Show More
@@ -1,198 +1,185 b'' | |||||
1 | """serialization utilities for apply messages |
|
1 | """serialization utilities for apply messages""" | |
2 |
|
2 | |||
3 | Authors: |
|
3 | # Copyright (c) IPython Development Team. | |
4 |
|
4 | # Distributed under the terms of the Modified BSD License. | ||
5 | * Min RK |
|
|||
6 | """ |
|
|||
7 | #----------------------------------------------------------------------------- |
|
|||
8 | # Copyright (C) 2010-2011 The IPython Development Team |
|
|||
9 | # |
|
|||
10 | # Distributed under the terms of the BSD License. The full license is in |
|
|||
11 | # the file COPYING, distributed as part of this software. |
|
|||
12 | #----------------------------------------------------------------------------- |
|
|||
13 |
|
||||
14 | #----------------------------------------------------------------------------- |
|
|||
15 | # Imports |
|
|||
16 | #----------------------------------------------------------------------------- |
|
|||
17 |
|
5 | |||
18 | try: |
|
6 | try: | |
19 | import cPickle |
|
7 | import cPickle | |
20 | pickle = cPickle |
|
8 | pickle = cPickle | |
21 | except: |
|
9 | except: | |
22 | cPickle = None |
|
10 | cPickle = None | |
23 | import pickle |
|
11 | import pickle | |
24 |
|
12 | |||
25 |
|
||||
26 | # IPython imports |
|
13 | # IPython imports | |
27 | from IPython.utils import py3compat |
|
14 | from IPython.utils import py3compat | |
28 | from IPython.utils.data import flatten |
|
15 | from IPython.utils.data import flatten | |
29 | from IPython.utils.pickleutil import ( |
|
16 | from IPython.utils.pickleutil import ( | |
30 | can, uncan, can_sequence, uncan_sequence, CannedObject, |
|
17 | can, uncan, can_sequence, uncan_sequence, CannedObject, | |
31 | istype, sequence_types, |
|
18 | istype, sequence_types, PICKLE_PROTOCOL, | |
32 | ) |
|
19 | ) | |
33 |
|
20 | |||
34 | if py3compat.PY3: |
|
21 | if py3compat.PY3: | |
35 | buffer = memoryview |
|
22 | buffer = memoryview | |
36 |
|
23 | |||
37 | #----------------------------------------------------------------------------- |
|
24 | #----------------------------------------------------------------------------- | |
38 | # Serialization Functions |
|
25 | # Serialization Functions | |
39 | #----------------------------------------------------------------------------- |
|
26 | #----------------------------------------------------------------------------- | |
40 |
|
27 | |||
41 | # default values for the thresholds: |
|
28 | # default values for the thresholds: | |
42 | MAX_ITEMS = 64 |
|
29 | MAX_ITEMS = 64 | |
43 | MAX_BYTES = 1024 |
|
30 | MAX_BYTES = 1024 | |
44 |
|
31 | |||
45 | def _extract_buffers(obj, threshold=MAX_BYTES): |
|
32 | def _extract_buffers(obj, threshold=MAX_BYTES): | |
46 | """extract buffers larger than a certain threshold""" |
|
33 | """extract buffers larger than a certain threshold""" | |
47 | buffers = [] |
|
34 | buffers = [] | |
48 | if isinstance(obj, CannedObject) and obj.buffers: |
|
35 | if isinstance(obj, CannedObject) and obj.buffers: | |
49 | for i,buf in enumerate(obj.buffers): |
|
36 | for i,buf in enumerate(obj.buffers): | |
50 | if len(buf) > threshold: |
|
37 | if len(buf) > threshold: | |
51 | # buffer larger than threshold, prevent pickling |
|
38 | # buffer larger than threshold, prevent pickling | |
52 | obj.buffers[i] = None |
|
39 | obj.buffers[i] = None | |
53 | buffers.append(buf) |
|
40 | buffers.append(buf) | |
54 | elif isinstance(buf, buffer): |
|
41 | elif isinstance(buf, buffer): | |
55 | # buffer too small for separate send, coerce to bytes |
|
42 | # buffer too small for separate send, coerce to bytes | |
56 | # because pickling buffer objects just results in broken pointers |
|
43 | # because pickling buffer objects just results in broken pointers | |
57 | obj.buffers[i] = bytes(buf) |
|
44 | obj.buffers[i] = bytes(buf) | |
58 | return buffers |
|
45 | return buffers | |
59 |
|
46 | |||
60 | def _restore_buffers(obj, buffers): |
|
47 | def _restore_buffers(obj, buffers): | |
61 | """restore buffers extracted by """ |
|
48 | """restore buffers extracted by """ | |
62 | if isinstance(obj, CannedObject) and obj.buffers: |
|
49 | if isinstance(obj, CannedObject) and obj.buffers: | |
63 | for i,buf in enumerate(obj.buffers): |
|
50 | for i,buf in enumerate(obj.buffers): | |
64 | if buf is None: |
|
51 | if buf is None: | |
65 | obj.buffers[i] = buffers.pop(0) |
|
52 | obj.buffers[i] = buffers.pop(0) | |
66 |
|
53 | |||
67 | def serialize_object(obj, buffer_threshold=MAX_BYTES, item_threshold=MAX_ITEMS): |
|
54 | def serialize_object(obj, buffer_threshold=MAX_BYTES, item_threshold=MAX_ITEMS): | |
68 | """Serialize an object into a list of sendable buffers. |
|
55 | """Serialize an object into a list of sendable buffers. | |
69 |
|
56 | |||
70 | Parameters |
|
57 | Parameters | |
71 | ---------- |
|
58 | ---------- | |
72 |
|
59 | |||
73 | obj : object |
|
60 | obj : object | |
74 | The object to be serialized |
|
61 | The object to be serialized | |
75 | buffer_threshold : int |
|
62 | buffer_threshold : int | |
76 | The threshold (in bytes) for pulling out data buffers |
|
63 | The threshold (in bytes) for pulling out data buffers | |
77 | to avoid pickling them. |
|
64 | to avoid pickling them. | |
78 | item_threshold : int |
|
65 | item_threshold : int | |
79 | The maximum number of items over which canning will iterate. |
|
66 | The maximum number of items over which canning will iterate. | |
80 | Containers (lists, dicts) larger than this will be pickled without |
|
67 | Containers (lists, dicts) larger than this will be pickled without | |
81 | introspection. |
|
68 | introspection. | |
82 |
|
69 | |||
83 | Returns |
|
70 | Returns | |
84 | ------- |
|
71 | ------- | |
85 | [bufs] : list of buffers representing the serialized object. |
|
72 | [bufs] : list of buffers representing the serialized object. | |
86 | """ |
|
73 | """ | |
87 | buffers = [] |
|
74 | buffers = [] | |
88 | if istype(obj, sequence_types) and len(obj) < item_threshold: |
|
75 | if istype(obj, sequence_types) and len(obj) < item_threshold: | |
89 | cobj = can_sequence(obj) |
|
76 | cobj = can_sequence(obj) | |
90 | for c in cobj: |
|
77 | for c in cobj: | |
91 | buffers.extend(_extract_buffers(c, buffer_threshold)) |
|
78 | buffers.extend(_extract_buffers(c, buffer_threshold)) | |
92 | elif istype(obj, dict) and len(obj) < item_threshold: |
|
79 | elif istype(obj, dict) and len(obj) < item_threshold: | |
93 | cobj = {} |
|
80 | cobj = {} | |
94 | for k in sorted(obj): |
|
81 | for k in sorted(obj): | |
95 | c = can(obj[k]) |
|
82 | c = can(obj[k]) | |
96 | buffers.extend(_extract_buffers(c, buffer_threshold)) |
|
83 | buffers.extend(_extract_buffers(c, buffer_threshold)) | |
97 | cobj[k] = c |
|
84 | cobj[k] = c | |
98 | else: |
|
85 | else: | |
99 | cobj = can(obj) |
|
86 | cobj = can(obj) | |
100 | buffers.extend(_extract_buffers(cobj, buffer_threshold)) |
|
87 | buffers.extend(_extract_buffers(cobj, buffer_threshold)) | |
101 |
|
88 | |||
102 |
buffers.insert(0, pickle.dumps(cobj, |
|
89 | buffers.insert(0, pickle.dumps(cobj, PICKLE_PROTOCOL)) | |
103 | return buffers |
|
90 | return buffers | |
104 |
|
91 | |||
105 | def unserialize_object(buffers, g=None): |
|
92 | def unserialize_object(buffers, g=None): | |
106 | """reconstruct an object serialized by serialize_object from data buffers. |
|
93 | """reconstruct an object serialized by serialize_object from data buffers. | |
107 |
|
94 | |||
108 | Parameters |
|
95 | Parameters | |
109 | ---------- |
|
96 | ---------- | |
110 |
|
97 | |||
111 | bufs : list of buffers/bytes |
|
98 | bufs : list of buffers/bytes | |
112 |
|
99 | |||
113 | g : globals to be used when uncanning |
|
100 | g : globals to be used when uncanning | |
114 |
|
101 | |||
115 | Returns |
|
102 | Returns | |
116 | ------- |
|
103 | ------- | |
117 |
|
104 | |||
118 | (newobj, bufs) : unpacked object, and the list of remaining unused buffers. |
|
105 | (newobj, bufs) : unpacked object, and the list of remaining unused buffers. | |
119 | """ |
|
106 | """ | |
120 | bufs = list(buffers) |
|
107 | bufs = list(buffers) | |
121 | pobj = bufs.pop(0) |
|
108 | pobj = bufs.pop(0) | |
122 | if not isinstance(pobj, bytes): |
|
109 | if not isinstance(pobj, bytes): | |
123 | # a zmq message |
|
110 | # a zmq message | |
124 | pobj = bytes(pobj) |
|
111 | pobj = bytes(pobj) | |
125 | canned = pickle.loads(pobj) |
|
112 | canned = pickle.loads(pobj) | |
126 | if istype(canned, sequence_types) and len(canned) < MAX_ITEMS: |
|
113 | if istype(canned, sequence_types) and len(canned) < MAX_ITEMS: | |
127 | for c in canned: |
|
114 | for c in canned: | |
128 | _restore_buffers(c, bufs) |
|
115 | _restore_buffers(c, bufs) | |
129 | newobj = uncan_sequence(canned, g) |
|
116 | newobj = uncan_sequence(canned, g) | |
130 | elif istype(canned, dict) and len(canned) < MAX_ITEMS: |
|
117 | elif istype(canned, dict) and len(canned) < MAX_ITEMS: | |
131 | newobj = {} |
|
118 | newobj = {} | |
132 | for k in sorted(canned): |
|
119 | for k in sorted(canned): | |
133 | c = canned[k] |
|
120 | c = canned[k] | |
134 | _restore_buffers(c, bufs) |
|
121 | _restore_buffers(c, bufs) | |
135 | newobj[k] = uncan(c, g) |
|
122 | newobj[k] = uncan(c, g) | |
136 | else: |
|
123 | else: | |
137 | _restore_buffers(canned, bufs) |
|
124 | _restore_buffers(canned, bufs) | |
138 | newobj = uncan(canned, g) |
|
125 | newobj = uncan(canned, g) | |
139 |
|
126 | |||
140 | return newobj, bufs |
|
127 | return newobj, bufs | |
141 |
|
128 | |||
142 | def pack_apply_message(f, args, kwargs, buffer_threshold=MAX_BYTES, item_threshold=MAX_ITEMS): |
|
129 | def pack_apply_message(f, args, kwargs, buffer_threshold=MAX_BYTES, item_threshold=MAX_ITEMS): | |
143 | """pack up a function, args, and kwargs to be sent over the wire |
|
130 | """pack up a function, args, and kwargs to be sent over the wire | |
144 |
|
131 | |||
145 | Each element of args/kwargs will be canned for special treatment, |
|
132 | Each element of args/kwargs will be canned for special treatment, | |
146 | but inspection will not go any deeper than that. |
|
133 | but inspection will not go any deeper than that. | |
147 |
|
134 | |||
148 | Any object whose data is larger than `threshold` will not have their data copied |
|
135 | Any object whose data is larger than `threshold` will not have their data copied | |
149 | (only numpy arrays and bytes/buffers support zero-copy) |
|
136 | (only numpy arrays and bytes/buffers support zero-copy) | |
150 |
|
137 | |||
151 | Message will be a list of bytes/buffers of the format: |
|
138 | Message will be a list of bytes/buffers of the format: | |
152 |
|
139 | |||
153 | [ cf, pinfo, <arg_bufs>, <kwarg_bufs> ] |
|
140 | [ cf, pinfo, <arg_bufs>, <kwarg_bufs> ] | |
154 |
|
141 | |||
155 | With length at least two + len(args) + len(kwargs) |
|
142 | With length at least two + len(args) + len(kwargs) | |
156 | """ |
|
143 | """ | |
157 |
|
144 | |||
158 | arg_bufs = flatten(serialize_object(arg, buffer_threshold, item_threshold) for arg in args) |
|
145 | arg_bufs = flatten(serialize_object(arg, buffer_threshold, item_threshold) for arg in args) | |
159 |
|
146 | |||
160 | kw_keys = sorted(kwargs.keys()) |
|
147 | kw_keys = sorted(kwargs.keys()) | |
161 | kwarg_bufs = flatten(serialize_object(kwargs[key], buffer_threshold, item_threshold) for key in kw_keys) |
|
148 | kwarg_bufs = flatten(serialize_object(kwargs[key], buffer_threshold, item_threshold) for key in kw_keys) | |
162 |
|
149 | |||
163 | info = dict(nargs=len(args), narg_bufs=len(arg_bufs), kw_keys=kw_keys) |
|
150 | info = dict(nargs=len(args), narg_bufs=len(arg_bufs), kw_keys=kw_keys) | |
164 |
|
151 | |||
165 |
msg = [pickle.dumps(can(f), |
|
152 | msg = [pickle.dumps(can(f), PICKLE_PROTOCOL)] | |
166 |
msg.append(pickle.dumps(info, |
|
153 | msg.append(pickle.dumps(info, PICKLE_PROTOCOL)) | |
167 | msg.extend(arg_bufs) |
|
154 | msg.extend(arg_bufs) | |
168 | msg.extend(kwarg_bufs) |
|
155 | msg.extend(kwarg_bufs) | |
169 |
|
156 | |||
170 | return msg |
|
157 | return msg | |
171 |
|
158 | |||
172 | def unpack_apply_message(bufs, g=None, copy=True): |
|
159 | def unpack_apply_message(bufs, g=None, copy=True): | |
173 | """unpack f,args,kwargs from buffers packed by pack_apply_message() |
|
160 | """unpack f,args,kwargs from buffers packed by pack_apply_message() | |
174 | Returns: original f,args,kwargs""" |
|
161 | Returns: original f,args,kwargs""" | |
175 | bufs = list(bufs) # allow us to pop |
|
162 | bufs = list(bufs) # allow us to pop | |
176 | assert len(bufs) >= 2, "not enough buffers!" |
|
163 | assert len(bufs) >= 2, "not enough buffers!" | |
177 | if not copy: |
|
164 | if not copy: | |
178 | for i in range(2): |
|
165 | for i in range(2): | |
179 | bufs[i] = bufs[i].bytes |
|
166 | bufs[i] = bufs[i].bytes | |
180 | f = uncan(pickle.loads(bufs.pop(0)), g) |
|
167 | f = uncan(pickle.loads(bufs.pop(0)), g) | |
181 | info = pickle.loads(bufs.pop(0)) |
|
168 | info = pickle.loads(bufs.pop(0)) | |
182 | arg_bufs, kwarg_bufs = bufs[:info['narg_bufs']], bufs[info['narg_bufs']:] |
|
169 | arg_bufs, kwarg_bufs = bufs[:info['narg_bufs']], bufs[info['narg_bufs']:] | |
183 |
|
170 | |||
184 | args = [] |
|
171 | args = [] | |
185 | for i in range(info['nargs']): |
|
172 | for i in range(info['nargs']): | |
186 | arg, arg_bufs = unserialize_object(arg_bufs, g) |
|
173 | arg, arg_bufs = unserialize_object(arg_bufs, g) | |
187 | args.append(arg) |
|
174 | args.append(arg) | |
188 | args = tuple(args) |
|
175 | args = tuple(args) | |
