Can't pickle overpy.Result

[NValerij]

Issue type

• Bug Report

OverPy version

0.4

OS

• Windows 7 x64 HP

Python version

• Python 3.5 x64

Summary

Neither pickle nor _pickle Python modules can't dump query result.

Steps to reproduce

import overpy
import pickle

op = overpy.Overpass()
sample = op.query("""
    way(50.746,7.154,50.748,7.157) ["highway"];
    (._;>;);
    out body;
    """)

with open('sample.pcl', 'wb') as pcl:
    pickle.dump(sample, pcl)

Expected results

pcl-file full of usefull data =)

Actual results

Empty plc-file and exception:

     11 with open('sample.pcl', 'wb') as pcl:
---> 12     pickle.dump(sample, pcl)

AttributeError: Can't pickle local object 'Element.__init__.<locals>.<lambda>'

[kxv]

This is true for Python 2.7, too.

[phibos]

Looks like the reason for the issues are the lambda functions used as attribute_modifiers

[phibos]

Thanks for reporting the issue.

I have merged a fix into the master branch and it will be in the next release.

[wottpal]

(kindly asking 4 years later) will there ever be a "next release"? really need that fix :) thanks!

Update: For all others coming here, I've found a way using @phibos fix without a new version:

python -m pip uninstall overpy  
python -m pip install --upgrade git+git://github.com/DinoTools/python-overpy.git
Collecting git+git://github.com/DinoTools/python-overpy.git

[peci1]

Here's our workaround for Python 2.7:

import overpy
import pickle

op = overpy.Overpass()
sample = op.query("""
    way(50.746,7.154,50.748,7.157) ["highway"];
    (._;>;);
    out body;
    """)

def make_overpy_picklable(data):
    if hasattr(data, '_attribute_modifiers'):
        data._attribute_modifiers = None
    return data


def make_overpy_result_picklable(sample):
    sample = make_overpy_picklable(sample)
    for i in range(len(sample.nodes)):
        sample.nodes[i] = make_overpy_picklable(sample.nodes[i])
    for i in range(len(sample.ways)):
        sample.ways[i] = make_overpy_picklable(sample.ways[i])
    for i in range(len(sample.areas)):
        sample.areas[i] = make_overpy_picklable(sample.areas[i])
    for i in range(len(sample.relations)):
        sample.relations[i] = make_overpy_picklable(sample.relations[i])
    return sample

sample = make_overpy_result_picklable(sample)

with open('sample.pcl', 'wb') as pcl:
    pickle.dump(sample, pcl)

Deleting the dictionary is not a problem as it is only needed in the constructor.