root / snf-pithos-backend / pithos / backends / lib / sqlalchemy / node.py @ 63092950
History | View | Annotate | Download (53.1 kB)
| 1 |
# Copyright 2011-2012 GRNET S.A. All rights reserved.
|
|---|---|
| 2 |
#
|
| 3 |
# Redistribution and use in source and binary forms, with or
|
| 4 |
# without modification, are permitted provided that the following
|
| 5 |
# conditions are met:
|
| 6 |
#
|
| 7 |
# 1. Redistributions of source code must retain the above
|
| 8 |
# copyright notice, this list of conditions and the following
|
| 9 |
# disclaimer.
|
| 10 |
#
|
| 11 |
# 2. Redistributions in binary form must reproduce the above
|
| 12 |
# copyright notice, this list of conditions and the following
|
| 13 |
# disclaimer in the documentation and/or other materials
|
| 14 |
# provided with the distribution.
|
| 15 |
#
|
| 16 |
# THIS SOFTWARE IS PROVIDED BY GRNET S.A. ``AS IS'' AND ANY EXPRESS
|
| 17 |
# OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
|
| 18 |
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
|
| 19 |
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL GRNET S.A OR
|
| 20 |
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
| 21 |
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
| 22 |
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
|
| 23 |
# USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
|
| 24 |
# AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
| 25 |
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
|
| 26 |
# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
| 27 |
# POSSIBILITY OF SUCH DAMAGE.
|
| 28 |
#
|
| 29 |
# The views and conclusions contained in the software and
|
| 30 |
# documentation are those of the authors and should not be
|
| 31 |
# interpreted as representing official policies, either expressed
|
| 32 |
# or implied, of GRNET S.A.
|
| 33 |
|
| 34 |
from time import time |
| 35 |
from operator import itemgetter |
| 36 |
from itertools import groupby |
| 37 |
|
| 38 |
from sqlalchemy import (Table, Integer, BigInteger, DECIMAL, Boolean, |
| 39 |
Column, String, MetaData, ForeignKey) |
| 40 |
from sqlalchemy.schema import Index |
| 41 |
from sqlalchemy.sql import func, and_, or_, not_, select, bindparam, exists |
| 42 |
from sqlalchemy.sql.expression import true |
| 43 |
from sqlalchemy.exc import NoSuchTableError |
| 44 |
|
| 45 |
from dbworker import DBWorker, ESCAPE_CHAR |
| 46 |
|
| 47 |
from pithos.backends.filter import parse_filters |
| 48 |
|
| 49 |
|
| 50 |
ROOTNODE = 0
|
| 51 |
|
| 52 |
(SERIAL, NODE, HASH, SIZE, TYPE, SOURCE, MTIME, MUSER, UUID, CHECKSUM, |
| 53 |
CLUSTER) = range(11) |
| 54 |
|
| 55 |
(MATCH_PREFIX, MATCH_EXACT) = range(2) |
| 56 |
|
| 57 |
inf = float('inf') |
| 58 |
|
| 59 |
|
| 60 |
def strnextling(prefix): |
| 61 |
"""Return the first unicode string
|
| 62 |
greater than but not starting with given prefix.
|
| 63 |
strnextling('hello') -> 'hellp'
|
| 64 |
"""
|
| 65 |
if not prefix: |
| 66 |
## all strings start with the null string,
|
| 67 |
## therefore we have to approximate strnextling('')
|
| 68 |
## with the last unicode character supported by python
|
| 69 |
## 0x10ffff for wide (32-bit unicode) python builds
|
| 70 |
## 0x00ffff for narrow (16-bit unicode) python builds
|
| 71 |
## We will not autodetect. 0xffff is safe enough.
|
| 72 |
return unichr(0xffff) |
| 73 |
s = prefix[:-1]
|
| 74 |
c = ord(prefix[-1]) |
| 75 |
if c >= 0xffff: |
| 76 |
raise RuntimeError |
| 77 |
s += unichr(c + 1) |
| 78 |
return s
|
| 79 |
|
| 80 |
|
| 81 |
def strprevling(prefix): |
| 82 |
"""Return an approximation of the last unicode string
|
| 83 |
less than but not starting with given prefix.
