python package for finding the timezone of any point on earth (coordinates) offline
Home Page: https://timezonefinder.michelfe.it/
License: MIT License
I am Jannik, a data scientist with a strong background in Python, hyperparameter optimisation and scientific computing.
I love to design and implement efficient algorithms especially in abstract domains.
My other interests include: AGI, computational geometry, knowledge representation and reasoning, machine and reinforment learning.
With my open source projects I want to give back to the community I owe so much to.
You can contact me on LinkedIn or buy me a coffee.
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
Make algorithm work when closest polygon is on the 'other end of earth'.
Until now polygons across the +-180deg longitude border are not being considered.
To support this, the distances would need to be computed differently.
Don't check every coordinate of the possible polygons, but rather preselect them.
(The computation of the distance from one line section of a polygon is quite complicated)
I had the idea to only select the sequence of points in the polygon which actually face the input point, but I did not come up with a valid algorithm for this (since polygons could be of various shapes and too complicated algorithms might take away the benefit of optimization)
At first i wanted to just look for the points with the most extreme phase (angle) from the input point, but then I realized that the polygon may surround the point and the phase might overflow ( e.g. from > pi to <pi). So this approach did not work.
With the following coordinates: 42.88783N 93.63952E, located in China (https://www.openstreetmap.org/?mlat=42.88783&mlon=93.63918#map=5/42.88783/93.63918), using:
from timezonefinder import TimezoneFinder
from pytz import timezone
import pytz
from datetime import datetime, timedelta
utc = pytz.utc
tf = TimezoneFinder()
def get_timezone_offset(lat, lon):
"""
returns the supposed offset from UTC in hour.
Stolen from https://github.com/MrMinimal64/timezonefinder
"""
today = datetime.now()
tz_target = timezone(tf.certain_timezone_at(lat=lat, lng=lon))
today_target = tz_target.localize(today)
today_utc = utc.localize(today)
return (today_utc - today_target).total_seconds() / 3600The following code returns 6.0 while calling:
>>> get_timezone_offset(42.88783, 93.63952)
6.0the expected result is 8.0, since China is in the same timezone (GMT+8) for the whole country, relying on https://www.timeanddate.com/time/map/
Tested with timezonefinder v3.0.1
Hi,
Is it possible to run this program directly from a command line (bash prompt)?
Something like:
python timezonefinder.py lat=45 lon=-1
Best,
Chris
Avoid problems like #37 from gotcha'ing anyone ever again. There's no real standard for the order of lat and lng so we can fix this with keyword args only
this function is not yet covered in the tests.
Make use of the pycc feature of Numba:
http://numba.pydata.org/numba-doc/latest/user/pycc.html
It would be ideal if the helper functions were compiled on installation on the user's system once (when Numba is installed). This would probably speed up the algorithms.
I encountered the issue that on my system the produced (compiled) code is for the wrong target architecture and hence cannot be run.
I would be glad if someone else could tell me if they are encountering the same issues.
A draft for how this could work is already included in the latest version of timezonefinder.
Procedure would be:
I'm trying to track down a potential issue (either with my own code base (more likely) or the library (less likely)). When running unit tests on a Django application that makes use of the timezonefinder library, I'm getting the following warning printed out:
C:\Python27\lib\site-packages\timezonefinder\helpers.py:42: RuntimeWarning: overflow encountered in long_scalars
if (x1GEx and x2GEx) or ((x1GEx or x2GEx) and (y2 - y) * (x2 - x1) >= (y2 - y1) * (x2 - x)):
My first thought is that I was somehow passing it bad data, which would seem logical given the warning - if you pass a really large value for y (or x), it could result in an overflow. That makes sense. But what doesn't make sense it it seems almost non-deterministic. If I put a single six.print() statement inside helpers.py, just before that statement, the warning goes away. I am at a bit of a loss as to why this would change anything. Is the six library somehow changing precision of something? Is there any way to tell the timezonefinder library to run using more bits to get around this?
For some reason certain_timezone_at() fails for a small location in the middle of Saskatchewan on version 2.1.2. This can be recreated with the following snippet:
from timezonefinder import TimezoneFinder
tf = TimezoneFinder()
tf.certain_timezone_at(lat=53.0742, lng=-108.9086) # Returns none. Should return u'America/Regina'
closest_timezone_at works for this location, so we were able to fall back to that, but for some reason that particular latitude/longitude fail. Given that Saskatchewan is basically a giant rectangle, this caught me off guard. Is this perhaps an issue with the underlying data? It shouldn't be anywhere near a coastline, so that shouldn't be an issue.