189 | assert not arg_bufs, "Shouldn't be any arg bufs left over" |
|
176 | assert not arg_bufs, "Shouldn't be any arg bufs left over" | |
190 |
|
177 | |||
191 | kwargs = {} |
|
178 | kwargs = {} | |
192 | for key in info['kw_keys']: |
|
179 | for key in info['kw_keys']: | |
193 | kwarg, kwarg_bufs = unserialize_object(kwarg_bufs, g) |
|
180 | kwarg, kwarg_bufs = unserialize_object(kwarg_bufs, g) | |
194 | kwargs[key] = kwarg |
|
181 | kwargs[key] = kwarg | |
195 | assert not kwarg_bufs, "Shouldn't be any kwarg bufs left over" |
|
182 | assert not kwarg_bufs, "Shouldn't be any kwarg bufs left over" | |
196 |
|
183 | |||
197 | return f,args,kwargs |
|
184 | return f,args,kwargs | |
198 |
|
185 |
@@ -1,856 +1,857 b'' | |||||
1 | """Session object for building, serializing, sending, and receiving messages in |
|
1 | """Session object for building, serializing, sending, and receiving messages in | |
2 | IPython. The Session object supports serialization, HMAC signatures, and |
|
2 | IPython. The Session object supports serialization, HMAC signatures, and | |
3 | metadata on messages. |
|
3 | metadata on messages. | |
4 |
|
4 | |||
5 | Also defined here are utilities for working with Sessions: |
|
5 | Also defined here are utilities for working with Sessions: | |
6 | * A SessionFactory to be used as a base class for configurables that work with |
|
6 | * A SessionFactory to be used as a base class for configurables that work with | |
7 | Sessions. |
|
7 | Sessions. | |
8 | * A Message object for convenience that allows attribute-access to the msg dict. |
|
8 | * A Message object for convenience that allows attribute-access to the msg dict. | |
9 | """ |
|
9 | """ | |
10 |
|
10 | |||
11 | # Copyright (c) IPython Development Team. |
|
11 | # Copyright (c) IPython Development Team. | |
12 | # Distributed under the terms of the Modified BSD License. |
|
12 | # Distributed under the terms of the Modified BSD License. | |
13 |
|
13 | |||
14 | import hashlib |
|
14 | import hashlib | |
15 | import hmac |
|
15 | import hmac | |
16 | import logging |
|
16 | import logging | |
17 | import os |
|
17 | import os | |
18 | import pprint |
|
18 | import pprint | |
19 | import random |
|
19 | import random | |
20 | import uuid |
|
20 | import uuid | |
21 | from datetime import datetime |
|
21 | from datetime import datetime | |
22 |
|
22 | |||
23 | try: |
|
23 | try: | |
24 | import cPickle |
|
24 | import cPickle | |
25 | pickle = cPickle |
|
25 | pickle = cPickle | |
26 | except: |
|
26 | except: | |
27 | cPickle = None |
|
27 | cPickle = None | |
28 | import pickle |
|
28 | import pickle | |
29 |
|
29 | |||
30 | import zmq |
|
30 | import zmq | |
31 | from zmq.utils import jsonapi |
|
31 | from zmq.utils import jsonapi | |
32 | from zmq.eventloop.ioloop import IOLoop |
|
32 | from zmq.eventloop.ioloop import IOLoop | |
33 | from zmq.eventloop.zmqstream import ZMQStream |
|
33 | from zmq.eventloop.zmqstream import ZMQStream | |
34 |
|
34 | |||
35 | from IPython.core.release import kernel_protocol_version, kernel_protocol_version_info |
|
35 | from IPython.core.release import kernel_protocol_version, kernel_protocol_version_info | |
36 | from IPython.config.configurable import Configurable, LoggingConfigurable |
|
36 | from IPython.config.configurable import Configurable, LoggingConfigurable | |
37 | from IPython.utils import io |
|
37 | from IPython.utils import io | |
38 | from IPython.utils.importstring import import_item |
|
38 | from IPython.utils.importstring import import_item | |
39 | from IPython.utils.jsonutil import extract_dates, squash_dates, date_default |
|
39 | from IPython.utils.jsonutil import extract_dates, squash_dates, date_default | |
40 | from IPython.utils.py3compat import (str_to_bytes, str_to_unicode, unicode_type, |
|
40 | from IPython.utils.py3compat import (str_to_bytes, str_to_unicode, unicode_type, | |
41 | iteritems) |
|
41 | iteritems) | |
42 | from IPython.utils.traitlets import (CBytes, Unicode, Bool, Any, Instance, Set, |
|
42 | from IPython.utils.traitlets import (CBytes, Unicode, Bool, Any, Instance, Set, | |
43 | DottedObjectName, CUnicode, Dict, Integer, |
|
43 | DottedObjectName, CUnicode, Dict, Integer, | |
44 | TraitError, |
|
44 | TraitError, | |
45 | ) |
|
45 | ) | |
|
46 | from IPython.utils.pickleutil import PICKLE_PROTOCOL | |||
46 | from IPython.kernel.adapter import adapt |
|
47 | from IPython.kernel.adapter import adapt | |
47 | from IPython.kernel.zmq.serialize import MAX_ITEMS, MAX_BYTES |
|
48 | from IPython.kernel.zmq.serialize import MAX_ITEMS, MAX_BYTES | |
48 |
|
49 | |||
49 | #----------------------------------------------------------------------------- |
|
50 | #----------------------------------------------------------------------------- | |
50 | # utility functions |
|
51 | # utility functions | |
51 | #----------------------------------------------------------------------------- |
|
52 | #----------------------------------------------------------------------------- | |
52 |
|
53 | |||
53 | def squash_unicode(obj): |
|
54 | def squash_unicode(obj): | |
54 | """coerce unicode back to bytestrings.""" |
|
55 | """coerce unicode back to bytestrings.""" | |
55 | if isinstance(obj,dict): |
|
56 | if isinstance(obj,dict): | |
56 | for key in obj.keys(): |
|
57 | for key in obj.keys(): | |
57 | obj[key] = squash_unicode(obj[key]) |
|
58 | obj[key] = squash_unicode(obj[key]) | |
58 | if isinstance(key, unicode_type): |
|
59 | if isinstance(key, unicode_type): | |
59 | obj[squash_unicode(key)] = obj.pop(key) |
|
60 | obj[squash_unicode(key)] = obj.pop(key) | |
60 | elif isinstance(obj, list): |
|
61 | elif isinstance(obj, list): | |
61 | for i,v in enumerate(obj): |
|
62 | for i,v in enumerate(obj): | |
62 | obj[i] = squash_unicode(v) |
|
63 | obj[i] = squash_unicode(v) | |
63 | elif isinstance(obj, unicode_type): |
|
64 | elif isinstance(obj, unicode_type): | |
64 | obj = obj.encode('utf8') |
|
65 | obj = obj.encode('utf8') | |
65 | return obj |
|
66 | return obj | |
66 |
|
67 | |||
67 | #----------------------------------------------------------------------------- |
|
68 | #----------------------------------------------------------------------------- | |
68 | # globals and defaults |
|
69 | # globals and defaults | |
69 | #----------------------------------------------------------------------------- |
|
70 | #----------------------------------------------------------------------------- | |
70 |
|
71 | |||
71 | # ISO8601-ify datetime objects |
|
72 | # ISO8601-ify datetime objects | |
72 | # allow unicode |
|
73 | # allow unicode | |
73 | # disallow nan, because it's not actually valid JSON |
|
74 | # disallow nan, because it's not actually valid JSON | |
74 | json_packer = lambda obj: jsonapi.dumps(obj, default=date_default, |
|
75 | json_packer = lambda obj: jsonapi.dumps(obj, default=date_default, | |
75 | ensure_ascii=False, allow_nan=False, |
|
76 | ensure_ascii=False, allow_nan=False, | |
76 | ) |
|
77 | ) | |
77 | json_unpacker = lambda s: jsonapi.loads(s) |
|
78 | json_unpacker = lambda s: jsonapi.loads(s) | |
78 |
|
79 | |||
79 |
pickle_packer = lambda o: pickle.dumps(squash_dates(o), |
|
80 | pickle_packer = lambda o: pickle.dumps(squash_dates(o), PICKLE_PROTOCOL) | |
80 | pickle_unpacker = pickle.loads |
|
81 | pickle_unpacker = pickle.loads | |
81 |
|
82 | |||
82 | default_packer = json_packer |
|
83 | default_packer = json_packer | |
83 | default_unpacker = json_unpacker |
|
84 | default_unpacker = json_unpacker | |
84 |
|
85 | |||
85 | DELIM = b"<IDS|MSG>" |
|
86 | DELIM = b"<IDS|MSG>" | |
86 | # singleton dummy tracker, which will always report as done |
|
87 | # singleton dummy tracker, which will always report as done | |
87 | DONE = zmq.MessageTracker() |
|
88 | DONE = zmq.MessageTracker() | |
88 |
|
89 | |||
89 | #----------------------------------------------------------------------------- |
|
90 | #----------------------------------------------------------------------------- | |
90 | # Mixin tools for apps that use Sessions |
|
91 | # Mixin tools for apps that use Sessions | |
91 | #----------------------------------------------------------------------------- |
|
92 | #----------------------------------------------------------------------------- | |
92 |
|
93 | |||
93 | session_aliases = dict( |
|
94 | session_aliases = dict( | |
94 | ident = 'Session.session', |
|
95 | ident = 'Session.session', | |
95 | user = 'Session.username', |
|
96 | user = 'Session.username', | |
96 | keyfile = 'Session.keyfile', |
|
97 | keyfile = 'Session.keyfile', | |
97 | ) |
|
98 | ) | |
98 |
|
99 | |||
99 | session_flags = { |
|
100 | session_flags = { | |
100 | 'secure' : ({'Session' : { 'key' : str_to_bytes(str(uuid.uuid4())), |
|
101 | 'secure' : ({'Session' : { 'key' : str_to_bytes(str(uuid.uuid4())), | |
101 | 'keyfile' : '' }}, |
|
102 | 'keyfile' : '' }}, | |
102 | """Use HMAC digests for authentication of messages. |
|
103 | """Use HMAC digests for authentication of messages. | |
103 | Setting this flag will generate a new UUID to use as the HMAC key. |
|
104 | Setting this flag will generate a new UUID to use as the HMAC key. | |
104 | """), |
|
105 | """), | |
105 | 'no-secure' : ({'Session' : { 'key' : b'', 'keyfile' : '' }}, |
|
106 | 'no-secure' : ({'Session' : { 'key' : b'', 'keyfile' : '' }}, | |
106 | """Don't authenticate messages."""), |
|
107 | """Don't authenticate messages."""), | |
107 | } |
|
108 | } | |
108 |
|
109 | |||
109 | def default_secure(cfg): |
|
110 | def default_secure(cfg): | |
110 | """Set the default behavior for a config environment to be secure. |
|
111 | """Set the default behavior for a config environment to be secure. | |
111 |
|
112 | |||
112 | If Session.key/keyfile have not been set, set Session.key to |
|
113 | If Session.key/keyfile have not been set, set Session.key to | |
113 | a new random UUID. |
|
114 | a new random UUID. | |
114 | """ |
|
115 | """ | |
115 |
|
116 | |||
116 | if 'Session' in cfg: |
|
117 | if 'Session' in cfg: | |
117 | if 'key' in cfg.Session or 'keyfile' in cfg.Session: |
|
118 | if 'key' in cfg.Session or 'keyfile' in cfg.Session: | |
118 | return |
|
119 | return | |
119 | # key/keyfile not specified, generate new UUID: |
|
120 | # key/keyfile not specified, generate new UUID: | |
120 | cfg.Session.key = str_to_bytes(str(uuid.uuid4())) |
|
121 | cfg.Session.key = str_to_bytes(str(uuid.uuid4())) | |
121 |
|
122 | |||
122 |
|
123 | |||
123 | #----------------------------------------------------------------------------- |
|
124 | #----------------------------------------------------------------------------- | |
124 | # Classes |
|
125 | # Classes | |
125 | #----------------------------------------------------------------------------- |
|
126 | #----------------------------------------------------------------------------- | |
126 |
|
127 | |||
127 | class SessionFactory(LoggingConfigurable): |
|
128 | class SessionFactory(LoggingConfigurable): | |
128 | """The Base class for configurables that have a Session, Context, logger, |
|
129 | """The Base class for configurables that have a Session, Context, logger, | |
129 | and IOLoop. |
|
130 | and IOLoop. | |
130 | """ |
|
131 | """ | |
131 |
|
132 | |||
132 | logname = Unicode('') |
|
133 | logname = Unicode('') | |
133 | def _logname_changed(self, name, old, new): |
|
134 | def _logname_changed(self, name, old, new): | |
134 | self.log = logging.getLogger(new) |
|
135 | self.log = logging.getLogger(new) | |
135 |
|
136 | |||
136 | # not configurable: |
|
137 | # not configurable: | |
137 | context = Instance('zmq.Context') |
|
138 | context = Instance('zmq.Context') | |
138 | def _context_default(self): |
|
139 | def _context_default(self): | |
139 | return zmq.Context.instance() |
|
140 | return zmq.Context.instance() | |
140 |
|
141 | |||
141 | session = Instance('IPython.kernel.zmq.session.Session') |
|
142 | session = Instance('IPython.kernel.zmq.session.Session') | |
142 |
|
143 | |||
143 | loop = Instance('zmq.eventloop.ioloop.IOLoop', allow_none=False) |
|
144 | loop = Instance('zmq.eventloop.ioloop.IOLoop', allow_none=False) | |
144 | def _loop_default(self): |
|
145 | def _loop_default(self): | |
145 | return IOLoop.instance() |
|
146 | return IOLoop.instance() | |
146 |
|
147 | |||
147 | def __init__(self, **kwargs): |
|
148 | def __init__(self, **kwargs): | |
148 | super(SessionFactory, self).__init__(**kwargs) |
|
149 | super(SessionFactory, self).__init__(**kwargs) | |
149 |
|
150 | |||
150 | if self.session is None: |
|
151 | if self.session is None: | |
151 | # construct the session |
|
152 | # construct the session | |
152 | self.session = Session(**kwargs) |
|
153 | self.session = Session(**kwargs) | |
153 |
|
154 | |||
154 |
|
155 | |||
155 | class Message(object): |
|
156 | class Message(object): | |
156 | """A simple message object that maps dict keys to attributes. |
|
157 | """A simple message object that maps dict keys to attributes. | |
157 |
|
158 | |||
158 | A Message can be created from a dict and a dict from a Message instance |
|
159 | A Message can be created from a dict and a dict from a Message instance | |
159 | simply by calling dict(msg_obj).""" |
|
160 | simply by calling dict(msg_obj).""" | |
160 |
|
161 | |||
161 | def __init__(self, msg_dict): |
|
162 | def __init__(self, msg_dict): | |
162 | dct = self.__dict__ |
|
163 | dct = self.__dict__ | |
163 | for k, v in iteritems(dict(msg_dict)): |
|
164 | for k, v in iteritems(dict(msg_dict)): | |
164 | if isinstance(v, dict): |
|
165 | if isinstance(v, dict): | |
165 | v = Message(v) |
|
166 | v = Message(v) | |
