt.Error_GetLastErrorNum.restype = ctypes.c_int
  File "/glade/apps/opt/python/2.7.7/gnu-westmere/4.8.2/lib/python2.7/ctypes/__init__.py", line 378, in __getattr__
    func = self.__getitem__(name)
  File "/glade/apps/opt/python/2.7.7/gnu-westmere/4.8.2/lib/python2.7/ctypes/__init__.py", line 383, in __getitem__
    func = self._FuncPtr((name_or_ordinal, self))
AttributeError: python: undefined symbol: Error_GetLastErrorNum

My plan to let rtree know about spatialindex was via the C compiler, but it failed before. How should I proceed?

Please do not use find_library from ctypes to find libspatialindex_c. This only works if it is installed on in a standard location. On HPC clusters, it is very likely that the library is not installed in a standard location. Adding the installation directory to LD_LIBRARY_PATH, or specifying the installation directory in the setup.cfg file should be sufficient, but it does not work.

The only way to make rtree find libspatialindex_c on a Linux HPC cluster is to "hack" core.py, comment out the line

lib_name = find_library('spatialindex_c')

and replace lib_name in

rt = ctypes.CDLL(lib_name)

with the absolut path of the library.

I understand that the problem is introduced by ctypes and not by the rtree package, but please use another mechanism than find_library to detect libspatialindex_c on Linux systems. The two most practical options would be to either use LD_LIBRARY_PATH or the setup.cfg file.

Thank you very much and best regards

Samuel Fux

I'm new to rtree so sorry if this is a basic question, but is there away to leverage streamloading to a file-based rtree?

The constrcutor to rtree.Index checks if the first argument is a basestring or a generator function. Is there a way to provide both?

For example, if I have the generator function

def generator_function():
... for i, obj in enumerate(somedata):
... yield (i, (obj.xmin, obj.ymin, obj.xmax, obj.ymax), obj)

And I want to save to "somefile"

r = index.Index("somefile", generator_function())

results in an exception.

Using rtree 0.8.2. The small example below recreates the problem.

!/usr/bin/env python

from rtree import index

def gen(d):
for i,obj in enumerate(d):
yield (i, obj[0], obj[1])

data = [((0,0,1,1), 25),
((1,1,2,2), 30),
((2,2,3,3), 35)]

idx = index.Index('foo', gen(data))

m = idx.intersection((2,3,3,3), objects=True)
for a in m:
print a.id
print a.bbox
print a.object

It correctly prints the correct bbox/object but the wrong id (should be 2). It seems like when stream loading is used to create the index, the ID is always returned as 0. It seems to work OK if idx.insert() is used to create the tree. Thanks.

$ ./test.py
0
[2.0, 2.0, 3.0, 3.0]
35

Hi, if I run the following:

import rtree
import cPickle
x = rtree.index.Index()
cPickle.loads(cPickle.dumps(x))

I get a segfault when I quit python, (or at another random time later, depends how long it runs for). I've tried using plain 'pickle' as well, but get the same behaviour.

I'm using rtree 0.8.2 (installed via pip) against libspatialite 4.2.0-2 on arch linux x64, python 2.7.9

I'm trying to implement a custom storage for rtree, and get the following exception:

Traceback (most recent call last):
  File "test.py", line 5, in <module>
    tree = rtree.index.Rtree(RTreeRedis())
  File "/Users/gunnar/projects/rtree-redis/env.rtreeredis/lib/python2.7/site-packages/rtree/index.py", line 225, in __init__
    storage.registerCallbacks( self.properties )
  File "/Users/gunnar/projects/rtree-redis/env.rtreeredis/lib/python2.7/site-packages/rtree/index.py", line 1158, in registerCallbacks
    properties.custom_storage_callbacks      = ctypes.cast( ctypes.pointer(callbacks), ctypes.c_void_p )
  File "/Users/gunnar/projects/rtree-redis/env.rtreeredis/lib/python2.7/site-packages/rtree/index.py", line 1028, in set_custom_storage_callbacks
    return core.rt.IndexProperty_SetCustomStorageCallbacks(self.handle, value)
  File "/Users/gunnar/projects/rtree-redis/env.rtreeredis/lib/python2.7/site-packages/rtree/core.py", line 17, in check_return
    raise RTreeError(msg)
rtree.core.RTreeError: LASError in "IndexProperty_SetCustomStorageCallbacks": The supplied storage callbacks size is wrong, expected 56, got 48