Looks like there might be a missing dependency in v4.0.0. Just importing the package throws a ModuleNotFoundError for importlib_resources
Steps to reproduce:
pip install timezonefinderfrom timezonefinder import TimezoneFinderObserve ModuleNotFoundError.. image:: https://anaconda.org/conda-forge/timezonefinder/badges/version.svg :target: https://anaconda.org/conda-forge/timezonefinder
Might me interesting for some users (using pytzwhere).
When I run closest_timezone_at, I thought that it would return one distance value per timezone. However, when I run it I sometimes get multiple results for one timezone. For example, the following:
tf.closest_timezone_at(lng=-85.2692, lat=31.6575, delta_degree=8,return_distances=True, force_evaluation=True, exact_computation=True)
returns two hits for the New York timezone
('America/New_York', array([ 14.21944659, 754.67091603, 749.29385281, 741.37317121, 704.2011185 , 724.31508828, 699.01969573, 781.7355791 , 551.5451686 , 771.86319838, 14.21944659, 791.60531965]), ['America/Chicago', 'America/Indiana/Indianapolis', 'America/Indiana/Petersburg', 'America/Indiana/Vincennes', 'America/Kentucky/Louisville', 'America/Indiana/Marengo', 'America/Indiana/Tell_City', 'America/Indiana/Vevay', 'America/Kentucky/Monticello', 'America/Nassau', 'America/New_York', 'America/New_York'])
Is this a bug or a feature?
s. the section for it on
http://efele.net/maps/tz/world/
Do people need this?
I am trying to understand the implications of using one vs the other, and have a few questions:
With memory mode = false, wouldn't using timezonefinderL be equivalent to using timezonefinder's timezone_at, in terms of speed and memory load? Since that function supposedly uses only shortcuts, same as timezonefinderL.
Wouldn't it be possible to provide a "limited" memory mode that only loads shortcuts for timezonefinder, to get the best of both world: fast lookups in most cases, and the option to do a deeper lookup when the tz returned is None.
using from timezonefinder import TimezoneFinder. Returns a type error. Not sure if it is OS dependent? I am currently running a Win 8.1, python 2.7 (Anaconda dist)
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-187-ffb6df41970c> in <module>()
----> 1 from timezonefinder import TimezoneFinder
C:\Users\Shetty\Anaconda\lib\site-packages\timezonefinder\__init__.py in <module>()
1 from __future__ import absolute_import, division, print_function, unicode_literals
2
----> 3 from .timezonefinder import TimezoneFinder
4
5 __version__ = '1.5.6'
C:\Users\Shetty\Anaconda\lib\site-packages\timezonefinder\timezonefinder.py in <module>()
15
16 if numba is not None:
---> 17 from .helpers_numba import coord2int, distance_to_polygon_exact, distance_to_polygon, inside_polygon, all_the_same
18 else:
19 from .helpers import coord2int, distance_to_polygon_exact, inside_polygon, all_the_same, distance_to_polygon
C:\Users\Shetty\Anaconda\lib\site-packages\timezonefinder\helpers_numba.py in <module>()
6
7
----> 8 @jit('i1(i4,i4,i4,i4,i4,i4)', nopython=True, cache=True)
9 def position_to_line(x, y, x1, x2, y1, y2):
10 """tests if a point pX(x,y) is Left|On|Right of an infinite line from p1 to p2
C:\Users\Shetty\Anaconda\lib\site-packages\numba\decorators.pyc in jit(signature_or_function, locals, target, cache, **options)
147 targetoptions=options)
148 if pyfunc is not None:
--> 149 return wrapper(pyfunc)
150 else:
151 return wrapper
C:\Users\Shetty\Anaconda\lib\site-packages\numba\decorators.pyc in wrapper(func)
161 return _DisableJitWrapper(func)
162 disp = dispatcher(py_func=func, locals=locals,
--> 163 targetoptions=targetoptions)
164 if cache:
165 disp.enable_caching()
C:\Users\Shetty\Anaconda\lib\site-packages\numba\dispatcher.pyc in __init__(self, py_func, locals, targetoptions)
274 self.targetctx = self.targetdescr.target_context
275
--> 276 pysig = utils.pysignature(py_func)
277 arg_count = len(pysig.parameters)
278
C:\Users\Shetty\Anaconda\lib\site-packages\funcsigs\__init__.pyc in signature(obj)
57
58 if not callable(obj):
---> 59 raise TypeError('{0!r} is not a callable object'.format(obj))
60
61 if isinstance(obj, types.MethodType):
TypeError: u'i1(i4,i4,i4,i4,i4,i4)' is not a callable object
When trying to import timezonefinder with numba installed (without numba it works fine) following error is raised:
LoweringError: Failed in nopython mode pipeline (step: nopython mode backend)
Cannot lower constant of type 'readonly bytes(uint8, 1d, C)'
File "../../../.local/lib/python3.6/site-packages/timezonefinder/helpers_numba.py", line 312:
def rectify_coordinates(lng, lat):
<source elided>
if lng > 180.0 or lng < -180.0 or lat > 90.0 or lat < -90.0:
raise ValueError(b'The coordinates should be given in degrees. They are out ouf bounds.')