166 | dct[k] = v |
|
167 | dct[k] = v | |
167 |
|
168 | |||
168 | # Having this iterator lets dict(msg_obj) work out of the box. |
|
169 | # Having this iterator lets dict(msg_obj) work out of the box. | |
169 | def __iter__(self): |
|
170 | def __iter__(self): | |
170 | return iter(iteritems(self.__dict__)) |
|
171 | return iter(iteritems(self.__dict__)) | |
171 |
|
172 | |||
172 | def __repr__(self): |
|
173 | def __repr__(self): | |
173 | return repr(self.__dict__) |
|
174 | return repr(self.__dict__) | |
174 |
|
175 | |||
175 | def __str__(self): |
|
176 | def __str__(self): | |
176 | return pprint.pformat(self.__dict__) |
|
177 | return pprint.pformat(self.__dict__) | |
177 |
|
178 | |||
178 | def __contains__(self, k): |
|
179 | def __contains__(self, k): | |
179 | return k in self.__dict__ |
|
180 | return k in self.__dict__ | |
180 |
|
181 | |||
181 | def __getitem__(self, k): |
|
182 | def __getitem__(self, k): | |
182 | return self.__dict__[k] |
|
183 | return self.__dict__[k] | |
183 |
|
184 | |||
184 |
|
185 | |||
185 | def msg_header(msg_id, msg_type, username, session): |
|
186 | def msg_header(msg_id, msg_type, username, session): | |
186 | date = datetime.now() |
|
187 | date = datetime.now() | |
187 | version = kernel_protocol_version |
|
188 | version = kernel_protocol_version | |
188 | return locals() |
|
189 | return locals() | |
189 |
|
190 | |||
190 | def extract_header(msg_or_header): |
|
191 | def extract_header(msg_or_header): | |
191 | """Given a message or header, return the header.""" |
|
192 | """Given a message or header, return the header.""" | |
192 | if not msg_or_header: |
|
193 | if not msg_or_header: | |
193 | return {} |
|
194 | return {} | |
194 | try: |
|
195 | try: | |
195 | # See if msg_or_header is the entire message. |
|
196 | # See if msg_or_header is the entire message. | |
196 | h = msg_or_header['header'] |
|
197 | h = msg_or_header['header'] | |
197 | except KeyError: |
|
198 | except KeyError: | |
198 | try: |
|
199 | try: | |
199 | # See if msg_or_header is just the header |
|
200 | # See if msg_or_header is just the header | |
200 | h = msg_or_header['msg_id'] |
|
201 | h = msg_or_header['msg_id'] | |
201 | except KeyError: |
|
202 | except KeyError: | |
202 | raise |
|
203 | raise | |
203 | else: |
|
204 | else: | |
204 | h = msg_or_header |
|
205 | h = msg_or_header | |
205 | if not isinstance(h, dict): |
|
206 | if not isinstance(h, dict): | |
206 | h = dict(h) |
|
207 | h = dict(h) | |
207 | return h |
|
208 | return h | |
208 |
|
209 | |||
209 | class Session(Configurable): |
|
210 | class Session(Configurable): | |
210 | """Object for handling serialization and sending of messages. |
|
211 | """Object for handling serialization and sending of messages. | |
211 |
|
212 | |||
212 | The Session object handles building messages and sending them |
|
213 | The Session object handles building messages and sending them | |
213 | with ZMQ sockets or ZMQStream objects. Objects can communicate with each |
|
214 | with ZMQ sockets or ZMQStream objects. Objects can communicate with each | |
214 | other over the network via Session objects, and only need to work with the |
|
215 | other over the network via Session objects, and only need to work with the | |
215 | dict-based IPython message spec. The Session will handle |
|
216 | dict-based IPython message spec. The Session will handle | |
216 | serialization/deserialization, security, and metadata. |
|
217 | serialization/deserialization, security, and metadata. | |
217 |
|
218 | |||
218 | Sessions support configurable serialiization via packer/unpacker traits, |
|
219 | Sessions support configurable serialiization via packer/unpacker traits, | |
219 | and signing with HMAC digests via the key/keyfile traits. |
|
220 | and signing with HMAC digests via the key/keyfile traits. | |
220 |
|
221 | |||
221 | Parameters |
|
222 | Parameters | |
222 | ---------- |
|
223 | ---------- | |
223 |
|
224 | |||
224 | debug : bool |
|
225 | debug : bool | |
225 | whether to trigger extra debugging statements |
|
226 | whether to trigger extra debugging statements | |
226 | packer/unpacker : str : 'json', 'pickle' or import_string |
|
227 | packer/unpacker : str : 'json', 'pickle' or import_string | |
227 | importstrings for methods to serialize message parts. If just |
|
228 | importstrings for methods to serialize message parts. If just | |
228 | 'json' or 'pickle', predefined JSON and pickle packers will be used. |
|
229 | 'json' or 'pickle', predefined JSON and pickle packers will be used. | |
229 | Otherwise, the entire importstring must be used. |
|
230 | Otherwise, the entire importstring must be used. | |
230 |
|
231 | |||
231 | The functions must accept at least valid JSON input, and output *bytes*. |
|
232 | The functions must accept at least valid JSON input, and output *bytes*. | |
232 |
|
233 | |||
233 | For example, to use msgpack: |
|
234 | For example, to use msgpack: | |
234 | packer = 'msgpack.packb', unpacker='msgpack.unpackb' |
|
235 | packer = 'msgpack.packb', unpacker='msgpack.unpackb' | |
235 | pack/unpack : callables |
|
236 | pack/unpack : callables | |
236 | You can also set the pack/unpack callables for serialization directly. |
|
237 | You can also set the pack/unpack callables for serialization directly. | |
237 | session : bytes |
|
238 | session : bytes | |
238 | the ID of this Session object. The default is to generate a new UUID. |
|
239 | the ID of this Session object. The default is to generate a new UUID. | |
239 | username : unicode |
|
240 | username : unicode | |
240 | username added to message headers. The default is to ask the OS. |
|
241 | username added to message headers. The default is to ask the OS. | |
241 | key : bytes |
|
242 | key : bytes | |
242 | The key used to initialize an HMAC signature. If unset, messages |
|
243 | The key used to initialize an HMAC signature. If unset, messages | |
243 | will not be signed or checked. |
|
244 | will not be signed or checked. | |
244 | keyfile : filepath |
|
245 | keyfile : filepath | |
245 | The file containing a key. If this is set, `key` will be initialized |
|
246 | The file containing a key. If this is set, `key` will be initialized | |
246 | to the contents of the file. |
|
247 | to the contents of the file. | |
247 |
|
248 | |||
248 | """ |
|
249 | """ | |
249 |
|
250 | |||
250 | debug=Bool(False, config=True, help="""Debug output in the Session""") |
|
251 | debug=Bool(False, config=True, help="""Debug output in the Session""") | |
251 |
|
252 | |||
252 | packer = DottedObjectName('json',config=True, |
|
253 | packer = DottedObjectName('json',config=True, | |
253 | help="""The name of the packer for serializing messages. |
|
254 | help="""The name of the packer for serializing messages. | |
254 | Should be one of 'json', 'pickle', or an import name |
|
255 | Should be one of 'json', 'pickle', or an import name | |
255 | for a custom callable serializer.""") |
|
256 | for a custom callable serializer.""") | |
256 | def _packer_changed(self, name, old, new): |
|
257 | def _packer_changed(self, name, old, new): | |
257 | if new.lower() == 'json': |
|
258 | if new.lower() == 'json': | |
258 | self.pack = json_packer |
|
259 | self.pack = json_packer | |
259 | self.unpack = json_unpacker |
|
260 | self.unpack = json_unpacker | |
260 | self.unpacker = new |
|
261 | self.unpacker = new | |
261 | elif new.lower() == 'pickle': |
|
262 | elif new.lower() == 'pickle': | |
262 | self.pack = pickle_packer |
|
263 | self.pack = pickle_packer | |
263 | self.unpack = pickle_unpacker |
|
264 | self.unpack = pickle_unpacker | |
264 | self.unpacker = new |
|
265 | self.unpacker = new | |
265 | else: |
|
266 | else: | |
266 | self.pack = import_item(str(new)) |
|
267 | self.pack = import_item(str(new)) | |
267 |
|
268 | |||
268 | unpacker = DottedObjectName('json', config=True, |
|
269 | unpacker = DottedObjectName('json', config=True, | |
269 | help="""The name of the unpacker for unserializing messages. |
|
270 | help="""The name of the unpacker for unserializing messages. | |
270 | Only used with custom functions for `packer`.""") |
|
271 | Only used with custom functions for `packer`.""") | |
271 | def _unpacker_changed(self, name, old, new): |
|
272 | def _unpacker_changed(self, name, old, new): | |
272 | if new.lower() == 'json': |
|
273 | if new.lower() == 'json': | |
273 | self.pack = json_packer |
|
274 | self.pack = json_packer | |
274 | self.unpack = json_unpacker |
|
275 | self.unpack = json_unpacker | |
275 | self.packer = new |
|
276 | self.packer = new | |
276 | elif new.lower() == 'pickle': |
|
277 | elif new.lower() == 'pickle': | |
277 | self.pack = pickle_packer |
|
278 | self.pack = pickle_packer | |
278 | self.unpack = pickle_unpacker |
|
279 | self.unpack = pickle_unpacker | |
279 | self.packer = new |
|
280 | self.packer = new | |
280 | else: |
|
281 | else: | |
281 | self.unpack = import_item(str(new)) |
|
282 | self.unpack = import_item(str(new)) | |
282 |
|
283 | |||
283 | session = CUnicode(u'', config=True, |
|
284 | session = CUnicode(u'', config=True, | |
284 | help="""The UUID identifying this session.""") |
|
285 | help="""The UUID identifying this session.""") | |
285 | def _session_default(self): |
|
286 | def _session_default(self): | |
286 | u = unicode_type(uuid.uuid4()) |
|
287 | u = unicode_type(uuid.uuid4()) | |
287 | self.bsession = u.encode('ascii') |
|
288 | self.bsession = u.encode('ascii') | |
288 | return u |
|
289 | return u | |
289 |
|
290 | |||
290 | def _session_changed(self, name, old, new): |
|
291 | def _session_changed(self, name, old, new): | |
291 | self.bsession = self.session.encode('ascii') |
|
292 | self.bsession = self.session.encode('ascii') | |
292 |
|
293 | |||
293 | # bsession is the session as bytes |
|
294 | # bsession is the session as bytes | |
294 | bsession = CBytes(b'') |
|
295 | bsession = CBytes(b'') | |
295 |
|
296 | |||
296 | username = Unicode(str_to_unicode(os.environ.get('USER', 'username')), |
|
297 | username = Unicode(str_to_unicode(os.environ.get('USER', 'username')), | |
297 | help="""Username for the Session. Default is your system username.""", |
|
298 | help="""Username for the Session. Default is your system username.""", | |
298 | config=True) |
|
299 | config=True) | |
299 |
|
300 | |||
300 | metadata = Dict({}, config=True, |
|
301 | metadata = Dict({}, config=True, | |
301 | help="""Metadata dictionary, which serves as the default top-level metadata dict for each message.""") |
|
302 | help="""Metadata dictionary, which serves as the default top-level metadata dict for each message.""") | |
302 |
|
303 | |||
303 | # if 0, no adapting to do. |
|
304 | # if 0, no adapting to do. | |
304 | adapt_version = Integer(0) |
|
305 | adapt_version = Integer(0) | |
305 |
|
306 | |||
306 | # message signature related traits: |
|
307 | # message signature related traits: | |
307 |
|
308 | |||
308 | key = CBytes(b'', config=True, |
|
309 | key = CBytes(b'', config=True, | |
309 | help="""execution key, for extra authentication.""") |
|
310 | help="""execution key, for extra authentication.""") | |
310 | def _key_changed(self): |
|
311 | def _key_changed(self): | |
311 | self._new_auth() |
|
312 | self._new_auth() | |
312 |
|
313 | |||
313 | signature_scheme = Unicode('hmac-sha256', config=True, |
|
314 | signature_scheme = Unicode('hmac-sha256', config=True, | |
314 | help="""The digest scheme used to construct the message signatures. |
|
315 | help="""The digest scheme used to construct the message signatures. | |
315 | Must have the form 'hmac-HASH'.""") |
|
316 | Must have the form 'hmac-HASH'.""") | |
316 | def _signature_scheme_changed(self, name, old, new): |
|
317 | def _signature_scheme_changed(self, name, old, new): | |
317 | if not new.startswith('hmac-'): |
|
318 | if not new.startswith('hmac-'): | |
318 | raise TraitError("signature_scheme must start with 'hmac-', got %r" % new) |
|
319 | raise TraitError("signature_scheme must start with 'hmac-', got %r" % new) | |
319 | hash_name = new.split('-', 1)[1] |
|
320 | hash_name = new.split('-', 1)[1] | |
320 | try: |
|
321 | try: | |
321 | self.digest_mod = getattr(hashlib, hash_name) |
|
322 | self.digest_mod = getattr(hashlib, hash_name) | |
322 | except AttributeError: |
|
323 | except AttributeError: | |
323 | raise TraitError("hashlib has no such attribute: %s" % hash_name) |
|
324 | raise TraitError("hashlib has no such attribute: %s" % hash_name) | |
324 | self._new_auth() |
|
325 | self._new_auth() | |
325 |
|
326 | |||
326 | digest_mod = Any() |
|
327 | digest_mod = Any() | |
327 | def _digest_mod_default(self): |
|
328 | def _digest_mod_default(self): | |
328 | return hashlib.sha256 |
|
329 | return hashlib.sha256 | |
329 |
|
330 | |||
330 | auth = Instance(hmac.HMAC) |
|
331 | auth = Instance(hmac.HMAC) | |
331 |
|
332 | |||
332 | def _new_auth(self): |
|
333 | def _new_auth(self): | |
333 | if self.key: |
|
334 | if self.key: | |
334 | self.auth = hmac.HMAC(self.key, digestmod=self.digest_mod) |
|
335 | self.auth = hmac.HMAC(self.key, digestmod=self.digest_mod) | |
335 | else: |
|
336 | else: | |
336 | self.auth = None |
|
337 | self.auth = None | |
337 |
|
338 | |||
338 | digest_history = Set() |
|
339 | digest_history = Set() | |
339 | digest_history_size = Integer(2**16, config=True, |
|
340 | digest_history_size = Integer(2**16, config=True, | |
340 | help="""The maximum number of digests to remember. |
|
341 | help="""The maximum number of digests to remember. | |
341 |
|
342 | |||
