json-streamer 
jsonstreamer provides a SAX-like push parser via the JSONStreamer class and a 'object' parser via the ObjectStreamer class which emits top level entities in any JSON object. Based on the fast c libary 'yajl'. Great for parsing streaming json over a network as it comes in or json objects that are too large to hold in memory altogether.
Dependencies
git clone [email protected]:lloyd/yajl.git
cd yajl
./configure && make install
Setup
pip3 install jsonstreamer
Also available at PyPi - https://pypi.python.org/pypi/jsonstreamer
Example
Shell
python -m jsonstreamer.jsonstreamer < some_file.json
Code
variables which contain the input we want to parse
json_object = """
{
"fruits":["apple","banana", "cherry"],
"calories":[100,200,50]
}
"""
json_array = """[1,2,true,[4,5],"a"]"""
a catch-all event listener function which prints the events
def _catch_all(event_name, *args):
print('\t{} : {}'.format(event_name, args))
JSONStreamer Example
Event listeners get events in their parameters and must have appropriate signatures for receiving their specific event of interest.
JSONStreamer provides the following events:
- doc_start
- doc_end
- object_start
- object_end
- array_start
- array_end
- key - this also carries the name of the key as a string param
- value - this also carries the value as a string|int|float|boolean|None param
- element - this also carries the value as a string|int|float|boolean|None param
Listener methods must have signatures that match
For example for events: doc_start, doc_end, object_start, object_end, array_start and array_end the listener must be as such, note no params required
def listener():
pass
OR, if your listener is a class method, it can have an additional 'self' param as such
def listener(self):
pass
For events: key, value, element listeners must also receive an additional payload and must be declared as such
def key_listener(key_string):
pass
import and run jsonstreamer on 'json_object'
from jsonstreamer import JSONStreamer
print("\nParsing the json object:")
streamer = JSONStreamer()
streamer.add_catch_all_listener(_catch_all)
streamer.consume(json_object[0:10]) #note that partial input is possible
streamer.consume(json_object[10:])
streamer.close()
output
Parsing the json object:
doc_start : ()
object_start : ()
key : ('fruits',)
array_start : ()
element : ('apple',)
element : ('banana',)
element : ('cherry',)
array_end : ()
key : ('calories',)
array_start : ()
element : (100,)
element : (200,)
element : (50,)
array_end : ()
object_end : ()
doc_end : ()
run jsonstreamer on 'json_array'
print("\nParsing the json array:")
streamer = JSONStreamer() #can't reuse old object, make a fresh one
streamer.add_catch_all_listener(_catch_all)
streamer.consume(json_array[0:5])
streamer.consume(json_array[5:])
streamer.close()
output
Parsing the json array:
doc_start : ()
array_start : ()
element : (1,)
element : (2,)
element : (True,)
array_start : ()
element : (4,)
element : (5,)
array_end : ()
element : ('a',)
array_end : ()
doc_end : ()
ObjectStreamer Example
ObjectStreamer provides the following events:
- object_stream_start
- object_stream_end
- array_stream_start
- array_stream_end
- pair
- element
import and run ObjectStreamer on 'json_object'
from jsonstreamer import ObjectStreamer
print("\nParsing the json object:")
object_streamer = ObjectStreamer()
object_streamer.add_catch_all_listener(_catch_all)
object_streamer.consume(json_object[0:9])
object_streamer.consume(json_object[9:])
object_streamer.close()
output
Parsing the json object:
object_stream_start : ()
pair : (('fruits', ['apple', 'banana', 'cherry']),)
pair : (('calories', [100, 200, 50]),)
object_stream_end : ()
run the ObjectStreamer on the 'json_array'
print("\nParsing the json array:")
object_streamer = ObjectStreamer()
object_streamer.add_catch_all_listener(_catch_all)
object_streamer.consume(json_array[0:4])
object_streamer.consume(json_array[4:])
object_streamer.close()
output - note that the events are different for an array
Parsing the json array:
array_stream_start : ()
element : (1,)
element : (2,)
element : (True,)
element : ([4, 5],)
element : ('a',)
array_stream_end : ()
Example on attaching listeners for various events
ob_streamer = ObjectStreamer()
def pair_listener(pair):
print('Explicit listener: Key: {} - Value: {}'.format(pair[0],pair[1]))
ob_streamer.add_listener('pair', pair_listener) #same for JSONStreamer
ob_streamer.consume(json_object)
ob_streamer.remove_listener(pair_listener) #if you need to remove the listener explicitly
Even easier way of attaching listeners
class MyClass:
def __init__(self):
self._obj_streamer = ObjectStreamer() #same for JSONStreamer
# this automatically finds listeners in this class and attaches them if they are named
# using the following convention '_on_eventname'. Note method names in this class
self._obj_streamer.auto_listen(self)
def _on_object_stream_start(self):
print ('Root Object Started')
def _on_pair(self, pair):
print('Key: {} - Value: {}'.format(pair[0],pair[1]))
def parse(self, data):
self._obj_streamer.consume(data)
m = MyClass()
m.parse(json_object)
Troubleshooting
- If you get an
OSError('Yajl cannot be found.')Please ensure that libyajl is available in the relevant directory. For example, on mac(osx) /usr/local/lib should have a "libyajl.dylib" Linux -> libyajl.so Windows -> yajl.dll
5 Jul 12, 2019
235 Dec 29, 2022
6 Dec 07, 2021
2 Feb 02, 2022
1 Jan 07, 2022
56 Dec 13, 2022
2 Oct 19, 2022
4 Jul 20, 2022
51 Dec 01, 2021
4 Sep 19, 2022
12 Mar 09, 2022
2 Feb 10, 2022
3 Aug 08, 2022
150 Nov 23, 2022
14 Dec 22, 2022
44 Dec 29, 2022
4 Mar 22, 2022
1 Dec 14, 2022
4 Jan 01, 2022
16.3k Dec 29, 2022