^
[1] During: lowering "$const32.2 = const(bytes, b'The coordinates should be given in degrees. They are out ouf bounds.')" at python3.6/site-packages/timezonefinder/helpers_numba.py (312)
This was most probably introduced in 3.4.2 by fix of #70 (does not occur in 3.4.1)
Readme says current data set in use: precompiled timezone-boundary-builder (without oceans).
All my points are in ocean, Is there any way I can use data with ocean?
I tried several locations in sea. It returns None.
From our unit tests:
troupe.utils.geo_lookup: ERROR: Failed to get timezone for 13,100
Traceback (most recent call last):
File "/home/saltmaster/wod.git/troupe/src/utils/geo_lookup.py", line 41, in timezone_at_location
tz_name = TIMEZONE_FINDER.closest_timezone_at(lat=lat, lng=lon)
File "/home/saltmaster/conda/envs/troupe/lib/python3.6/site-packages/timezonefinder/functional.py", line 27, in wrapper
return func(*args, **kwargs)
File "/home/saltmaster/conda/envs/troupe/lib/python3.6/site-packages/timezonefinder/timezonefinder.py", line 442, in closest_timezone_at
while pointer < polygons_in_list:
UnboundLocalError: local variable 'pointer' referenced before assignment
Also, it's just super clear from reading the code.
Hi MrMinimal64,
I came across your timezonefinder and realized it would be perfect for my project, so I tried using it. I have been trying to use the timezone_at() method, but I receive an answer of 'uninhabited' for every location I put in.
To test this further, I tried using the geo-coordinates of New York, NY for timezone_at(), certain_timezone_at(), and closest_timezone_at(), and received a response of 'uninhabited' for all three. The response to my TimezoneFinder.using_numba() is False. My version of timezonefinder is 1.5.6
Any clues as to how fix this would be greatly appreciated. Thanks for the hard work on this library; it seems like it would be really great if it worked properly for me!
Is this really necessary? Queries are quite fast already with numba and parallelism introduces new challenges and dependencies...
For closest_timezone_at() it could offer a big additional speed boost, because distances from a the polygons are independent and (almost) all need to be computed.
Also with the timezone_at algorithms it may be interesting to check if the point is inside of multiple polygons at once and just stop if one matches.
What about:
http://www.parallelpython.com/
This should work automatically if the dependencies are all installed (like with numba). And there should be a function for checking if parallelism is being used.