342 | The digest history will be culled when it exceeds this value. |
|
343 | The digest history will be culled when it exceeds this value. | |
343 | """ |
|
344 | """ | |
344 | ) |
|
345 | ) | |
345 |
|
346 | |||
346 | keyfile = Unicode('', config=True, |
|
347 | keyfile = Unicode('', config=True, | |
347 | help="""path to file containing execution key.""") |
|
348 | help="""path to file containing execution key.""") | |
348 | def _keyfile_changed(self, name, old, new): |
|
349 | def _keyfile_changed(self, name, old, new): | |
349 | with open(new, 'rb') as f: |
|
350 | with open(new, 'rb') as f: | |
350 | self.key = f.read().strip() |
|
351 | self.key = f.read().strip() | |
351 |
|
352 | |||
352 | # for protecting against sends from forks |
|
353 | # for protecting against sends from forks | |
353 | pid = Integer() |
|
354 | pid = Integer() | |
354 |
|
355 | |||
355 | # serialization traits: |
|
356 | # serialization traits: | |
356 |
|
357 | |||
357 | pack = Any(default_packer) # the actual packer function |
|
358 | pack = Any(default_packer) # the actual packer function | |
358 | def _pack_changed(self, name, old, new): |
|
359 | def _pack_changed(self, name, old, new): | |
359 | if not callable(new): |
|
360 | if not callable(new): | |
360 | raise TypeError("packer must be callable, not %s"%type(new)) |
|
361 | raise TypeError("packer must be callable, not %s"%type(new)) | |
361 |
|
362 | |||
362 | unpack = Any(default_unpacker) # the actual packer function |
|
363 | unpack = Any(default_unpacker) # the actual packer function | |
363 | def _unpack_changed(self, name, old, new): |
|
364 | def _unpack_changed(self, name, old, new): | |
364 | # unpacker is not checked - it is assumed to be |
|
365 | # unpacker is not checked - it is assumed to be | |
365 | if not callable(new): |
|
366 | if not callable(new): | |
366 | raise TypeError("unpacker must be callable, not %s"%type(new)) |
|
367 | raise TypeError("unpacker must be callable, not %s"%type(new)) | |
367 |
|
368 | |||
368 | # thresholds: |
|
369 | # thresholds: | |
369 | copy_threshold = Integer(2**16, config=True, |
|
370 | copy_threshold = Integer(2**16, config=True, | |
370 | help="Threshold (in bytes) beyond which a buffer should be sent without copying.") |
|
371 | help="Threshold (in bytes) beyond which a buffer should be sent without copying.") | |
371 | buffer_threshold = Integer(MAX_BYTES, config=True, |
|
372 | buffer_threshold = Integer(MAX_BYTES, config=True, | |
372 | help="Threshold (in bytes) beyond which an object's buffer should be extracted to avoid pickling.") |
|
373 | help="Threshold (in bytes) beyond which an object's buffer should be extracted to avoid pickling.") | |
373 | item_threshold = Integer(MAX_ITEMS, config=True, |
|
374 | item_threshold = Integer(MAX_ITEMS, config=True, | |
374 | help="""The maximum number of items for a container to be introspected for custom serialization. |
|
375 | help="""The maximum number of items for a container to be introspected for custom serialization. | |
375 | Containers larger than this are pickled outright. |
|
376 | Containers larger than this are pickled outright. | |
376 | """ |
|
377 | """ | |
377 | ) |
|
378 | ) | |
378 |
|
379 | |||
379 |
|
380 | |||
380 | def __init__(self, **kwargs): |
|
381 | def __init__(self, **kwargs): | |
381 | """create a Session object |
|
382 | """create a Session object | |
382 |
|
383 | |||
383 | Parameters |
|
384 | Parameters | |
384 | ---------- |
|
385 | ---------- | |
385 |
|
386 | |||
386 | debug : bool |
|
387 | debug : bool | |
387 | whether to trigger extra debugging statements |
|
388 | whether to trigger extra debugging statements | |
388 | packer/unpacker : str : 'json', 'pickle' or import_string |
|
389 | packer/unpacker : str : 'json', 'pickle' or import_string | |
389 | importstrings for methods to serialize message parts. If just |
|
390 | importstrings for methods to serialize message parts. If just | |
390 | 'json' or 'pickle', predefined JSON and pickle packers will be used. |
|
391 | 'json' or 'pickle', predefined JSON and pickle packers will be used. | |
391 | Otherwise, the entire importstring must be used. |
|
392 | Otherwise, the entire importstring must be used. | |
392 |
|
393 | |||
393 | The functions must accept at least valid JSON input, and output |
|
394 | The functions must accept at least valid JSON input, and output | |
394 | *bytes*. |
|
395 | *bytes*. | |
395 |
|
396 | |||
396 | For example, to use msgpack: |
|
397 | For example, to use msgpack: | |
397 | packer = 'msgpack.packb', unpacker='msgpack.unpackb' |
|
398 | packer = 'msgpack.packb', unpacker='msgpack.unpackb' | |
398 | pack/unpack : callables |
|
399 | pack/unpack : callables | |
399 | You can also set the pack/unpack callables for serialization |
|
400 | You can also set the pack/unpack callables for serialization | |
400 | directly. |
|
401 | directly. | |
401 | session : unicode (must be ascii) |
|
402 | session : unicode (must be ascii) | |
402 | the ID of this Session object. The default is to generate a new |
|
403 | the ID of this Session object. The default is to generate a new | |
403 | UUID. |
|
404 | UUID. | |
404 | bsession : bytes |
|
405 | bsession : bytes | |
405 | The session as bytes |
|
406 | The session as bytes | |
406 | username : unicode |
|
407 | username : unicode | |
407 | username added to message headers. The default is to ask the OS. |
|
408 | username added to message headers. The default is to ask the OS. | |
408 | key : bytes |
|
409 | key : bytes | |
409 | The key used to initialize an HMAC signature. If unset, messages |
|
410 | The key used to initialize an HMAC signature. If unset, messages | |
410 | will not be signed or checked. |
|
411 | will not be signed or checked. | |
411 | signature_scheme : str |
|
412 | signature_scheme : str | |
412 | The message digest scheme. Currently must be of the form 'hmac-HASH', |
|
413 | The message digest scheme. Currently must be of the form 'hmac-HASH', | |
413 | where 'HASH' is a hashing function available in Python's hashlib. |
|
414 | where 'HASH' is a hashing function available in Python's hashlib. | |
414 | The default is 'hmac-sha256'. |
|
415 | The default is 'hmac-sha256'. | |
415 | This is ignored if 'key' is empty. |
|
416 | This is ignored if 'key' is empty. | |
416 | keyfile : filepath |
|
417 | keyfile : filepath | |
417 | The file containing a key. If this is set, `key` will be |
|
418 | The file containing a key. If this is set, `key` will be | |
418 | initialized to the contents of the file. |
|
419 | initialized to the contents of the file. | |
419 | """ |
|
420 | """ | |
420 | super(Session, self).__init__(**kwargs) |
|
421 | super(Session, self).__init__(**kwargs) | |
421 | self._check_packers() |
|
422 | self._check_packers() | |
422 | self.none = self.pack({}) |
|
423 | self.none = self.pack({}) | |
423 | # ensure self._session_default() if necessary, so bsession is defined: |
|
424 | # ensure self._session_default() if necessary, so bsession is defined: | |
424 | self.session |
|
425 | self.session | |
425 | self.pid = os.getpid() |
|
426 | self.pid = os.getpid() | |
426 |
|
427 | |||
427 | @property |
|
428 | @property | |
428 | def msg_id(self): |
|
429 | def msg_id(self): | |
429 | """always return new uuid""" |
|
430 | """always return new uuid""" | |
430 | return str(uuid.uuid4()) |
|
431 | return str(uuid.uuid4()) | |
431 |
|
432 | |||
432 | def _check_packers(self): |
|
433 | def _check_packers(self): | |
433 | """check packers for datetime support.""" |
|
434 | """check packers for datetime support.""" | |
434 | pack = self.pack |
|
435 | pack = self.pack | |
435 | unpack = self.unpack |
|
436 | unpack = self.unpack | |
436 |
|
437 | |||
437 | # check simple serialization |
|
438 | # check simple serialization | |
438 | msg = dict(a=[1,'hi']) |
|
439 | msg = dict(a=[1,'hi']) | |
439 | try: |
|
440 | try: | |
440 | packed = pack(msg) |
|
441 | packed = pack(msg) | |
441 | except Exception as e: |
|
442 | except Exception as e: | |
442 | msg = "packer '{packer}' could not serialize a simple message: {e}{jsonmsg}" |
|
443 | msg = "packer '{packer}' could not serialize a simple message: {e}{jsonmsg}" | |
443 | if self.packer == 'json': |
|
444 | if self.packer == 'json': | |
444 | jsonmsg = "\nzmq.utils.jsonapi.jsonmod = %s" % jsonapi.jsonmod |
|
445 | jsonmsg = "\nzmq.utils.jsonapi.jsonmod = %s" % jsonapi.jsonmod | |
445 | else: |
|
446 | else: | |
446 | jsonmsg = "" |
|
447 | jsonmsg = "" | |
447 | raise ValueError( |
|
448 | raise ValueError( | |
448 | msg.format(packer=self.packer, e=e, jsonmsg=jsonmsg) |
|
449 | msg.format(packer=self.packer, e=e, jsonmsg=jsonmsg) | |
449 | ) |
|
450 | ) | |
450 |
|
451 | |||
451 | # ensure packed message is bytes |
|
452 | # ensure packed message is bytes | |
452 | if not isinstance(packed, bytes): |
|
453 | if not isinstance(packed, bytes): | |
453 | raise ValueError("message packed to %r, but bytes are required"%type(packed)) |
|
454 | raise ValueError("message packed to %r, but bytes are required"%type(packed)) | |
454 |
|
455 | |||
455 | # check that unpack is pack's inverse |
|
456 | # check that unpack is pack's inverse | |
456 | try: |
|
457 | try: | |
457 | unpacked = unpack(packed) |
|
458 | unpacked = unpack(packed) | |
458 | assert unpacked == msg |
|
459 | assert unpacked == msg | |
459 | except Exception as e: |
|
460 | except Exception as e: | |
460 | msg = "unpacker '{unpacker}' could not handle output from packer '{packer}': {e}{jsonmsg}" |
|
461 | msg = "unpacker '{unpacker}' could not handle output from packer '{packer}': {e}{jsonmsg}" | |
461 | if self.packer == 'json': |
|
462 | if self.packer == 'json': | |
462 | jsonmsg = "\nzmq.utils.jsonapi.jsonmod = %s" % jsonapi.jsonmod |
|
463 | jsonmsg = "\nzmq.utils.jsonapi.jsonmod = %s" % jsonapi.jsonmod | |
463 | else: |
|
464 | else: | |
464 | jsonmsg = "" |
|
465 | jsonmsg = "" | |
465 | raise ValueError( |
|
466 | raise ValueError( | |
466 | msg.format(packer=self.packer, unpacker=self.unpacker, e=e, jsonmsg=jsonmsg) |
|
467 | msg.format(packer=self.packer, unpacker=self.unpacker, e=e, jsonmsg=jsonmsg) | |
467 | ) |
|
468 | ) | |
468 |
|
469 | |||
469 | # check datetime support |
|
470 | # check datetime support | |
470 | msg = dict(t=datetime.now()) |
|
471 | msg = dict(t=datetime.now()) | |
471 | try: |
|
472 | try: | |
472 | unpacked = unpack(pack(msg)) |
|
473 | unpacked = unpack(pack(msg)) | |
473 | if isinstance(unpacked['t'], datetime): |
|
474 | if isinstance(unpacked['t'], datetime): | |
474 | raise ValueError("Shouldn't deserialize to datetime") |
|
475 | raise ValueError("Shouldn't deserialize to datetime") | |
475 | except Exception: |
|
476 | except Exception: | |
476 | self.pack = lambda o: pack(squash_dates(o)) |
|
477 | self.pack = lambda o: pack(squash_dates(o)) | |
477 | self.unpack = lambda s: unpack(s) |
|
478 | self.unpack = lambda s: unpack(s) | |
478 |
|
479 | |||
479 | def msg_header(self, msg_type): |
|
480 | def msg_header(self, msg_type): | |
480 | return msg_header(self.msg_id, msg_type, self.username, self.session) |
|
481 | return msg_header(self.msg_id, msg_type, self.username, self.session) | |
481 |
|
482 | |||
482 | def msg(self, msg_type, content=None, parent=None, header=None, metadata=None): |
|
483 | def msg(self, msg_type, content=None, parent=None, header=None, metadata=None): | |
483 | """Return the nested message dict. |
|
484 | """Return the nested message dict. | |
484 |
|
485 | |||
485 | This format is different from what is sent over the wire. The |
|
486 | This format is different from what is sent over the wire. The | |
486 | serialize/unserialize methods converts this nested message dict to the wire |
|
487 | serialize/unserialize methods converts this nested message dict to the wire | |
487 | format, which is a list of message parts. |
|
488 | format, which is a list of message parts. | |
488 | """ |
|
489 | """ | |
489 | msg = {} |
|
490 | msg = {} | |
490 | header = self.msg_header(msg_type) if header is None else header |
|
491 | header = self.msg_header(msg_type) if header is None else header | |
491 | msg['header'] = header |
|
492 | msg['header'] = header | |
492 | msg['msg_id'] = header['msg_id'] |
|
493 | msg['msg_id'] = header['msg_id'] | |
493 | msg['msg_type'] = header['msg_type'] |
|
494 | msg['msg_type'] = header['msg_type'] | |
494 | msg['parent_header'] = {} if parent is None else extract_header(parent) |
|
495 | msg['parent_header'] = {} if parent is None else extract_header(parent) | |
495 | msg['content'] = {} if content is None else content |
|
496 | msg['content'] = {} if content is None else content | |
496 | msg['metadata'] = self.metadata.copy() |
|
497 | msg['metadata'] = self.metadata.copy() | |
497 | if metadata is not None: |
|
498 | if metadata is not None: | |
498 | msg['metadata'].update(metadata) |
|
499 | msg['metadata'].update(metadata) | |
499 | return msg |
|
500 | return msg | |
500 |
|
501 | |||
501 | def sign(self, msg_list): |
|
502 | def sign(self, msg_list): | |
502 | """Sign a message with HMAC digest. If no auth, return b''. |
|
503 | """Sign a message with HMAC digest. If no auth, return b''. | |
503 |
|
504 | |||
504 | Parameters |
|
505 | Parameters | |
505 | ---------- |
|
506 | ---------- | |
506 | msg_list : list |
|
507 | msg_list : list | |
507 | The [p_header,p_parent,p_content] part of the message list. |
|
508 | The [p_header,p_parent,p_content] part of the message list. | |
508 | """ |
|
509 | """ | |
509 | if self.auth is None: |
|
510 | if self.auth is None: | |
510 | return b'' |
|
511 | return b'' | |
