1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
|
#! /usr/bin/env python3
# A database storing dictionaries, keyed on a timestamp.
# key = 8 octets timestamp | 1 octet version
# struct.pack('!dB', time.time(), 0)
# value = A dict which will be stored as a JSON object encoded in
# UTF-8. Note that dict keys of type integer or float will become
# strings while values will keep their type.
# Note that there's a (slim) chance that you'd stomp on the previous
# value if you're too quick with generating the timestamps, ie
# invoking time.time() several times quickly enough.
from store import KVStore
import json
from pathlib import PurePath
import time
import struct
import os
class DictDB(KVStore):
VERSION = 0
# TODO: implement indexes
# TODO: implement search(dict key) for indexed fields
def __init__(self, name, basedir='.'):
super().__init__(name, basedir)
self._ts = time.time()
self._index = {}
def unique_key(self):
ts = time.time()
while ts == self._ts:
ts = time.time()
self._ts = ts
return struct.pack('!dB', ts, DictDB.VERSION)
def index_add(self, path):
name = PurePath(path).name
self._index[name] = DBIndex(path)
def add(self, data, batch_write=False):
if type(data) is list:
ret = []
if batch_write: # Supposedly makes the update atomic.
batch = self.batch()
for e in data:
ret += [self.add(e)]
if batch_write:
self.write(batch)
return ret
else:
key = self.unique_key()
json_data = json.dumps(data).encode('UTF-8')
self.put(key, json_data)
return key
def get(self, key):
enc = super().get(key) # name collision, specify super class
ret = json.loads(enc.decode('UTF-8'))
return ret
def slice(self, key_from=None, key_to=None):
ret = []
for key_raw, val_raw in list(self.range_iter(key_from=key_from, key_to=key_to)):
(key, ver) = struct.unpack('!dB', key_raw)
if ver == DictDB.VERSION:
val = json.loads(val_raw.decode('UTF-8'))
ret.append((key, val))
return ret
# Search top level dict for objects with a name matching DICT_NAME
# and optionally value DICT_VAL.
def search(self, dict_name, dict_val=None):
res = []
for key, obj in self.slice():
if dict_name in obj:
if dict_val is None:
res.append((key, obj))
elif dict_val == obj[dict_name]:
res.append((key, obj))
return res
def timestamp_from_key(key):
ts, _ = struct.unpack('!dB', key)
return ts
class DBIndex(KVStore):
def __init__(self, name, basedir='.'):
super().__init__(name, basedir)
# def update_or_create(self, key, val):
# curval = []
# try:
# curval = self._db.Get(key)
# except KeyError:
# self._db.Put(key, [val])
# return
# if curval is list:
# self._db.Put(key, curval + [val])
# else:
# self._db.Put(key, json.dumps([curval, val]))
# def index_full(self, name):
# kv = self._index[name])
# for key, val in self.range_iter():
# valdict = json.loads(val)
# field = valdict.get(name)
# if field:
# ix = kv.get(key)
# if ix:
# kv.put(ix + [key])
# else:
# kv.put([key])
from operator import itemgetter
def dict_eq(a, b):
sort_on_key = lambda d: sorted(d.items(), key=itemgetter(0))
return sort_on_key(a) == sort_on_key(b)
if __name__ == '__main__':
DBDIR = 'test_db.db'
# TODO: rm -r DBDIR
db = DictDB('db', basedir = DBDIR)
#ix = DBIndex('foo', basedir = DBDIR)
key = db.add({'foo': 'Bar'})
assert(db.get(key) == {'foo': 'Bar'})
key = db.add({0: 'Foo'}) # NOTE: int keys become strings
assert(db.get(key) == {'0': 'Foo'})
d = {'4711': 'Large number', '7': 'Small number', '0': 'Bar'}
key = db.add(d)
res = db.get(key)
assert(dict_eq(d, res))
key = db.add({'an int': 0})
assert(db.get(key) == {'an int': 0})
key = db.add({'a float': 1.1})
assert(db.get(key) == {'a float': 1.1})
# TODO: verify slice() too
for key, val in db.slice():
print(key, val)
res = db.search('an int')
assert(dict_eq(res[0][1], {'an int': 0}))
res = db.search('0')
assert(dict_eq(res[0][1], {'0': 'Foo'}))
assert(dict_eq(res[1][1], d))
res = db.search('7', dict_val = 'Small number') # FIXME: verify better -- do we hit only '7' here, f.ex.?
assert('7' in res[0][1])
res = db.search('7', dict_val = 'not matching')
assert(not res)
N = 10 * 1000 # 10k takes ~0.2s.
data = [{str(x): x} for x in range(N)]
keys = db.add(data, batch_write = False)
assert(len(keys) == N)
for k in range(len(keys)):
assert(db.get(keys[k]) == data[k])
|