Hi, this might not be super important but for the sake of correctness:
Some of extreme values of latitude/longitude cannot be handled correctly by timezonefinder. For instance, following examples would result in an error instead of returning 'None'.
python -m timezonefinder.timezonefinder 180.0 0.0
python -m timezonefinder.timezonefinder 179.0 -90.0
Error:
Traceback (most recent call last):
File "/usr/lib/python3.6/runpy.py", line 193, in _run_module_as_main
"__main__", mod_spec)
File "/usr/lib/python3.6/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/home/dudnik/Projects/timezonefinder/timezonefinder/timezonefinder.py", line 587, in <module>
tz = tf.timezone_at(lng=longitude, lat=latitude)
File "/home/dudnik/Projects/timezonefinder/timezonefinder/functional.py", line 27, in wrapper
return func(*args, **kwargs)
File "/home/dudnik/Projects/timezonefinder/timezonefinder/timezonefinder.py", line 487, in timezone_at
return timezone_names[unpack(b'<H', self.shortcuts_unique_id.read(2))[0]]
struct.error: unpack requires a buffer of 2 bytes
Basically some values of longitude and latitude would cause 'shortcut_id_x' and 'shortcut_id_y' result in a position which is out of bounds of 'shortcuts_unique_id.bin' file.
Test if loading polygons (or caching them) into Memory gives a speed advantage.
Mode should then be configurable by the user.
Might be interesting for people who only care about performance.
I have not seen a package for timezonefinder on conda. More information about it can be found here https://www.continuum.io/downloads. I am in the process of making one on conda-forge and was wondering if @MrMinimal64 would like to be a maintainer of this as well?
Thanks!
Test results for tzwhere(shapely=True)
TIMES for 10000 realistic queries:
tzwhere: 0:00:00.772074
timezonefinder: 0:00:00.867903
0.89 times faster
TIMES for 10000 random queries:
tzwhere: 0:00:00.563584
timezonefinder: 0:00:00.570806
0.99 times faster
Do you have any idea how much faster you are than pytzwhere when using shapely with pytzwhere?
I like your usage of numba though. That makes a lot of sense :)
Using the package kwargs_only prevents builds on conda-forge as it is not currently a package contained in conda. Are there alternative workaround options (e.g. Do future releases of timezonefinder need to support Python 2 or can they be Python 3 only)?
As mentioned in the README.rst, library six is now required but, not set up as a required dependency. It is not resolved automatically when installing timezonefinder.
Suggested:
Just as a heads-up:
As many other projects (cf. https://python3statement.org/) I decided to drop the support for python2 with the next timezonefinder release. This is justified by the past compatibility issues and the increased effort of maintaining the compatibility (more tests, dependencies, code...).
The new release hence will be "4.0.0".
copy and modify the tests in helpers_test.py to test the precompiled helper functions
(all input has to converted into numpy arrays)
Unless I missed something, the only way to use custom map files currently, is to use the package as a standalone program.
It would be great if it was possible to provide a path to custom files, when using it as a package. I have modified the code to do that here and it seems to work fine (note that you might need to change the way the timezone_names is loaded, since it currently gets set once at the global level).
Let me know if you'd prefer a pull request, and I can probably provide one:
def __init__(self, bin_file_location, in_memory=False):
self.in_memory = in_memory
if self.in_memory:
self.fromfile = tzf.fromfile_memory
else:
self.fromfile = tzf.fromfile
bin_file_names = ('poly_zone_ids', 'poly_coord_amount', 'poly_adr2data',
'poly_data', 'poly_max_values', 'poly_nr2zone_id', 'hole_poly_ids',
'hole_coord_amount', 'hole_adr2data', 'hole_data', 'shortcuts_entry_amount',
'shortcuts_adr2data', 'shortcuts_data', 'shortcuts_unique_id')
for bf_name in bin_file_names:
if bin_file_location is None:
bf = self.open_binary('timezonefinder', f'{bf_name}.bin')
else:
bf = open(os.path.join(bin_file_location, f'{bf_name}.bin'), mode='rb')
setattr(self, bf_name, bf)
[...]
# Overwrite JSON file being loaded in the parent class:
if bin_file_location is not None:
with open(os.path.join(bin_file_location, TIMEZONE_NAMES_FILE), 'r') as f:
timezone_names = json.loads(f.read())
I am currently working on a simple GUI for this package and while visualizing the actual timezone polygons used internally in a map, I realized that for the most use cases, I can think of, the data is unnecessarily complex (e.g. small detailed islands on coastlines...)
Unfortunately the GUI is not finished yet, so I cannot really show the exact level of detail, but hopefully my point is understandable. (the JSON of a polygon of one zone is usually more than 100KB big)
So my question is:
Are there any objections to (just?!) including simplified data (=geometries) starting with the next release? Simplification just means certain points would additionally be classified as "inside a zone", without reassigning points which originally belong to a different zone (<- less complex polygons).