511 | h = self.auth.copy() |
|
512 | h = self.auth.copy() | |
512 | for m in msg_list: |
|
513 | for m in msg_list: | |
513 | h.update(m) |
|
514 | h.update(m) | |
514 | return str_to_bytes(h.hexdigest()) |
|
515 | return str_to_bytes(h.hexdigest()) | |
515 |
|
516 | |||
516 | def serialize(self, msg, ident=None): |
|
517 | def serialize(self, msg, ident=None): | |
517 | """Serialize the message components to bytes. |
|
518 | """Serialize the message components to bytes. | |
518 |
|
519 | |||
519 | This is roughly the inverse of unserialize. The serialize/unserialize |
|
520 | This is roughly the inverse of unserialize. The serialize/unserialize | |
520 | methods work with full message lists, whereas pack/unpack work with |
|
521 | methods work with full message lists, whereas pack/unpack work with | |
521 | the individual message parts in the message list. |
|
522 | the individual message parts in the message list. | |
522 |
|
523 | |||
523 | Parameters |
|
524 | Parameters | |
524 | ---------- |
|
525 | ---------- | |
525 | msg : dict or Message |
|
526 | msg : dict or Message | |
526 | The nexted message dict as returned by the self.msg method. |
|
527 | The nexted message dict as returned by the self.msg method. | |
527 |
|
528 | |||
528 | Returns |
|
529 | Returns | |
529 | ------- |
|
530 | ------- | |
530 | msg_list : list |
|
531 | msg_list : list | |
531 | The list of bytes objects to be sent with the format:: |
|
532 | The list of bytes objects to be sent with the format:: | |
532 |
|
533 | |||
533 | [ident1, ident2, ..., DELIM, HMAC, p_header, p_parent, |
|
534 | [ident1, ident2, ..., DELIM, HMAC, p_header, p_parent, | |
534 | p_metadata, p_content, buffer1, buffer2, ...] |
|
535 | p_metadata, p_content, buffer1, buffer2, ...] | |
535 |
|
536 | |||
536 | In this list, the ``p_*`` entities are the packed or serialized |
|
537 | In this list, the ``p_*`` entities are the packed or serialized | |
537 | versions, so if JSON is used, these are utf8 encoded JSON strings. |
|
538 | versions, so if JSON is used, these are utf8 encoded JSON strings. | |
538 | """ |
|
539 | """ | |
539 | content = msg.get('content', {}) |
|
540 | content = msg.get('content', {}) | |
540 | if content is None: |
|
541 | if content is None: | |
541 | content = self.none |
|
542 | content = self.none | |
542 | elif isinstance(content, dict): |
|
543 | elif isinstance(content, dict): | |
543 | content = self.pack(content) |
|
544 | content = self.pack(content) | |
544 | elif isinstance(content, bytes): |
|
545 | elif isinstance(content, bytes): | |
545 | # content is already packed, as in a relayed message |
|
546 | # content is already packed, as in a relayed message | |
546 | pass |
|
547 | pass | |
547 | elif isinstance(content, unicode_type): |
|
548 | elif isinstance(content, unicode_type): | |
548 | # should be bytes, but JSON often spits out unicode |
|
549 | # should be bytes, but JSON often spits out unicode | |
549 | content = content.encode('utf8') |
|
550 | content = content.encode('utf8') | |
550 | else: |
|
551 | else: | |
551 | raise TypeError("Content incorrect type: %s"%type(content)) |
|
552 | raise TypeError("Content incorrect type: %s"%type(content)) | |
552 |
|
553 | |||
553 | real_message = [self.pack(msg['header']), |
|
554 | real_message = [self.pack(msg['header']), | |
554 | self.pack(msg['parent_header']), |
|
555 | self.pack(msg['parent_header']), | |
555 | self.pack(msg['metadata']), |
|
556 | self.pack(msg['metadata']), | |
556 | content, |
|
557 | content, | |
557 | ] |
|
558 | ] | |
558 |
|
559 | |||
559 | to_send = [] |
|
560 | to_send = [] | |
560 |
|
561 | |||
561 | if isinstance(ident, list): |
|
562 | if isinstance(ident, list): | |
562 | # accept list of idents |
|
563 | # accept list of idents | |
563 | to_send.extend(ident) |
|
564 | to_send.extend(ident) | |
564 | elif ident is not None: |
|
565 | elif ident is not None: | |
565 | to_send.append(ident) |
|
566 | to_send.append(ident) | |
566 | to_send.append(DELIM) |
|
567 | to_send.append(DELIM) | |
567 |
|
568 | |||
568 | signature = self.sign(real_message) |
|
569 | signature = self.sign(real_message) | |
569 | to_send.append(signature) |
|
570 | to_send.append(signature) | |
570 |
|
571 | |||
571 | to_send.extend(real_message) |
|
572 | to_send.extend(real_message) | |
572 |
|
573 | |||
573 | return to_send |
|
574 | return to_send | |
574 |
|
575 | |||
575 | def send(self, stream, msg_or_type, content=None, parent=None, ident=None, |
|
576 | def send(self, stream, msg_or_type, content=None, parent=None, ident=None, | |
576 | buffers=None, track=False, header=None, metadata=None): |
|
577 | buffers=None, track=False, header=None, metadata=None): | |
577 | """Build and send a message via stream or socket. |
|
578 | """Build and send a message via stream or socket. | |
578 |
|
579 | |||
579 | The message format used by this function internally is as follows: |
|
580 | The message format used by this function internally is as follows: | |
580 |
|
581 | |||
581 | [ident1,ident2,...,DELIM,HMAC,p_header,p_parent,p_content, |
|
582 | [ident1,ident2,...,DELIM,HMAC,p_header,p_parent,p_content, | |
582 | buffer1,buffer2,...] |
|
583 | buffer1,buffer2,...] | |
583 |
|
584 | |||
584 | The serialize/unserialize methods convert the nested message dict into this |
|
585 | The serialize/unserialize methods convert the nested message dict into this | |
585 | format. |
|
586 | format. | |
586 |
|
587 | |||
587 | Parameters |
|
588 | Parameters | |
588 | ---------- |
|
589 | ---------- | |
589 |
|
590 | |||
590 | stream : zmq.Socket or ZMQStream |
|
591 | stream : zmq.Socket or ZMQStream | |
591 | The socket-like object used to send the data. |
|
592 | The socket-like object used to send the data. | |
592 | msg_or_type : str or Message/dict |
|
593 | msg_or_type : str or Message/dict | |
593 | Normally, msg_or_type will be a msg_type unless a message is being |
|
594 | Normally, msg_or_type will be a msg_type unless a message is being | |
594 | sent more than once. If a header is supplied, this can be set to |
|
595 | sent more than once. If a header is supplied, this can be set to | |
595 | None and the msg_type will be pulled from the header. |
|
596 | None and the msg_type will be pulled from the header. | |
596 |
|
597 | |||
597 | content : dict or None |
|
598 | content : dict or None | |
598 | The content of the message (ignored if msg_or_type is a message). |
|
599 | The content of the message (ignored if msg_or_type is a message). | |
599 | header : dict or None |
|
600 | header : dict or None | |
600 | The header dict for the message (ignored if msg_to_type is a message). |
|
601 | The header dict for the message (ignored if msg_to_type is a message). | |
601 | parent : Message or dict or None |
|
602 | parent : Message or dict or None | |
602 | The parent or parent header describing the parent of this message |
|
603 | The parent or parent header describing the parent of this message | |
603 | (ignored if msg_or_type is a message). |
|
604 | (ignored if msg_or_type is a message). | |
604 | ident : bytes or list of bytes |
|
605 | ident : bytes or list of bytes | |
605 | The zmq.IDENTITY routing path. |
|
606 | The zmq.IDENTITY routing path. | |
606 | metadata : dict or None |
|
607 | metadata : dict or None | |
607 | The metadata describing the message |
|
608 | The metadata describing the message | |
608 | buffers : list or None |
|
609 | buffers : list or None | |
609 | The already-serialized buffers to be appended to the message. |
|
610 | The already-serialized buffers to be appended to the message. | |
610 | track : bool |
|
611 | track : bool | |
611 | Whether to track. Only for use with Sockets, because ZMQStream |
|
612 | Whether to track. Only for use with Sockets, because ZMQStream | |
612 | objects cannot track messages. |
|
613 | objects cannot track messages. | |
613 |
|
614 | |||
614 |
|
615 | |||
615 | Returns |
|
616 | Returns | |
616 | ------- |
|
617 | ------- | |
617 | msg : dict |
|
618 | msg : dict | |
618 | The constructed message. |
|
619 | The constructed message. | |
619 | """ |
|
620 | """ | |
620 | if not isinstance(stream, zmq.Socket): |
|
621 | if not isinstance(stream, zmq.Socket): | |
621 | # ZMQStreams and dummy sockets do not support tracking. |
|
622 | # ZMQStreams and dummy sockets do not support tracking. | |
622 | track = False |
|
623 | track = False | |
623 |
|
624 | |||
624 | if isinstance(msg_or_type, (Message, dict)): |
|
625 | if isinstance(msg_or_type, (Message, dict)): | |
625 | # We got a Message or message dict, not a msg_type so don't |
|
626 | # We got a Message or message dict, not a msg_type so don't | |
626 | # build a new Message. |
|
627 | # build a new Message. | |
627 | msg = msg_or_type |
|
628 | msg = msg_or_type | |
628 | else: |
|
629 | else: | |
629 | msg = self.msg(msg_or_type, content=content, parent=parent, |
|
630 | msg = self.msg(msg_or_type, content=content, parent=parent, | |
630 | header=header, metadata=metadata) |
|
631 | header=header, metadata=metadata) | |
631 | if not os.getpid() == self.pid: |
|
632 | if not os.getpid() == self.pid: | |
632 | io.rprint("WARNING: attempted to send message from fork") |
|
633 | io.rprint("WARNING: attempted to send message from fork") | |
633 | io.rprint(msg) |
|
634 | io.rprint(msg) | |
634 | return |
|
635 | return | |
635 | buffers = [] if buffers is None else buffers |
|
636 | buffers = [] if buffers is None else buffers | |
636 | if self.adapt_version: |
|
637 | if self.adapt_version: | |
637 | msg = adapt(msg, self.adapt_version) |
|
638 | msg = adapt(msg, self.adapt_version) | |
638 | to_send = self.serialize(msg, ident) |
|
639 | to_send = self.serialize(msg, ident) | |
639 | to_send.extend(buffers) |
|
640 | to_send.extend(buffers) | |
640 | longest = max([ len(s) for s in to_send ]) |
|
641 | longest = max([ len(s) for s in to_send ]) | |
641 | copy = (longest < self.copy_threshold) |
|
642 | copy = (longest < self.copy_threshold) | |
642 |
|
643 | |||
643 | if buffers and track and not copy: |
|
644 | if buffers and track and not copy: | |
644 | # only really track when we are doing zero-copy buffers |
|
645 | # only really track when we are doing zero-copy buffers | |
645 | tracker = stream.send_multipart(to_send, copy=False, track=True) |
|
646 | tracker = stream.send_multipart(to_send, copy=False, track=True) | |
646 | else: |
|
647 | else: | |
647 | # use dummy tracker, which will be done immediately |
|
648 | # use dummy tracker, which will be done immediately | |
648 | tracker = DONE |
|
649 | tracker = DONE | |
649 | stream.send_multipart(to_send, copy=copy) |
|
650 | stream.send_multipart(to_send, copy=copy) | |
650 |
|
651 | |||
651 | if self.debug: |
|
652 | if self.debug: | |
652 | pprint.pprint(msg) |
|
653 | pprint.pprint(msg) | |
653 | pprint.pprint(to_send) |
|
654 | pprint.pprint(to_send) | |
654 | pprint.pprint(buffers) |
|
655 | pprint.pprint(buffers) | |
655 |
|
656 | |||
656 | msg['tracker'] = tracker |
|
657 | msg['tracker'] = tracker | |
657 |
|
658 | |||
658 | return msg |
|
659 | return msg | |
659 |
|
660 | |||
660 | def send_raw(self, stream, msg_list, flags=0, copy=True, ident=None): |
|
661 | def send_raw(self, stream, msg_list, flags=0, copy=True, ident=None): | |
661 | """Send a raw message via ident path. |
|
662 | """Send a raw message via ident path. | |
662 |
|
663 | |||
663 | This method is used to send a already serialized message. |
|
664 | This method is used to send a already serialized message. | |
664 |
|
665 | |||
665 | Parameters |
|
666 | Parameters | |
666 | ---------- |
|
667 | ---------- | |
667 | stream : ZMQStream or Socket |
|
668 | stream : ZMQStream or Socket | |
668 | The ZMQ stream or socket to use for sending the message. |
|
669 | The ZMQ stream or socket to use for sending the message. | |
669 | msg_list : list |
|
670 | msg_list : list | |
670 | The serialized list of messages to send. This only includes the |
|
671 | The serialized list of messages to send. This only includes the | |
671 | [p_header,p_parent,p_metadata,p_content,buffer1,buffer2,...] portion of |
|
672 | [p_header,p_parent,p_metadata,p_content,buffer1,buffer2,...] portion of | |
672 | the message. |
|
673 | the message. | |
673 | ident : ident or list |
|
674 | ident : ident or list | |
674 | A single ident or a list of idents to use in sending. |
|
675 | A single ident or a list of idents to use in sending. | |
675 | """ |
|
676 | """ | |
676 | to_send = [] |
|
677 | to_send = [] | |
677 | if isinstance(ident, bytes): |
|
678 | if isinstance(ident, bytes): | |
678 | ident = [ident] |
|
679 | ident = [ident] | |
679 | if ident is not None: |
|
680 | if ident is not None: | |
680 | to_send.extend(ident) |
|
681 | to_send.extend(ident) | |
681 |
|
682 | |||
682 | to_send.append(DELIM) |
|
683 | to_send.append(DELIM) | |
683 | to_send.append(self.sign(msg_list)) |
|
684 | to_send.append(self.sign(msg_list)) | |
684 | to_send.extend(msg_list) |
|
685 | to_send.extend(msg_list) | |
685 | stream.send_multipart(to_send, flags, copy=copy) |
|
686 | stream.send_multipart(to_send, flags, copy=copy) | |
686 |
|
687 | |||
687 | def recv(self, socket, mode=zmq.NOBLOCK, content=True, copy=True): |
|
688 | def recv(self, socket, mode=zmq.NOBLOCK, content=True, copy=True): | |
688 | """Receive and unpack a message. |
|
689 | """Receive and unpack a message. | |
689 |
|
690 | |||
690 | Parameters |
|
691 | Parameters | |
691 | ---------- |
|
692 | ---------- | |
692 | socket : ZMQStream or Socket |
|
693 | socket : ZMQStream or Socket | |