I based my code on this: https://github.com/Toblerity/zope.index.rtree/blob/master/rtree/storage.py

and since it fails already when constructing the rtree index, very little of my code actually runs.

This on python 2.7.5 on OSX 12.4.0, i.e. x86_64

Thanks for any pointers!

As far as I can tell from the documentation and the code, Rtree lets you either specify a filename for disk-based index, or a stream for stream loading. But how can I do both? What if I want a disk-based index which is populated by reading from a stream?

When checking nearest neighbours, rtree can return more than the specified number of results, as if two points are equidistant, it returns both them. To check this equidistance, it must have some threshold since the distances are floats. I want to be able to control this threshold.

Could anyone explain this error?
Thank you in advance.
M@rco

2015-04-10 12:41:17 ERROR Zope.SiteErrorLog 1428662477.860.832295083576
Traceback (innermost last):
...........
Module Products.LinguaPlone.patches, line 34, in searchResults
Module Products.CMFPlone.CatalogTool, line 393, in searchResults
Module Products.ZCatalog.ZCatalog, line 604, in searchResults
Module Products.ZCatalog.Catalog, line 925, in searchResults
Module Products.ZCatalog.Catalog, line 547, in search
Module collective.geo.index.geometryindex, line 169, in _apply_index
Module collective.geo.index.index, line 69, in rtree
Module rtree.index, line 211, in init
Module rtree.index, line 985, in get_overwrite
Module rtree.core, line 42, in check_value
RTreeError: Error in "IndexProperty_GetOverwrite": Property IndexIdentifier was empty

Hey is this of any interest to submit back as a pull request?

import zlib

from rtree import Rtree

class CompressedRtree(Rtree):
    '''
    adds zlib compression to rtree
    '''

    def __init__(self, *args, **kwargs):
        Rtree.__init__(self, *args, **kwargs)

    def dumps(self, obj):
        result = super(CompressedRtree, self).dumps(obj)
        return zlib.compress(result)

    def loads(self, string):
        return super(CompressedRtree, self).loads(zlib.decompress(string))

after doing nearest with objects=False, I end up with a list of IDs.

Is there a way to later go back and retrieve either the raw object or an Item from this id? i.e. index.get_item_by_id ?

I would maybe also be nice to have a way to get the id and bounds, without getting the objects

Trying to set GeoPandas up with Rtree support today I have run into the following:

RTreeError: Error in "Index_CreateWithStream": IllegalArgumentException: RTree::BulkLoader::bulkLoadUsingSTR: Empty data stream given.

If the generator you send to the bulk loader turns out to be empty, say you filter all its element out for good reasons, leaving only an expression equivalent to

idx = Index(iter(()))

the bulk loader quits and raises an exception. More graceful failure would be better.

There is a bug in the index dumping logic when you do not have cPickle available. If you use this simple script:

#!/usr/bin/env python

from rtree import index
import pickle as pickle

idx = index.Index()
left, bottom, right, top = (0.0, 0.0, 1.0, 1.0)
idx.insert(0, (left, bottom, right, top), "SomePickleable")
print list(idx.intersection((1.0, 1.0, 2.0, 2.0)))
re_idx = pickle.loads(pickle.dumps(idx))
print list(re_idx.intersection((1.0, 1.0, 2.0, 2.0)))

It runs fine with cPickle available:
$ ./rtree_bug.py
[0]
[0]

But if you force things to use (regular) Pickle:
mv lib/python2.7/lib-dynload/cPickle.so lib/python2.7/lib-dynload/cPickle.so.outoftheway

You get a protocol error:
$ ./rtree_bug.py
...
raise ValueError("pickle protocol must be <= %d" % HIGHEST_PROTOCOL)
ValueError: pickle protocol must be <= 2

The problem is in the dumps( ) aliasing, which I have a pull request for.