In general I am interested in what you use this package for.
Your answers would simply influence where and how I would upload the results. Thanks for your feedback.
instead of using the vanilla precompiled boundaries in the timezone-boundary-builder
releases, which are fairly huge and include territorial waters, it could make sense to use the timezone-boundary-builder
tool itself to extract different boundaries.
desired properties:
fewer vertices (less detail in the timezone polygons -> smaller data, faster algorithms)
exclude territorial waters ( polygons follow the shorelines -> looking for closest timezone with closest_timezone_at() possible again)
I have not looked into it yet, so i cannot predict how easy it is to accomplish this.
I also do not know how relevant this is to you users.
Hi everyone, I'm using this library to get timezone from user addresses, I'm getting this error for the following coordinates:
tf.timezone_at(lng=-87.6330737, lat=41.9112897)
Can you help me please telling what it means in this case?, this seems a numpy issue, but not sure what's wrong in this case.
Thanks in advance.
Hello,
Depending on Numba version, this test fails. Works with Numba 0.42 but crashes with 0.34:
self = <numba.runtime.context.NRTContext object at 0x7f00c42f2cc0>
builder = <llvmlite.llvmpy.core.Builder object at 0x7f00c42f9f28>
root_type = list(list(float64))
typ = memory-managed *payload(list(list(float64)))
value = <ir.ExtractValue 'extracted.meminfo.61' of type 'i8*', opname 'extractvalue', operands [<ir.LoadInstr '.979' of type '{i8*, i8*}', opname 'load', operands [<ir.AllocaInstr '$0.17' of type '{i8*, i8*}*', opname 'alloca', operands ()>]>]>
funcname = 'NRT_decref'
getters = (functools.partial(<function StructModel.traverse.<locals>.getter at 0x7f00c42fdea0>, 'meminfo'),)
def _call_incref_decref(self, builder, root_type, typ, value,
funcname, getters=()):
self._require_nrt()
from numba.runtime.nrtdynmod import incref_decref_ty
data_model = self._context.data_model_manager[typ]
members = data_model.traverse(builder)
for mtyp, getter in members:
self._call_incref_decref(builder, root_type, mtyp, value,
funcname, getters + (getter,))
if data_model.has_nrt_meminfo():
# Call the chain of getters to compute the member value
for getter in getters:
value = getter(value)
try:
meminfo = data_model.get_nrt_meminfo(builder, value)
except NotImplementedError as e:
> raise NotImplementedError("%s: %s" % (root_type, str(e)))
E numba.errors.LoweringError: Failed at nopython (nopython mode backend)
E list(list(float64)): unsupported nested memory-managed object
E File "timezonefinder/helpers_numba.py", line 330
E [1] During: lowering "$0.17 = build_list(items=[Var($0.8, /home/acecile/packaging/python-timezonefinder/timezonefinder-4.0.2/.pybuild/pythonX.Y_3.5/build/timezonefinder/helpers_numba.py (330)), Var($0.16, /home/acecile/packaging/python-timezonefinder/timezonefinder-4.0.2/.pybuild/pythonX.Y_3.5/build/timezonefinder/helpers_numba.py (330))])" at /home/acecile/packaging/python-timezonefinder/timezonefinder-4.0.2/.pybuild/pythonX.Y_3.5/build/timezonefinder/helpers_numba.py (330)
/usr/lib/python3/dist-packages/numba/runtime/context.py:188: LoweringError
You should probably detects this and disable Numba for this part of code if the version is too old.
Best regards, Adam.
First, thanks for timezonefinder!
I'm using the PyPI v3.2.0 of timezonefinder which has a setup requirement of numpy. The pinned requirements.txt for version v3.2.0 is numpy==1.15.3. On 2019/01/13 v1.16.0 of numpy was released. Since the version was not pinned in setup.py to 1.15.3, my project was automatically updated to numpy 1.16.0, which broke the application.
Please consider pinning install_requires package versions so timezonefinder doesn't inadvertently automatically upgrade dependencies to the latest version.