693 | The socket or stream to use in receiving. |
|
694 | The socket or stream to use in receiving. | |
694 |
|
695 | |||
695 | Returns |
|
696 | Returns | |
696 | ------- |
|
697 | ------- | |
697 | [idents], msg |
|
698 | [idents], msg | |
698 | [idents] is a list of idents and msg is a nested message dict of |
|
699 | [idents] is a list of idents and msg is a nested message dict of | |
699 | same format as self.msg returns. |
|
700 | same format as self.msg returns. | |
700 | """ |
|
701 | """ | |
701 | if isinstance(socket, ZMQStream): |
|
702 | if isinstance(socket, ZMQStream): | |
702 | socket = socket.socket |
|
703 | socket = socket.socket | |
703 | try: |
|
704 | try: | |
704 | msg_list = socket.recv_multipart(mode, copy=copy) |
|
705 | msg_list = socket.recv_multipart(mode, copy=copy) | |
705 | except zmq.ZMQError as e: |
|
706 | except zmq.ZMQError as e: | |
706 | if e.errno == zmq.EAGAIN: |
|
707 | if e.errno == zmq.EAGAIN: | |
707 | # We can convert EAGAIN to None as we know in this case |
|
708 | # We can convert EAGAIN to None as we know in this case | |
708 | # recv_multipart won't return None. |
|
709 | # recv_multipart won't return None. | |
709 | return None,None |
|
710 | return None,None | |
710 | else: |
|
711 | else: | |
711 | raise |
|
712 | raise | |
712 | # split multipart message into identity list and message dict |
|
713 | # split multipart message into identity list and message dict | |
713 | # invalid large messages can cause very expensive string comparisons |
|
714 | # invalid large messages can cause very expensive string comparisons | |
714 | idents, msg_list = self.feed_identities(msg_list, copy) |
|
715 | idents, msg_list = self.feed_identities(msg_list, copy) | |
715 | try: |
|
716 | try: | |
716 | return idents, self.unserialize(msg_list, content=content, copy=copy) |
|
717 | return idents, self.unserialize(msg_list, content=content, copy=copy) | |
717 | except Exception as e: |
|
718 | except Exception as e: | |
718 | # TODO: handle it |
|
719 | # TODO: handle it | |
719 | raise e |
|
720 | raise e | |
720 |
|
721 | |||
721 | def feed_identities(self, msg_list, copy=True): |
|
722 | def feed_identities(self, msg_list, copy=True): | |
722 | """Split the identities from the rest of the message. |
|
723 | """Split the identities from the rest of the message. | |
723 |
|
724 | |||
724 | Feed until DELIM is reached, then return the prefix as idents and |
|
725 | Feed until DELIM is reached, then return the prefix as idents and | |
725 | remainder as msg_list. This is easily broken by setting an IDENT to DELIM, |
|
726 | remainder as msg_list. This is easily broken by setting an IDENT to DELIM, | |
726 | but that would be silly. |
|
727 | but that would be silly. | |
727 |
|
728 | |||
728 | Parameters |
|
729 | Parameters | |
729 | ---------- |
|
730 | ---------- | |
730 | msg_list : a list of Message or bytes objects |
|
731 | msg_list : a list of Message or bytes objects | |
731 | The message to be split. |
|
732 | The message to be split. | |
732 | copy : bool |
|
733 | copy : bool | |
733 | flag determining whether the arguments are bytes or Messages |
|
734 | flag determining whether the arguments are bytes or Messages | |
734 |
|
735 | |||
735 | Returns |
|
736 | Returns | |
736 | ------- |
|
737 | ------- | |
737 | (idents, msg_list) : two lists |
|
738 | (idents, msg_list) : two lists | |
738 | idents will always be a list of bytes, each of which is a ZMQ |
|
739 | idents will always be a list of bytes, each of which is a ZMQ | |
739 | identity. msg_list will be a list of bytes or zmq.Messages of the |
|
740 | identity. msg_list will be a list of bytes or zmq.Messages of the | |
740 | form [HMAC,p_header,p_parent,p_content,buffer1,buffer2,...] and |
|
741 | form [HMAC,p_header,p_parent,p_content,buffer1,buffer2,...] and | |
741 | should be unpackable/unserializable via self.unserialize at this |
|
742 | should be unpackable/unserializable via self.unserialize at this | |
742 | point. |
|
743 | point. | |
743 | """ |
|
744 | """ | |
744 | if copy: |
|
745 | if copy: | |
745 | idx = msg_list.index(DELIM) |
|
746 | idx = msg_list.index(DELIM) | |
746 | return msg_list[:idx], msg_list[idx+1:] |
|
747 | return msg_list[:idx], msg_list[idx+1:] | |
747 | else: |
|
748 | else: | |
748 | failed = True |
|
749 | failed = True | |
749 | for idx,m in enumerate(msg_list): |
|
750 | for idx,m in enumerate(msg_list): | |
750 | if m.bytes == DELIM: |
|
751 | if m.bytes == DELIM: | |
751 | failed = False |
|
752 | failed = False | |
752 | break |
|
753 | break | |
753 | if failed: |
|
754 | if failed: | |
754 | raise ValueError("DELIM not in msg_list") |
|
755 | raise ValueError("DELIM not in msg_list") | |
755 | idents, msg_list = msg_list[:idx], msg_list[idx+1:] |
|
756 | idents, msg_list = msg_list[:idx], msg_list[idx+1:] | |
756 | return [m.bytes for m in idents], msg_list |
|
757 | return [m.bytes for m in idents], msg_list | |
757 |
|
758 | |||
758 | def _add_digest(self, signature): |
|
759 | def _add_digest(self, signature): | |
759 | """add a digest to history to protect against replay attacks""" |
|
760 | """add a digest to history to protect against replay attacks""" | |
760 | if self.digest_history_size == 0: |
|
761 | if self.digest_history_size == 0: | |
761 | # no history, never add digests |
|
762 | # no history, never add digests | |
762 | return |
|
763 | return | |
763 |
|
764 | |||
764 | self.digest_history.add(signature) |
|
765 | self.digest_history.add(signature) | |
765 | if len(self.digest_history) > self.digest_history_size: |
|
766 | if len(self.digest_history) > self.digest_history_size: | |
766 | # threshold reached, cull 10% |
|
767 | # threshold reached, cull 10% | |
767 | self._cull_digest_history() |
|
768 | self._cull_digest_history() | |
768 |
|
769 | |||
769 | def _cull_digest_history(self): |
|
770 | def _cull_digest_history(self): | |
770 | """cull the digest history |
|
771 | """cull the digest history | |
771 |
|
772 | |||
772 | Removes a randomly selected 10% of the digest history |
|
773 | Removes a randomly selected 10% of the digest history | |
773 | """ |
|
774 | """ | |
774 | current = len(self.digest_history) |
|
775 | current = len(self.digest_history) | |
775 | n_to_cull = max(int(current // 10), current - self.digest_history_size) |
|
776 | n_to_cull = max(int(current // 10), current - self.digest_history_size) | |
776 | if n_to_cull >= current: |
|
777 | if n_to_cull >= current: | |
777 | self.digest_history = set() |
|
778 | self.digest_history = set() | |
778 | return |
|
779 | return | |
779 | to_cull = random.sample(self.digest_history, n_to_cull) |
|
780 | to_cull = random.sample(self.digest_history, n_to_cull) | |
780 | self.digest_history.difference_update(to_cull) |
|
781 | self.digest_history.difference_update(to_cull) | |
781 |
|
782 | |||
782 | def unserialize(self, msg_list, content=True, copy=True): |
|
783 | def unserialize(self, msg_list, content=True, copy=True): | |
783 | """Unserialize a msg_list to a nested message dict. |
|
784 | """Unserialize a msg_list to a nested message dict. | |
784 |
|
785 | |||
785 | This is roughly the inverse of serialize. The serialize/unserialize |
|
786 | This is roughly the inverse of serialize. The serialize/unserialize | |
786 | methods work with full message lists, whereas pack/unpack work with |
|
787 | methods work with full message lists, whereas pack/unpack work with | |
787 | the individual message parts in the message list. |
|
788 | the individual message parts in the message list. | |
788 |
|
789 | |||
789 | Parameters |
|
790 | Parameters | |
790 | ---------- |
|
791 | ---------- | |
791 | msg_list : list of bytes or Message objects |
|
792 | msg_list : list of bytes or Message objects | |
792 | The list of message parts of the form [HMAC,p_header,p_parent, |
|
793 | The list of message parts of the form [HMAC,p_header,p_parent, | |
793 | p_metadata,p_content,buffer1,buffer2,...]. |
|
794 | p_metadata,p_content,buffer1,buffer2,...]. | |
794 | content : bool (True) |
|
795 | content : bool (True) | |
795 | Whether to unpack the content dict (True), or leave it packed |
|
796 | Whether to unpack the content dict (True), or leave it packed | |
796 | (False). |
|
797 | (False). | |
797 | copy : bool (True) |
|
798 | copy : bool (True) | |
798 | Whether to return the bytes (True), or the non-copying Message |
|
799 | Whether to return the bytes (True), or the non-copying Message | |
799 | object in each place (False). |
|
800 | object in each place (False). | |
800 |
|
801 | |||
801 | Returns |
|
802 | Returns | |
802 | ------- |
|
803 | ------- | |
803 | msg : dict |
|
804 | msg : dict | |
804 | The nested message dict with top-level keys [header, parent_header, |
|
805 | The nested message dict with top-level keys [header, parent_header, | |
805 | content, buffers]. |
|
806 | content, buffers]. | |
806 | """ |
|
807 | """ | |
807 | minlen = 5 |
|
808 | minlen = 5 | |
808 | message = {} |
|
809 | message = {} | |
809 | if not copy: |
|
810 | if not copy: | |
810 | for i in range(minlen): |
|
811 | for i in range(minlen): | |
811 | msg_list[i] = msg_list[i].bytes |
|
812 | msg_list[i] = msg_list[i].bytes | |
812 | if self.auth is not None: |
|
813 | if self.auth is not None: | |
813 | signature = msg_list[0] |
|
814 | signature = msg_list[0] | |
814 | if not signature: |
|
815 | if not signature: | |
815 | raise ValueError("Unsigned Message") |
|
816 | raise ValueError("Unsigned Message") | |
816 | if signature in self.digest_history: |
|
817 | if signature in self.digest_history: | |
817 | raise ValueError("Duplicate Signature: %r" % signature) |
|
818 | raise ValueError("Duplicate Signature: %r" % signature) | |
818 | self._add_digest(signature) |
|
819 | self._add_digest(signature) | |
819 | check = self.sign(msg_list[1:5]) |
|
820 | check = self.sign(msg_list[1:5]) | |
820 | if not signature == check: |
|
821 | if not signature == check: | |
821 | raise ValueError("Invalid Signature: %r" % signature) |
|
822 | raise ValueError("Invalid Signature: %r" % signature) | |
822 | if not len(msg_list) >= minlen: |
|
823 | if not len(msg_list) >= minlen: | |
823 | raise TypeError("malformed message, must have at least %i elements"%minlen) |
|
824 | raise TypeError("malformed message, must have at least %i elements"%minlen) | |
824 | header = self.unpack(msg_list[1]) |
|
825 | header = self.unpack(msg_list[1]) | |
825 | message['header'] = extract_dates(header) |
|
826 | message['header'] = extract_dates(header) | |
826 | message['msg_id'] = header['msg_id'] |
|
827 | message['msg_id'] = header['msg_id'] | |
827 | message['msg_type'] = header['msg_type'] |
|
828 | message['msg_type'] = header['msg_type'] | |
828 | message['parent_header'] = extract_dates(self.unpack(msg_list[2])) |
|
829 | message['parent_header'] = extract_dates(self.unpack(msg_list[2])) | |
829 | message['metadata'] = self.unpack(msg_list[3]) |
|
830 | message['metadata'] = self.unpack(msg_list[3]) | |
830 | if content: |
|
831 | if content: | |
831 | message['content'] = self.unpack(msg_list[4]) |
|
832 | message['content'] = self.unpack(msg_list[4]) | |
832 | else: |
|
833 | else: | |
833 | message['content'] = msg_list[4] |
|
834 | message['content'] = msg_list[4] | |
834 |
|
835 | |||
835 | message['buffers'] = msg_list[5:] |
|
836 | message['buffers'] = msg_list[5:] | |
836 | # print("received: %s: %s\n %s" % (message['msg_type'], message['header'], message['content'])) |
|
837 | # print("received: %s: %s\n %s" % (message['msg_type'], message['header'], message['content'])) | |
837 | # adapt to the current version |
|
838 | # adapt to the current version | |
838 | return adapt(message) |
|
839 | return adapt(message) | |
839 | # print("adapted: %s: %s\n %s" % (adapted['msg_type'], adapted['header'], adapted['content'])) |
|
840 | # print("adapted: %s: %s\n %s" % (adapted['msg_type'], adapted['header'], adapted['content'])) | |
840 |
|
841 | |||
841 | def test_msg2obj(): |
|
842 | def test_msg2obj(): | |
842 | am = dict(x=1) |
|
843 | am = dict(x=1) | |
843 | ao = Message(am) |
|
844 | ao = Message(am) | |
844 | assert ao.x == am['x'] |
|
845 | assert ao.x == am['x'] | |
845 |
|
846 | |||
846 | am['y'] = dict(z=1) |
|
847 | am['y'] = dict(z=1) | |
847 | ao = Message(am) |
|
848 | ao = Message(am) | |
848 | assert ao.y.z == am['y']['z'] |
|
849 | assert ao.y.z == am['y']['z'] | |
849 |
|
850 | |||
850 | k1, k2 = 'y', 'z' |
|
851 | k1, k2 = 'y', 'z' | |
851 | assert ao[k1][k2] == am[k1][k2] |
|
852 | assert ao[k1][k2] == am[k1][k2] | |
852 |
|
853 | |||
853 | am2 = dict(ao) |
|
854 | am2 = dict(ao) | |
854 | assert am['x'] == am2['x'] |
|
855 | assert am['x'] == am2['x'] | |
855 | assert am['y']['z'] == am2['y']['z'] |
|
856 | assert am['y']['z'] == am2['y']['z'] | |
856 |
|
857 |
@@ -1,433 +1,438 b'' | |||||
1 | # encoding: utf-8 |
|
1 | # encoding: utf-8 | |
2 | """Pickle related utilities. Perhaps this should be called 'can'.""" |
|
2 | """Pickle related utilities. Perhaps this should be called 'can'.""" | |
3 |
|
3 | |||
4 | # Copyright (c) IPython Development Team. |
|
4 | # Copyright (c) IPython Development Team. | |
5 | # Distributed under the terms of the Modified BSD License. |
|
5 | # Distributed under the terms of the Modified BSD License. | |
6 |
|
6 | |||
7 | import copy |
|
7 | import copy | |
8 | import logging |
|
8 | import logging | |
9 | import sys |
|
9 | import sys | |
10 | from types import FunctionType |
|
10 | from types import FunctionType | |
11 |
|
11 | |||
12 | try: |
|
12 | try: | |