Pointed out by @hobu in #21 (comment).

Largely due to naive changes I made while working on Python 3.4. Should be easy fixes.

HI,

Im not expert with linux and c libraries, so probaly my description is incomplete or not very handful.
Anyway Im experiencing a problem with libspatialindex.so.3.0.0 and rtree 8.2.
Now and then my plone instance dies with no clues and message. I've checked the dmesg report from linux and here is what I see:

[1476438.713131] traps: python[317] general protection ip:7fb04f5a2ff9 sp:7fb04a3440b0 error:0 in libspatialindex.so.3.0.0[7fb04f506000+b8000]
[1476763.870627] python[379]: segfault at 0 ip 00007f855e9ccaea sp 00007f85409640f8 error 4 in libc-2.19.so[7f855e944000+1bb000]
[1478334.385962] traps: python[1211] general protection ip:7ff88dbf5ff9 sp:7ff889198390 error:0 in libspatialindex.so.3.0.0[7ff88db59000+b8000]
[1526321.346560] traps: python[17381] general protection ip:7f33d0f48ff9 sp:7f33cc4eb390 error:0 in libspatialindex.so.3.0.0[7f33d0eac000+b8000]
[1527794.507131] traps: python[20645] general protection ip:7f8ccb8c6ff9 sp:7f8cc5766390 error:0 in libspatialindex.so.3.0.0[7f8ccb82a000+b8000]
[1528432.553798] python[20984]: segfault at 22 ip 00007f8f4de55ff2 sp 00007f8f404f6390 error 4 in libspatialindex.so.3.0.0[7f8f4ddb9000+b8000]
[1528892.762885] traps: python[21271] general protection ip:7f97170eeff2 sp:7f9711e90390 error:0 in libspatialindex.so.3.0.0[7f9717052000+b8000]
[1529146.029769] python[21388]: segfault at 0 ip 00007f29b5d25aea sp 00007f298bff9a88 error 4 in libc-2.19.so[7f29b5c9d000+1bb000]
[1529364.634798] python[21435]: segfault at 0 ip 00007fa394931aea sp 00007fa3780cc148 error 4 in libc-2.19.so[7fa3948a9000+1bb000]
[1529636.648433] python[21533]: segfault at 0 ip 00007f7c14cbaaea sp 00007f7bf37f6af8 error 4 in libc-2.19.so[7f7c14c32000+1bb000]

Im working on a linux instance (a VPS):

Linux 3.13.0-46-generic #79-Ubuntu SMP Tue Mar 10 20:06:50 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

Any suggestion is very appreciated. Thank you in advace.

I can't seem to get the install to work for me. Any thoughts about what I can do to debug this?

Building rtree on OS X either from the current release or from master results in a failure:

$ pip install git+git://github.com/Toblerity/rtree.git
Downloading/unpacking git+git://github.com/Toblerity/rtree.git
  Cloning git://github.com/Toblerity/rtree.git to /var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build
  Running setup.py (path:/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/setup.py) egg_info for package from git+git://github.com/Toblerity/rtree.git
    Traceback (most recent call last):
      File "<string>", line 17, in <module>
      File "/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/setup.py", line 4, in <module>
        import rtree
      File "/private/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/rtree/__init__.py", line 1, in <module>
        from .index import Rtree
      File "/private/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/rtree/index.py", line 6, in <module>
        from . import core
      File "/private/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/rtree/core.py", line 110, in <module>
        rt.Error_GetLastErrorNum.restype = ctypes.c_int
      File "/usr/local/Cellar/python3/3.4.2_1/Frameworks/Python.framework/Versions/3.4/lib/python3.4/ctypes/__init__.py", line 364, in __getattr__
        func = self.__getitem__(name)
      File "/usr/local/Cellar/python3/3.4.2_1/Frameworks/Python.framework/Versions/3.4/lib/python3.4/ctypes/__init__.py", line 369, in __getitem__
        func = self._FuncPtr((name_or_ordinal, self))
    AttributeError: dlsym(RTLD_DEFAULT, Error_GetLastErrorNum): symbol not found
    Complete output from command python setup.py egg_info:
    Traceback (most recent call last):

  File "<string>", line 17, in <module>

  File "/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/setup.py", line 4, in <module>

    import rtree

  File "/private/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/rtree/__init__.py", line 1, in <module>

    from .index import Rtree

  File "/private/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/rtree/index.py", line 6, in <module>

    from . import core

  File "/private/var/folders/z_/kgsm76k94s7_jhnkfydfnd380000gn/T/pip-p4_3c96t-build/rtree/core.py", line 110, in <module>

    rt.Error_GetLastErrorNum.restype = ctypes.c_int

  File "/usr/local/Cellar/python3/3.4.2_1/Frameworks/Python.framework/Versions/3.4/lib/python3.4/ctypes/__init__.py", line 364, in __getattr__

    func = self.__getitem__(name)

  File "/usr/local/Cellar/python3/3.4.2_1/Frameworks/Python.framework/Versions/3.4/lib/python3.4/ctypes/__init__.py", line 369, in __getitem__

    func = self._FuncPtr((name_or_ordinal, self))

AttributeError: dlsym(RTLD_DEFAULT, Error_GetLastErrorNum): symbol not found

Running Python 3.4.2 on OS X 10.10.

I am getting the following error when attempting installation on Ubuntu. I have tried via pip as well as downloading and installing from the github repo.

Here is the error:
Downloading/unpacking Rtree
Downloading Rtree-0.8.0.tar.gz (56kB): 56kB downloaded
Running setup.py (path:/tmp/pip_build_root/Rtree/setup.py) egg_info for package Rtree
Traceback (most recent call last):
File "", line 17, in
File "/tmp/pip_build_root/Rtree/setup.py", line 4, in
import rtree
File "rtree/init.py", line 1, in
from .index import Rtree
File "rtree/index.py", line 31, in
int(t) for t in c_api_version.decode('utf-8').split('.')]
ValueError: invalid literal for int() with base 10: 'SIDX_VERSION_MAJOR'
Complete output from command python setup.py egg_info:
Traceback (most recent call last):

File "", line 17, in

File "/tmp/pip_build_root/Rtree/setup.py", line 4, in

import rtree

File "rtree/init.py", line 1, in

from .index import Rtree

File "rtree/index.py", line 31, in

int(t) for t in __c_api_version__.decode('utf-8').split('.')]

ValueError: invalid literal for int() with base 10: 'SIDX_VERSION_MAJOR'

is it possible to get the points exacly in the same coordinate of the query input (not the nearest)?

Hello,

Perhaps I am missing some critical step, but I am attempting to serialize a 5-d index to disk using the standard method. Example code follows below. The problem I am running into is that when I close the index on disk and attempt to open again, the 5-d nature of the index appears to be lost. My Rtree version is 0.7.0.

import numpy as np
import rtree

random = np.random.random((5000,5))

def gen(r):
    for i in xrange(r.shape[0]):
        yield (i, (r[i,0], r[i,1], r[i,2], r[i,3], r[i,4], 
                   r[i,0], r[i,1], r[i,2], r[i,3], r[i,4]), None)

p = rtree.index.Property()
p.dimension = 5
p.overwrite = True
tree = rtree.index.Index('test2', gen(random), properties=p)
print tree.properties.dimension
print tree.bounds
tree.close()