If timezonefinder does not support Python 2.7, that's perfectly fine and this issue should be refiled against the conda-forge feedstock so that it only builds python 3.x packages.
pparente@vm4-002 ~ ❯❯❯ conda create -n tzfinder 'python=2.7*' 'timezonefinder>=3.4.1'
Solving environment: done
## Package Plan ##
environment location: /home/pparente/miniconda/envs/tzfinder
added / updated specs:
- python=2.7
- timezonefinder[version='>=3.4.1']
The following packages will be downloaded:
package | build
---------------------------|-----------------
numpy-1.16.0 |py27_blas_openblashb06ca3d_0 9.1 MB conda-forge
openblas-0.3.3 | ha44fe06_1 13.7 MB conda-forge
blas-1.1 | openblas 1 KB conda-forge
libgfortran-3.0.0 | 1 281 KB conda-forge
------------------------------------------------------------
Total: 23.0 MB
The following NEW packages will be INSTALLED:
blas: 1.1-openblas conda-forge
ca-certificates: 2018.11.29-ha4d7672_0 conda-forge
certifi: 2018.11.29-py27_1000 conda-forge
enum34: 1.1.6-py27_1001 conda-forge
funcsigs: 1.0.2-py_3 conda-forge
importlib_resources: 1.0.2-py27_1000 conda-forge
libffi: 3.2.1-hfc679d8_5 conda-forge
libgcc-ng: 7.2.0-hdf63c60_3 conda-forge
libgfortran: 3.0.0-1 conda-forge
libstdcxx-ng: 7.2.0-hdf63c60_3 conda-forge
llvmlite: 0.26.0-py27hd28b015_0 conda-forge
ncurses: 6.1-hfc679d8_2 conda-forge
numba: 0.41.0-py27hf8a1672_0 conda-forge
numpy: 1.16.0-py27_blas_openblashb06ca3d_0 conda-forge [blas_openblas]
openblas: 0.3.3-ha44fe06_1 conda-forge
openssl: 1.0.2p-h470a237_2 conda-forge
pathlib2: 2.3.3-py27_1000 conda-forge
pip: 18.1-py27_1000 conda-forge
python: 2.7.15-h33da82c_6 conda-forge
readline: 7.0-haf1bffa_1 conda-forge
scandir: 1.9.0-py27h470a237_0 conda-forge
setuptools: 40.6.3-py27_0 conda-forge
singledispatch: 3.4.0.3-py27_1000 conda-forge
six: 1.12.0-py27_1000 conda-forge
sqlite: 3.26.0-hb1c47c0_0 conda-forge
timezonefinder: 3.4.1-py_0 conda-forge
tk: 8.6.9-ha92aebf_0 conda-forge
typing: 3.6.6-py27_1000 conda-forge
wheel: 0.32.3-py27_0 conda-forge
zlib: 1.2.11-h470a237_4 conda-forge
Proceed ([y]/n)? y
Downloading and Extracting Packages
numpy-1.16.0 | 9.1 MB | ################################################################################### | 100%
openblas-0.3.3 | 13.7 MB | ################################################################################### | 100%
blas-1.1 | 1 KB | ################################################################################### | 100%
libgfortran-3.0.0 | 281 KB | ################################################################################### | 100%
Preparing transaction: done
Verifying transaction: done
Executing transaction: done
#
# To activate this environment, use
#
# $ conda activate tzfinder
#
# To deactivate an active environment, use
#
# $ conda deactivate
pparente@vm4-002 ~ ❯❯❯ conda activate tzfinder
(tzfinder) pparente@vm4-002 ~ ❯❯❯ python
Python 2.7.15 | packaged by conda-forge | (default, Nov 29 2018, 06:43:57)
[GCC 4.8.2 20140120 (Red Hat 4.8.2-15)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import timezonefinder
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/timezonefinder/__init__.py", line 3, in <module>
from .timezonefinder import TimezoneFinder
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/timezonefinder/timezonefinder.py", line 22, in <module>
from .helpers_numba import coord2int, distance_to_polygon_exact, distance_to_polygon, inside_polygon, \
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/timezonefinder/helpers_numba.py", line 309, in <module>
@njit(dtype_2float_tuple(f8, f8), cache=True)
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/decorators.py", line 198, in wrapper
disp.compile(sig)
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/compiler_lock.py", line 32, in _acquire_compile_lock
return func(*args, **kwargs)
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/dispatcher.py", line 655, in compile
cres = self._compiler.compile(args, return_type)
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/dispatcher.py", line 82, in compile
pipeline_class=self.pipeline_class)
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/compiler.py", line 926, in compile_extra
return pipeline.compile_extra(func)
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/compiler.py", line 374, in compile_extra
return self._compile_bytecode()
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/compiler.py", line 857, in _compile_bytecode
return self._compile_core()
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/compiler.py", line 844, in _compile_core
res = pm.run(self.status)
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/compiler_lock.py", line 32, in _acquire_compile_lock
return func(*args, **kwargs)
File "/home/pparente/miniconda/envs/tzfinder/lib/python2.7/site-packages/numba/compiler.py", line 255, in run
raise patched_exception
numba.errors.TypingError: Failed in nopython mode pipeline (step: nopython frontend)
cannot determine Numba type of <type 'unicode'>
File "miniconda/envs/tzfinder/lib/python2.7/site-packages/timezonefinder/helpers_numba.py", line 312:
def rectify_coordinates(lng, lat):
<source elided>
if lng > 180.0 or lng < -180.0 or lat > 90.0 or lat < -90.0:
raise ValueError('The coordinates should be given in degrees. They are out ouf bounds.')