13 | import cPickle as pickle |
|
13 | import cPickle as pickle | |
14 | except ImportError: |
|
14 | except ImportError: | |
15 | import pickle |
|
15 | import pickle | |
16 |
|
16 | |||
17 | from . import codeutil # This registers a hook when it's imported |
|
17 | from . import codeutil # This registers a hook when it's imported | |
18 | from . import py3compat |
|
18 | from . import py3compat | |
19 | from .importstring import import_item |
|
19 | from .importstring import import_item | |
20 | from .py3compat import string_types, iteritems |
|
20 | from .py3compat import string_types, iteritems | |
21 |
|
21 | |||
22 | from IPython.config import Application |
|
22 | from IPython.config import Application | |
23 |
|
23 | |||
24 | if py3compat.PY3: |
|
24 | if py3compat.PY3: | |
25 | buffer = memoryview |
|
25 | buffer = memoryview | |
26 | class_type = type |
|
26 | class_type = type | |
27 | else: |
|
27 | else: | |
28 | from types import ClassType |
|
28 | from types import ClassType | |
29 | class_type = (type, ClassType) |
|
29 | class_type = (type, ClassType) | |
30 |
|
30 | |||
|
31 | try: | |||
|
32 | PICKLE_PROTOCOL = pickle.DEFAULT_PROTOCOL | |||
|
33 | except AttributeError: | |||
|
34 | PICKLE_PROTOCOL = pickle.HIGHEST_PROTOCOL | |||
|
35 | ||||
31 | def _get_cell_type(a=None): |
|
36 | def _get_cell_type(a=None): | |
32 | """the type of a closure cell doesn't seem to be importable, |
|
37 | """the type of a closure cell doesn't seem to be importable, | |
33 | so just create one |
|
38 | so just create one | |
34 | """ |
|
39 | """ | |
35 | def inner(): |
|
40 | def inner(): | |
36 | return a |
|
41 | return a | |
37 | return type(py3compat.get_closure(inner)[0]) |
|
42 | return type(py3compat.get_closure(inner)[0]) | |
38 |
|
43 | |||
39 | cell_type = _get_cell_type() |
|
44 | cell_type = _get_cell_type() | |
40 |
|
45 | |||
41 | #------------------------------------------------------------------------------- |
|
46 | #------------------------------------------------------------------------------- | |
42 | # Functions |
|
47 | # Functions | |
43 | #------------------------------------------------------------------------------- |
|
48 | #------------------------------------------------------------------------------- | |
44 |
|
49 | |||
45 |
|
50 | |||
46 | def use_dill(): |
|
51 | def use_dill(): | |
47 | """use dill to expand serialization support |
|
52 | """use dill to expand serialization support | |
48 |
|
53 | |||
49 | adds support for object methods and closures to serialization. |
|
54 | adds support for object methods and closures to serialization. | |
50 | """ |
|
55 | """ | |
51 | # import dill causes most of the magic |
|
56 | # import dill causes most of the magic | |
52 | import dill |
|
57 | import dill | |
53 |
|
58 | |||
54 | # dill doesn't work with cPickle, |
|
59 | # dill doesn't work with cPickle, | |
55 | # tell the two relevant modules to use plain pickle |
|
60 | # tell the two relevant modules to use plain pickle | |
56 |
|
61 | |||
57 | global pickle |
|
62 | global pickle | |
58 | pickle = dill |
|
63 | pickle = dill | |
59 |
|
64 | |||
60 | try: |
|
65 | try: | |
61 | from IPython.kernel.zmq import serialize |
|
66 | from IPython.kernel.zmq import serialize | |
62 | except ImportError: |
|
67 | except ImportError: | |
63 | pass |
|
68 | pass | |
64 | else: |
|
69 | else: | |
65 | serialize.pickle = dill |
|
70 | serialize.pickle = dill | |
66 |
|
71 | |||
67 | # disable special function handling, let dill take care of it |
|
72 | # disable special function handling, let dill take care of it | |
68 | can_map.pop(FunctionType, None) |
|
73 | can_map.pop(FunctionType, None) | |
69 |
|
74 | |||
70 | def use_cloudpickle(): |
|
75 | def use_cloudpickle(): | |
71 | """use cloudpickle to expand serialization support |
|
76 | """use cloudpickle to expand serialization support | |
72 |
|
77 | |||
73 | adds support for object methods and closures to serialization. |
|
78 | adds support for object methods and closures to serialization. | |
74 | """ |
|
79 | """ | |
75 | from cloud.serialization import cloudpickle |
|
80 | from cloud.serialization import cloudpickle | |
76 |
|
81 | |||
77 | global pickle |
|
82 | global pickle | |
78 | pickle = cloudpickle |
|
83 | pickle = cloudpickle | |
79 |
|
84 | |||
80 | try: |
|
85 | try: | |
81 | from IPython.kernel.zmq import serialize |
|
86 | from IPython.kernel.zmq import serialize | |
82 | except ImportError: |
|
87 | except ImportError: | |
83 | pass |
|
88 | pass | |
84 | else: |
|
89 | else: | |
85 | serialize.pickle = cloudpickle |
|
90 | serialize.pickle = cloudpickle | |
86 |
|
91 | |||
87 | # disable special function handling, let cloudpickle take care of it |
|
92 | # disable special function handling, let cloudpickle take care of it | |
88 | can_map.pop(FunctionType, None) |
|
93 | can_map.pop(FunctionType, None) | |
89 |
|
94 | |||
90 |
|
95 | |||
91 | #------------------------------------------------------------------------------- |
|
96 | #------------------------------------------------------------------------------- | |
92 | # Classes |
|
97 | # Classes | |
93 | #------------------------------------------------------------------------------- |
|
98 | #------------------------------------------------------------------------------- | |
94 |
|
99 | |||
95 |
|
100 | |||
96 | class CannedObject(object): |
|
101 | class CannedObject(object): | |
97 | def __init__(self, obj, keys=[], hook=None): |
|
102 | def __init__(self, obj, keys=[], hook=None): | |
98 | """can an object for safe pickling |
|
103 | """can an object for safe pickling | |
99 |
|
104 | |||
100 | Parameters |
|
105 | Parameters | |
101 | ========== |
|
106 | ========== | |
102 |
|
107 | |||
103 | obj: |
|
108 | obj: | |
104 | The object to be canned |
|
109 | The object to be canned | |
105 | keys: list (optional) |
|
110 | keys: list (optional) | |
106 | list of attribute names that will be explicitly canned / uncanned |
|
111 | list of attribute names that will be explicitly canned / uncanned | |
107 | hook: callable (optional) |
|
112 | hook: callable (optional) | |
108 | An optional extra callable, |
|
113 | An optional extra callable, | |
109 | which can do additional processing of the uncanned object. |
|
114 | which can do additional processing of the uncanned object. | |
110 |
|
115 | |||
111 | large data may be offloaded into the buffers list, |
|
116 | large data may be offloaded into the buffers list, | |
112 | used for zero-copy transfers. |
|
117 | used for zero-copy transfers. | |
113 | """ |
|
118 | """ | |
114 | self.keys = keys |
|
119 | self.keys = keys | |
115 | self.obj = copy.copy(obj) |
|
120 | self.obj = copy.copy(obj) | |
116 | self.hook = can(hook) |
|
121 | self.hook = can(hook) | |
117 | for key in keys: |
|
122 | for key in keys: | |
118 | setattr(self.obj, key, can(getattr(obj, key))) |
|
123 | setattr(self.obj, key, can(getattr(obj, key))) | |
119 |
|
124 | |||
120 | self.buffers = [] |
|
125 | self.buffers = [] | |
121 |
|
126 | |||
122 | def get_object(self, g=None): |
|
127 | def get_object(self, g=None): | |
123 | if g is None: |
|
128 | if g is None: | |
124 | g = {} |
|
129 | g = {} | |
125 | obj = self.obj |
|
130 | obj = self.obj | |
126 | for key in self.keys: |
|
131 | for key in self.keys: | |
127 | setattr(obj, key, uncan(getattr(obj, key), g)) |
|
132 | setattr(obj, key, uncan(getattr(obj, key), g)) | |
128 |
|
133 | |||
129 | if self.hook: |
|
134 | if self.hook: | |
130 | self.hook = uncan(self.hook, g) |
|
135 | self.hook = uncan(self.hook, g) | |
131 | self.hook(obj, g) |
|
136 | self.hook(obj, g) | |
132 | return self.obj |
|
137 | return self.obj | |
133 |
|
138 | |||
134 |
|
139 | |||
135 | class Reference(CannedObject): |
|
140 | class Reference(CannedObject): | |
136 | """object for wrapping a remote reference by name.""" |
|
141 | """object for wrapping a remote reference by name.""" | |
137 | def __init__(self, name): |
|
142 | def __init__(self, name): | |
138 | if not isinstance(name, string_types): |
|
143 | if not isinstance(name, string_types): | |
139 | raise TypeError("illegal name: %r"%name) |
|
144 | raise TypeError("illegal name: %r"%name) | |
140 | self.name = name |
|
145 | self.name = name | |
141 | self.buffers = [] |
|
146 | self.buffers = [] | |
142 |
|
147 | |||
143 | def __repr__(self): |
|
148 | def __repr__(self): | |
144 | return "<Reference: %r>"%self.name |
|
149 | return "<Reference: %r>"%self.name | |
145 |
|
150 | |||
146 | def get_object(self, g=None): |
|
151 | def get_object(self, g=None): | |
147 | if g is None: |
|
152 | if g is None: | |
148 | g = {} |
|
153 | g = {} | |
149 |
|
154 | |||
150 | return eval(self.name, g) |
|
155 | return eval(self.name, g) | |
151 |
|
156 | |||
152 |
|
157 | |||
153 | class CannedCell(CannedObject): |
|
158 | class CannedCell(CannedObject): | |
154 | """Can a closure cell""" |
|
159 | """Can a closure cell""" | |
155 | def __init__(self, cell): |
|
160 | def __init__(self, cell): | |
156 | self.cell_contents = can(cell.cell_contents) |
|
161 | self.cell_contents = can(cell.cell_contents) | |
157 |
|
162 | |||
158 | def get_object(self, g=None): |
|
163 | def get_object(self, g=None): | |
159 | cell_contents = uncan(self.cell_contents, g) |
|
164 | cell_contents = uncan(self.cell_contents, g) | |
160 | def inner(): |
|
165 | def inner(): | |
161 | return cell_contents |
|
166 | return cell_contents | |
162 | return py3compat.get_closure(inner)[0] |
|
167 | return py3compat.get_closure(inner)[0] | |
163 |
|
168 | |||
164 |
|
169 | |||
165 | class CannedFunction(CannedObject): |
|
170 | class CannedFunction(CannedObject): | |
166 |
|
171 | |||
167 | def __init__(self, f): |
|
172 | def __init__(self, f): | |
168 | self._check_type(f) |
|
173 | self._check_type(f) | |
169 | self.code = f.__code__ |
|
174 | self.code = f.__code__ | |
170 | if f.__defaults__: |
|
175 | if f.__defaults__: | |
171 | self.defaults = [ can(fd) for fd in f.__defaults__ ] |
|
176 | self.defaults = [ can(fd) for fd in f.__defaults__ ] | |
172 | else: |
|
177 | else: | |
173 | self.defaults = None |
|
178 | self.defaults = None | |
174 |
|
179 | |||
175 | closure = py3compat.get_closure(f) |
|
180 | closure = py3compat.get_closure(f) | |
176 | if closure: |
|
181 | if closure: | |
177 | self.closure = tuple( can(cell) for cell in closure ) |
|
182 | self.closure = tuple( can(cell) for cell in closure ) | |
178 | else: |
|
183 | else: | |
179 | self.closure = None |
|
184 | self.closure = None | |
180 |
|
185 | |||
181 | self.module = f.__module__ or '__main__' |
|
186 | self.module = f.__module__ or '__main__' | |
182 | self.__name__ = f.__name__ |
|
187 | self.__name__ = f.__name__ | |
183 | self.buffers = [] |
|
188 | self.buffers = [] | |
184 |
|
189 | |||
185 | def _check_type(self, obj): |
|
190 | def _check_type(self, obj): | |
186 | assert isinstance(obj, FunctionType), "Not a function type" |
|
191 | assert isinstance(obj, FunctionType), "Not a function type" | |
187 |
|
192 | |||
188 | def get_object(self, g=None): |
|
193 | def get_object(self, g=None): | |
189 | # try to load function back into its module: |
|
194 | # try to load function back into its module: | |
190 | if not self.module.startswith('__'): |
|
195 | if not self.module.startswith('__'): | |
191 | __import__(self.module) |
|
196 | __import__(self.module) | |
192 | g = sys.modules[self.module].__dict__ |
|
197 | g = sys.modules[self.module].__dict__ | |
193 |
|
198 | |||
194 | if g is None: |
|
199 | if g is None: | |
195 | g = {} |
|
200 | g = {} | |
196 | if self.defaults: |
|
201 | if self.defaults: | |
197 | defaults = tuple(uncan(cfd, g) for cfd in self.defaults) |
|
202 | defaults = tuple(uncan(cfd, g) for cfd in self.defaults) | |
198 | else: |
|
203 | else: | |
199 | defaults = None |
|
204 | defaults = None | |
200 | if self.closure: |
|
205 | if self.closure: | |
201 | closure = tuple(uncan(cell, g) for cell in self.closure) |
|
206 | closure = tuple(uncan(cell, g) for cell in self.closure) | |
202 | else: |
|
207 | else: | |
203 | closure = None |
|
208 | closure = None | |
204 | newFunc = FunctionType(self.code, g, self.__name__, defaults, closure) |
|
209 | newFunc = FunctionType(self.code, g, self.__name__, defaults, closure) | |
205 | return newFunc |
|
210 | return newFunc | |
206 |
|
211 | |||
207 | class CannedClass(CannedObject): |
|
212 | class CannedClass(CannedObject): | |
208 |
|
213 | |||
209 | def __init__(self, cls): |
|
214 | def __init__(self, cls): | |
210 | self._check_type(cls) |
|
215 | self._check_type(cls) | |
211 | self.name = cls.__name__ |
|
216 | self.name = cls.__name__ | |
212 | self.old_style = not isinstance(cls, type) |
|
217 | self.old_style = not isinstance(cls, type) | |
213 | self._canned_dict = {} |
|
218 | self._canned_dict = {} | |
214 | for k,v in cls.__dict__.items(): |
|
219 | for k,v in cls.__dict__.items(): | |
215 | if k not in ('__weakref__', '__dict__'): |
|
220 | if k not in ('__weakref__', '__dict__'): | |
216 | self._canned_dict[k] = can(v) |
|
221 | self._canned_dict[k] = can(v) | |
217 | if self.old_style: |
|
222 | if self.old_style: | |
218 | mro = [] |
|
223 | mro = [] | |
219 | else: |
|
224 | else: | |
220 | mro = cls.mro() |
|
225 | mro = cls.mro() | |
221 |
|
226 | |||
222 | self.parents = [ can(c) for c in mro[1:] ] |
|
227 | self.parents = [ can(c) for c in mro[1:] ] | |
223 | self.buffers = [] |
|
228 | self.buffers = [] | |
224 |
|
229 | |||
225 | def _check_type(self, obj): |
|
230 | def _check_type(self, obj): | |
226 | assert isinstance(obj, class_type), "Not a class type" |