Result:
5
[0.0001399796428322997, 0.00012241192644635301, 7.3393899523988e-05, 0.0005358454728167095, 0.00015215451107841726, 0.9997935298037376, 0.9995067343095075, 0.9997580576087635, 0.9999079765788539, 0.999747647405568]

tree = rtree.index.Index('test2')
print tree.properties.dimension
print tree.bounds
tree.close()

Result:
2
[9.68817621993523e-06, 0.0001726496530731758, 3.065078523922793e-05, 0.0001633907971153592, 1.7523149461129428e-05, 0.9998708500048494, 0.9999690660514665, 0.9999390843951469, 0.9999013700514435, 0.9999794343619967]

Hey Toblerity,
I'm investigating a library called hickle which allows you to use hdf5 to store objects. https://github.com/telegraphic/hickle
I was trying to get this going and store rtrees inside of HDF for the compression options. Any first impressions of this? Do you think its as easy as adding an additional import statement and using hickle on index.dumps() & index.loads()? Any thoughts would be appreciated.

Sadly, I haven't found any out there.

is this library available for python 3.x as well?
Thanks

To get the memory leak fix out there, among other things.

cc @hobu

I am trying to build rtree with conda and make it pass all the tests but the test_stream_input() is failing.

Instead of [0, 4, 16, 27, 35, 40, 47, 50, 76, 80] the test is returning a list with all zeros. I am not familiar with rtree's code, but it seems that the zeros are coming from py_next_item.

I think this snippet from line 104 of rtree/core.py only works by coincidence.

elif os.name == 'posix':
    platform = os.uname()[0]
    lib_name = 'libspatialindex_c.so'
    if platform == 'Darwin':
        lib_name = 'libspatialindex_c.dylib'
        free = ctypes.CDLL(find_library('libc')).free
    else:
        free = ctypes.CDLL(find_library('libc.so.6')).free

If you're actually looking for libc with the find_library function, you'd want find_library('c'), but this code works because:

• find_library returns None if it fails to find the library
• dlopen (and thus, CDLL) returns a handle that can access libc if you pass None to it as the library parameter.

So, as far as I can tell, this:

    if platform == 'Darwin':
        lib_name = 'libspatialindex_c.dylib'
        free = ctypes.CDLL(find_library('libc')).free
    else:
        free = ctypes.CDLL(find_library('libc.so.6')).free

could be replaced with:

    if platform == 'Darwin':
        lib_name = 'libspatialindex_c.dylib'
    free = ctypes.CDLL(None).free

Please do correct me if I'm wrong -- or I guess if this is a deliberate choice for some reason?

Help. I have problem when instal Rtree on my windows.
I used cmd for install it.
But it's seem cannot found spatialindex_c.dll
Actually, i have spatialindex_c-64.dll
what sholud i do with it?

this is what i have done:
C:\Python27\Scripts\easy_install.exe Rtree
Best match: Rtree 0.8.2
Downloading https://pypi.python.org/packages/source/R/Rtree/Rtree-0.8.2.tar.gz#md5=4ba7276cbd4c51b373068004ae864d3
Processing Rtree-0.8.2.tar.gz
Writing c:\users.....\Rtree-0.8.02\setup.cfg
Running Rtree-0.8.2\setup.py -q dbist_egg --dist-dir ......
error: could not find or load spatialindex_c.dll

And I tried to download Rtree-0.8.2 and install it offline too, but stiil could not find or load spatialindex_c.dll.

On the mailing list Howard shows the way how to do this:

The base libspatialindex library itself absolutely supports this through it its visitor pattern, however. Presumably there would need to be some sort of callback-based method, much like the way the DataStream stuff in the C API works in the reverse direction, to connect up the Python side to the results visitor.