^
I am using this library within my python app hosted in a docker image, and when I started using it it was okay, and now I'm having this issue. My guess is that somewhere in the code the object is not being garbage collected/destroyed and the binaries remain open over time causing the issue. Any inputs?
Unable to find source-code formatter for language: text. Available languages are: actionscript, ada, applescript, bash, c, c#, c++, cpp, css, erlang, go, groovy, haskell, html, java, javascript, js, json, lua, none, nyan, objc, perl, php, python, r, rainbow, ruby, scala, sh, sql, swift, visualbasic, xml, yaml
OSError: [Errno 24] Too many open files: '/usr/local/lib/python3.6/site-packages/timezonefinder/poly_zone_ids.bin'
172.17.0.21 - - [08/May/2018 07:10:23] "POST /1.0/predict HTTP/1.1" 500 -
Traceback (most recent call last):
File "/usr/local/lib/python3.6/site-packages/flask/app.py", line 1997, in callFile "/usr/local/lib/python3.6/site-packages/flask/app.py", line 1985, in wsgi_app
File "/usr/local/lib/python3.6/site-packages/flask/app.py", line 1540, in handle_exception
File "/usr/local/lib/python3.6/site-packages/flask/_compat.py", line 33, in reraise
File "/usr/local/lib/python3.6/site-packages/flask/app.py", line 1982, in wsgi_app
File "/usr/local/lib/python3.6/site-packages/flask/app.py", line 1614, in full_dispatch_request
File "/usr/local/lib/python3.6/site-packages/flask/app.py", line 1517, in handle_user_exception
File "/usr/local/lib/python3.6/site-packages/flask/_compat.py", line 33, in reraise
File "/usr/local/lib/python3.6/site-packages/flask/app.py", line 1612, in full_dispatch_request
File "/usr/local/lib/python3.6/site-packages/flask/app.py", line 1598, in dispatch_request
File "/usr/src/app/app.py", line 141, in is_at_venue
File "/usr/src/app/utils/AtVenueModel.py", line 38, in init
File "/usr/src/app/utils/TemporalTools.py", line 12, in getLocalTime
File "/usr/local/lib/python3.6/site-packages/timezonefinder/timezonefinder.py", line 104, in init
modify the tox setting for also running the tests with Numba enabled.
Things might go wrong there.
Your package is very helpful, thanks a lot 😄
Just thinking, maybe you can vectorize all the functions so they can be use on a numpy array or pandas DataFrame, it'll be awesome for those who work with geo data
I know we can use .apply but why not vectorize 😉 ?
Hi.
One feature that I am missing is the possibility to compute the timezones from a pandas dataframe containing Latitude and Longitude.
If you think this might be a feature that suites into this package, I can try to implement it.
I can't see how the .bin file is created, is there a script you have that isn't checked in to Github? It would be nice to make it reproducible/updatable in the future.
Writing script to convert original timezone data from
http://efele.net/maps/tz/world/
directly into the needed binary file.
This is important to keep the data up to date (countries may change their timezones?!).