|
231 | assert isinstance(obj, class_type), "Not a class type" | |
227 |
|
232 | |||
228 | def get_object(self, g=None): |
|
233 | def get_object(self, g=None): | |
229 | parents = tuple(uncan(p, g) for p in self.parents) |
|
234 | parents = tuple(uncan(p, g) for p in self.parents) | |
230 | return type(self.name, parents, uncan_dict(self._canned_dict, g=g)) |
|
235 | return type(self.name, parents, uncan_dict(self._canned_dict, g=g)) | |
231 |
|
236 | |||
232 | class CannedArray(CannedObject): |
|
237 | class CannedArray(CannedObject): | |
233 | def __init__(self, obj): |
|
238 | def __init__(self, obj): | |
234 | from numpy import ascontiguousarray |
|
239 | from numpy import ascontiguousarray | |
235 | self.shape = obj.shape |
|
240 | self.shape = obj.shape | |
236 | self.dtype = obj.dtype.descr if obj.dtype.fields else obj.dtype.str |
|
241 | self.dtype = obj.dtype.descr if obj.dtype.fields else obj.dtype.str | |
237 | self.pickled = False |
|
242 | self.pickled = False | |
238 | if sum(obj.shape) == 0: |
|
243 | if sum(obj.shape) == 0: | |
239 | self.pickled = True |
|
244 | self.pickled = True | |
240 | elif obj.dtype == 'O': |
|
245 | elif obj.dtype == 'O': | |
241 | # can't handle object dtype with buffer approach |
|
246 | # can't handle object dtype with buffer approach | |
242 | self.pickled = True |
|
247 | self.pickled = True | |
243 | elif obj.dtype.fields and any(dt == 'O' for dt,sz in obj.dtype.fields.values()): |
|
248 | elif obj.dtype.fields and any(dt == 'O' for dt,sz in obj.dtype.fields.values()): | |
244 | self.pickled = True |
|
249 | self.pickled = True | |
245 | if self.pickled: |
|
250 | if self.pickled: | |
246 | # just pickle it |
|
251 | # just pickle it | |
247 |
self.buffers = [pickle.dumps(obj, |
|
252 | self.buffers = [pickle.dumps(obj, PICKLE_PROTOCOL)] | |
248 | else: |
|
253 | else: | |
249 | # ensure contiguous |
|
254 | # ensure contiguous | |
250 | obj = ascontiguousarray(obj, dtype=None) |
|
255 | obj = ascontiguousarray(obj, dtype=None) | |
251 | self.buffers = [buffer(obj)] |
|
256 | self.buffers = [buffer(obj)] | |
252 |
|
257 | |||
253 | def get_object(self, g=None): |
|
258 | def get_object(self, g=None): | |
254 | from numpy import frombuffer |
|
259 | from numpy import frombuffer | |
255 | data = self.buffers[0] |
|
260 | data = self.buffers[0] | |
256 | if self.pickled: |
|
261 | if self.pickled: | |
257 | # no shape, we just pickled it |
|
262 | # no shape, we just pickled it | |
258 | return pickle.loads(data) |
|
263 | return pickle.loads(data) | |
259 | else: |
|
264 | else: | |
260 | return frombuffer(data, dtype=self.dtype).reshape(self.shape) |
|
265 | return frombuffer(data, dtype=self.dtype).reshape(self.shape) | |
261 |
|
266 | |||
262 |
|
267 | |||
263 | class CannedBytes(CannedObject): |
|
268 | class CannedBytes(CannedObject): | |
264 | wrap = bytes |
|
269 | wrap = bytes | |
265 | def __init__(self, obj): |
|
270 | def __init__(self, obj): | |
266 | self.buffers = [obj] |
|
271 | self.buffers = [obj] | |
267 |
|
272 | |||
268 | def get_object(self, g=None): |
|
273 | def get_object(self, g=None): | |
269 | data = self.buffers[0] |
|
274 | data = self.buffers[0] | |
270 | return self.wrap(data) |
|
275 | return self.wrap(data) | |
271 |
|
276 | |||
272 | def CannedBuffer(CannedBytes): |
|
277 | def CannedBuffer(CannedBytes): | |
273 | wrap = buffer |
|
278 | wrap = buffer | |
274 |
|
279 | |||
275 | #------------------------------------------------------------------------------- |
|
280 | #------------------------------------------------------------------------------- | |
276 | # Functions |
|
281 | # Functions | |
277 | #------------------------------------------------------------------------------- |
|
282 | #------------------------------------------------------------------------------- | |
278 |
|
283 | |||
279 | def _logger(): |
|
284 | def _logger(): | |
280 | """get the logger for the current Application |
|
285 | """get the logger for the current Application | |
281 |
|
286 | |||
282 | the root logger will be used if no Application is running |
|
287 | the root logger will be used if no Application is running | |
283 | """ |
|
288 | """ | |
284 | if Application.initialized(): |
|
289 | if Application.initialized(): | |
285 | logger = Application.instance().log |
|
290 | logger = Application.instance().log | |
286 | else: |
|
291 | else: | |
287 | logger = logging.getLogger() |
|
292 | logger = logging.getLogger() | |
288 | if not logger.handlers: |
|
293 | if not logger.handlers: | |
289 | logging.basicConfig() |
|
294 | logging.basicConfig() | |
290 |
|
295 | |||
291 | return logger |
|
296 | return logger | |
292 |
|
297 | |||
293 | def _import_mapping(mapping, original=None): |
|
298 | def _import_mapping(mapping, original=None): | |
294 | """import any string-keys in a type mapping |
|
299 | """import any string-keys in a type mapping | |
295 |
|
300 | |||
296 | """ |
|
301 | """ | |
297 | log = _logger() |
|
302 | log = _logger() | |
298 | log.debug("Importing canning map") |
|
303 | log.debug("Importing canning map") | |
299 | for key,value in list(mapping.items()): |
|
304 | for key,value in list(mapping.items()): | |
300 | if isinstance(key, string_types): |
|
305 | if isinstance(key, string_types): | |
301 | try: |
|
306 | try: | |
302 | cls = import_item(key) |
|
307 | cls = import_item(key) | |
303 | except Exception: |
|
308 | except Exception: | |
304 | if original and key not in original: |
|
309 | if original and key not in original: | |
305 | # only message on user-added classes |
|
310 | # only message on user-added classes | |
306 | log.error("canning class not importable: %r", key, exc_info=True) |
|
311 | log.error("canning class not importable: %r", key, exc_info=True) | |
307 | mapping.pop(key) |
|
312 | mapping.pop(key) | |
308 | else: |
|
313 | else: | |
309 | mapping[cls] = mapping.pop(key) |
|
314 | mapping[cls] = mapping.pop(key) | |
310 |
|
315 | |||
311 | def istype(obj, check): |
|
316 | def istype(obj, check): | |
312 | """like isinstance(obj, check), but strict |
|
317 | """like isinstance(obj, check), but strict | |
313 |
|
318 | |||
314 | This won't catch subclasses. |
|
319 | This won't catch subclasses. | |
315 | """ |
|
320 | """ | |
316 | if isinstance(check, tuple): |
|
321 | if isinstance(check, tuple): | |
317 | for cls in check: |
|
322 | for cls in check: | |
318 | if type(obj) is cls: |
|
323 | if type(obj) is cls: | |
319 | return True |
|
324 | return True | |
320 | return False |
|
325 | return False | |
321 | else: |
|
326 | else: | |
322 | return type(obj) is check |
|
327 | return type(obj) is check | |
323 |
|
328 | |||
324 | def can(obj): |
|
329 | def can(obj): | |
325 | """prepare an object for pickling""" |
|
330 | """prepare an object for pickling""" | |
326 |
|
331 | |||
327 | import_needed = False |
|
332 | import_needed = False | |
328 |
|
333 | |||
329 | for cls,canner in iteritems(can_map): |
|
334 | for cls,canner in iteritems(can_map): | |
330 | if isinstance(cls, string_types): |
|
335 | if isinstance(cls, string_types): | |
331 | import_needed = True |
|
336 | import_needed = True | |
332 | break |
|
337 | break | |
333 | elif istype(obj, cls): |
|
338 | elif istype(obj, cls): | |
334 | return canner(obj) |
|
339 | return canner(obj) | |
335 |
|
340 | |||
336 | if import_needed: |
|
341 | if import_needed: | |
337 | # perform can_map imports, then try again |
|
342 | # perform can_map imports, then try again | |
338 | # this will usually only happen once |
|
343 | # this will usually only happen once | |
339 | _import_mapping(can_map, _original_can_map) |
|
344 | _import_mapping(can_map, _original_can_map) | |
340 | return can(obj) |
|
345 | return can(obj) | |
341 |
|
346 | |||
342 | return obj |
|
347 | return obj | |
343 |
|
348 | |||
344 | def can_class(obj): |
|
349 | def can_class(obj): | |
345 | if isinstance(obj, class_type) and obj.__module__ == '__main__': |
|
350 | if isinstance(obj, class_type) and obj.__module__ == '__main__': | |
346 | return CannedClass(obj) |
|
351 | return CannedClass(obj) | |
347 | else: |
|
352 | else: | |
348 | return obj |
|
353 | return obj | |
349 |
|
354 | |||
350 | def can_dict(obj): |
|
355 | def can_dict(obj): | |
351 | """can the *values* of a dict""" |
|
356 | """can the *values* of a dict""" | |
352 | if istype(obj, dict): |
|
357 | if istype(obj, dict): | |
353 | newobj = {} |
|
358 | newobj = {} | |
354 | for k, v in iteritems(obj): |
|
359 | for k, v in iteritems(obj): | |
355 | newobj[k] = can(v) |
|
360 | newobj[k] = can(v) | |
356 | return newobj |
|
361 | return newobj | |
357 | else: |
|
362 | else: | |
358 | return obj |
|
363 | return obj | |
359 |
|
364 | |||
360 | sequence_types = (list, tuple, set) |
|
365 | sequence_types = (list, tuple, set) | |
361 |
|
366 | |||
362 | def can_sequence(obj): |
|
367 | def can_sequence(obj): | |
363 | """can the elements of a sequence""" |
|
368 | """can the elements of a sequence""" | |
364 | if istype(obj, sequence_types): |
|
369 | if istype(obj, sequence_types): | |
365 | t = type(obj) |
|
370 | t = type(obj) | |
366 | return t([can(i) for i in obj]) |
|
371 | return t([can(i) for i in obj]) | |
367 | else: |
|
372 | else: | |
368 | return obj |
|
373 | return obj | |
369 |
|
374 | |||
370 | def uncan(obj, g=None): |
|
375 | def uncan(obj, g=None): | |
371 | """invert canning""" |
|
376 | """invert canning""" | |
372 |
|
377 | |||
373 | import_needed = False |
|
378 | import_needed = False | |
374 | for cls,uncanner in iteritems(uncan_map): |
|
379 | for cls,uncanner in iteritems(uncan_map): | |
375 | if isinstance(cls, string_types): |
|
380 | if isinstance(cls, string_types): | |
376 | import_needed = True |
|
381 | import_needed = True | |
377 | break |
|
382 | break | |
378 | elif isinstance(obj, cls): |
|
383 | elif isinstance(obj, cls): | |
379 | return uncanner(obj, g) |
|
384 | return uncanner(obj, g) | |
380 |
|
385 | |||
381 | if import_needed: |
|
386 | if import_needed: | |
382 | # perform uncan_map imports, then try again |
|
387 | # perform uncan_map imports, then try again | |
383 | # this will usually only happen once |
|
388 | # this will usually only happen once | |
384 | _import_mapping(uncan_map, _original_uncan_map) |
|
389 | _import_mapping(uncan_map, _original_uncan_map) | |
385 | return uncan(obj, g) |
|
390 | return uncan(obj, g) | |
386 |
|
391 | |||
387 | return obj |
|
392 | return obj | |
388 |
|
393 | |||
389 | def uncan_dict(obj, g=None): |
|
394 | def uncan_dict(obj, g=None): | |
390 | if istype(obj, dict): |
|
395 | if istype(obj, dict): | |
391 | newobj = {} |
|
396 | newobj = {} | |
392 | for k, v in iteritems(obj): |
|
397 | for k, v in iteritems(obj): | |
393 | newobj[k] = uncan(v,g) |
|
398 | newobj[k] = uncan(v,g) | |
394 | return newobj |
|
399 | return newobj | |
395 | else: |
|
400 | else: | |
396 | return obj |
|
401 | return obj | |
397 |
|
402 | |||
398 | def uncan_sequence(obj, g=None): |
|
403 | def uncan_sequence(obj, g=None): | |
399 | if istype(obj, sequence_types): |
|
404 | if istype(obj, sequence_types): | |
400 | t = type(obj) |
|
405 | t = type(obj) | |
401 | return t([uncan(i,g) for i in obj]) |
|
406 | return t([uncan(i,g) for i in obj]) | |
402 | else: |
|
407 | else: | |
403 | return obj |
|
408 | return obj | |
404 |
|
409 | |||
405 | def _uncan_dependent_hook(dep, g=None): |
|
410 | def _uncan_dependent_hook(dep, g=None): | |
406 | dep.check_dependency() |
|
411 | dep.check_dependency() | |
407 |
|
412 | |||
408 | def can_dependent(obj): |
|
413 | def can_dependent(obj): | |
409 | return CannedObject(obj, keys=('f', 'df'), hook=_uncan_dependent_hook) |
|
414 | return CannedObject(obj, keys=('f', 'df'), hook=_uncan_dependent_hook) | |
410 |
|
415 | |||
411 | #------------------------------------------------------------------------------- |
|
416 | #------------------------------------------------------------------------------- | |
412 | # API dictionaries |
|
417 | # API dictionaries | |
413 | #------------------------------------------------------------------------------- |
|
418 | #------------------------------------------------------------------------------- | |
414 |
|
419 | |||
415 | # These dicts can be extended for custom serialization of new objects |
|
420 | # These dicts can be extended for custom serialization of new objects | |
416 |
|
421 | |||
417 | can_map = { |
|
422 | can_map = { | |
418 | 'IPython.parallel.dependent' : can_dependent, |
|
423 | 'IPython.parallel.dependent' : can_dependent, | |
419 | 'numpy.ndarray' : CannedArray, |
|
424 | 'numpy.ndarray' : CannedArray, | |
420 | FunctionType : CannedFunction, |
|
425 | FunctionType : CannedFunction, | |
421 | bytes : CannedBytes, |
|
426 | bytes : CannedBytes, | |
422 | buffer : CannedBuffer, |
|
427 | buffer : CannedBuffer, | |
423 | cell_type : CannedCell, |
|
428 | cell_type : CannedCell, | |
424 | class_type : can_class, |
|
429 | class_type : can_class, | |
425 | } |
|
430 | } | |
426 |
|
431 | |||
427 | uncan_map = { |
|
432 | uncan_map = { | |
428 | CannedObject : lambda obj, g: obj.get_object(g), |
|
433 | CannedObject : lambda obj, g: obj.get_object(g), | |
429 | } |
|
434 | } | |
430 |
|
435 | |||
431 | # for use in _import_mapping: |
|
436 | # for use in _import_mapping: | |
432 | _original_can_map = can_map.copy() |
|
437 | _original_can_map = can_map.copy() | |
433 | _original_uncan_map = uncan_map.copy() |
|
438 | _original_uncan_map = uncan_map.copy() |
General Comments 0
You need to be logged in to leave comments.
Login now