Also make the timezone_names being loaded from a separate file (names and numeration are changing)
See this post: http://gis.stackexchange.com/questions/7339/converting-shapefiles-to-text-ascii-files
import shapefile
sf = shapefile.Reader("shapefiles/blockgroups")
shapes = sf.shapes()
points = shapes[0].points
http://www.gdal.org/drv_csv.html
For now only the conversion of the tzwhere data (as .csv) is supported with the file_converter.py
May not be that important.
The problem is in the REAME.rst under Installation, there is a broken link (missing space between "feedstock" and the <). This is causing readme_renderer to choke, which then causes long_description to fail to render on PyPI.
You can check the README yourself with:
pip install readme_renderer
python setup.py check -r -sHi MrMinimal64,
Thank you for your library, it seems perfect for my project ! Indeed I have to convert many local times to UTC times from latitude and longitude features.
That's why I installed your project.
However I can't load your code. When I write:
from timezonefinder import TimezoneFinder
I have the following output:
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-8-ffb6df41970c> in <module>()
----> 1 from timezonefinder import TimezoneFinder
/home/ecoast/anaconda2/lib/python2.7/site-packages/timezonefinder/__init__.py in <module>()
1 from __future__ import absolute_import, division, print_function, unicode_literals
2
----> 3 from .timezonefinder import TimezoneFinder
4
5 __version__ = '1.5.4'
/home/ecoast/anaconda2/lib/python2.7/site-packages/timezonefinder/timezonefinder.py in <module>()
15
16 if numba is not None:
---> 17 from .helpers_numba import coord2int, distance_to_polygon, inside_polygon
18 else:
19 from .helpers import coord2int, distance_to_polygon, inside_polygon
/home/ecoast/anaconda2/lib/python2.7/site-packages/timezonefinder/helpers_numba.py in <module>()
6
7
----> 8 @jit('uint(i4,i4,i4,i4,i4,i4)', nopython=True, cache=True)
9 def position_to_line(x, y, x1, x2, y1, y2):
10 """tests if a point pX(x,y) is Left|On|Right of an infinite line from p1 to p2
could
/home/ecoast/anaconda2/lib/python2.7/site-packages/numba/decorators.py in jit(signature_or_function, locals, target, cache, **options)
146
147 wrapper = _jit(sigs, locals=locals, target=target, cache=cache,
--> 148 targetoptions=options)
149 if pyfunc is not None:
150 return wrapper(pyfunc)
/home/ecoast/anaconda2/lib/python2.7/site-packages/numba/decorators.py in wrapper(func)
162 if config.DISABLE_JIT and not target == 'npyufunc':
163 return _DisableJitWrapper(func)
--> 164 disp = dispatcher(py_func=func, locals=locals,
165 targetoptions=targetoptions,
166 **dispatcher_args)
/home/ecoast/anaconda2/lib/python2.7/site-packages/numba/dispatcher.py in __init__(self, py_func, locals, targetoptions, impl_kind)
380 locals: dict, optional
381 Mapping of local variable names to Numba types. Used to override
--> 382 the types deduced by the type inference engine.
383 targetoptions: dict, optional
384 Target-specific config options.
/home/ecoast/anaconda2/lib/python2.7/site-packages/funcsigs/__init__.pyc in signature(obj)
57
58 if not callable(obj):
---> 59 raise TypeError('{0!r} is not a callable object'.format(obj))
60
61 if isinstance(obj, types.MethodType):
TypeError: u'uint(i4,i4,i4,i4,i4,i4)' is not a callable object
And yet I checked all requirements and if my libraries were updated.
If you can help me I will be very grateful to you !
Is there a way to get the Time Zone GMT offset (-1, +14, 0, ...), instead of the Time Zone name?
If instead of using the Python interpreter, timezonefinder is used embedded in CPython, an ImportError exception will occur, rendering it unusable:
File ".../lib/python2.7/site-packages/timezonefinder/__init__.py", line 3, in <module> from .timezonefinder import TimezoneFinder File ".../lib/python2.7/site-packages/timezonefinder/timezonefinder.py", line 7, in <module> from sys import argv, exit ImportError: cannot import name argv
This occurs because sys.path is not available to the embedded interpreter.
A simple fix to this would be to make the import a local one inside the block:
if __name__ == '__main__':
I'm using timezonefinder with Python 2.7 because it said Python 2 compatible in an older version :)
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.