|
| 84 |
strprevling(u'hello') -> u'helln\\xffff'
|
| 85 |
"""
|
| 86 |
if not prefix: |
| 87 |
## There is no prevling for the null string
|
| 88 |
return prefix
|
| 89 |
s = prefix[:-1]
|
| 90 |
c = ord(prefix[-1]) |
| 91 |
if c > 0: |
| 92 |
s += unichr(c - 1) + unichr(0xffff) |
| 93 |
return s
|
| 94 |
|
| 95 |
_propnames = {
|
| 96 |
'serial': 0, |
| 97 |
'node': 1, |
| 98 |
'hash': 2, |
| 99 |
'size': 3, |
| 100 |
'type': 4, |
| 101 |
'source': 5, |
| 102 |
'mtime': 6, |
| 103 |
'muser': 7, |
| 104 |
'uuid': 8, |
| 105 |
'checksum': 9, |
| 106 |
'cluster': 10, |
| 107 |
} |
| 108 |
|
| 109 |
|
| 110 |
def create_tables(engine): |
| 111 |
metadata = MetaData() |
| 112 |
|
| 113 |
#create nodes table
|
| 114 |
columns = [] |
| 115 |
columns.append(Column('node', Integer, primary_key=True)) |
| 116 |
columns.append(Column('parent', Integer,
|
| 117 |
ForeignKey('nodes.node',
|
| 118 |
ondelete='CASCADE',
|
| 119 |
onupdate='CASCADE'),
|
| 120 |
autoincrement=False))
|
| 121 |
columns.append(Column('latest_version', Integer))
|
| 122 |
columns.append(Column('path', String(2048), default='', nullable=False)) |
| 123 |
nodes = Table('nodes', metadata, *columns, mysql_engine='InnoDB') |
| 124 |
Index('idx_nodes_path', nodes.c.path, unique=True) |
| 125 |
Index('idx_nodes_parent', nodes.c.parent)
|
| 126 |
|
| 127 |
#create policy table
|
| 128 |
columns = [] |
| 129 |
columns.append(Column('node', Integer,
|
| 130 |
ForeignKey('nodes.node',
|
| 131 |
ondelete='CASCADE',
|
| 132 |
onupdate='CASCADE'),
|
| 133 |
primary_key=True))
|
| 134 |
columns.append(Column('key', String(128), primary_key=True)) |
| 135 |
columns.append(Column('value', String(256))) |
| 136 |
Table('policy', metadata, *columns, mysql_engine='InnoDB') |
| 137 |
|
| 138 |
#create statistics table
|
| 139 |
columns = [] |
| 140 |
columns.append(Column('node', Integer,
|
| 141 |
ForeignKey('nodes.node',
|
| 142 |
ondelete='CASCADE',
|
| 143 |
onupdate='CASCADE'),
|
| 144 |
primary_key=True))
|
| 145 |
columns.append(Column('population', Integer, nullable=False, default=0)) |
| 146 |
columns.append(Column('size', BigInteger, nullable=False, default=0)) |
| 147 |
columns.append(Column('mtime', DECIMAL(precision=16, scale=6))) |
| 148 |
columns.append(Column('cluster', Integer, nullable=False, default=0, |
| 149 |
primary_key=True, autoincrement=False)) |
| 150 |
Table('statistics', metadata, *columns, mysql_engine='InnoDB') |
| 151 |
|
| 152 |
#create versions table
|
| 153 |
columns = [] |
| 154 |
columns.append(Column('serial', Integer, primary_key=True)) |
| 155 |
columns.append(Column('node', Integer,
|
| 156 |
ForeignKey('nodes.node',
|
| 157 |
ondelete='CASCADE',
|
| 158 |
onupdate='CASCADE')))
|
| 159 |
columns.append(Column('hash', String(256))) |
| 160 |
columns.append(Column('size', BigInteger, nullable=False, default=0)) |
| 161 |
columns.append(Column('type', String(256), nullable=False, default='')) |
| 162 |
columns.append(Column('source', Integer))
|
| 163 |
columns.append(Column('mtime', DECIMAL(precision=16, scale=6))) |
| 164 |
columns.append(Column('muser', String(256), nullable=False, default='')) |
| 165 |
columns.append(Column('uuid', String(64), nullable=False, default='')) |
| 166 |
columns.append(Column('checksum', String(256), nullable=False, default='')) |
| 167 |
columns.append(Column('cluster', Integer, nullable=False, default=0)) |
| 168 |
versions = Table('versions', metadata, *columns, mysql_engine='InnoDB') |
| 169 |
Index('idx_versions_node_mtime', versions.c.node, versions.c.mtime)
|
| 170 |
Index('idx_versions_node_uuid', versions.c.uuid)
|
| 171 |
|
| 172 |
#create attributes table
|
| 173 |
columns = [] |
| 174 |
columns.append(Column('serial', Integer,
|
| 175 |
ForeignKey('versions.serial',
|
| 176 |
ondelete='CASCADE',
|
| 177 |
onupdate='CASCADE'),
|
| 178 |
primary_key=True))
|
| 179 |
columns.append(Column('domain', String(256), primary_key=True)) |
| 180 |
columns.append(Column('key', String(128), primary_key=True)) |
| 181 |
columns.append(Column('value', String(256))) |
| 182 |
columns.append(Column('node', Integer, nullable=False, default=0)) |
| 183 |
columns.append(Column('is_latest', Boolean, nullable=False, default=True)) |
| 184 |
attributes = Table('attributes', metadata, *columns, mysql_engine='InnoDB') |
| 185 |
Index('idx_attributes_domain', attributes.c.domain)
|
| 186 |
Index('idx_attributes_serial_node', attributes.c.serial, attributes.c.node)
|
| 187 |
|
| 188 |
metadata.create_all(engine) |
| 189 |
return metadata.sorted_tables
|
| 190 |
|
| 191 |
|
| 192 |
class Node(DBWorker): |
| 193 |
"""Nodes store path organization and have multiple versions.
|
| 194 |
Versions store object history and have multiple attributes.
|
| 195 |
Attributes store metadata.
|
| 196 |
"""
|
| 197 |
|
| 198 |
# TODO: Provide an interface for included and excluded clusters.
|
| 199 |
|
| 200 |
def __init__(self, **params): |
| 201 |
DBWorker.__init__(self, **params)
|
| 202 |
try:
|
| 203 |
metadata = MetaData(self.engine)
|
| 204 |
self.nodes = Table('nodes', metadata, autoload=True) |
| 205 |
self.policy = Table('policy', metadata, autoload=True) |
| 206 |
self.statistics = Table('statistics', metadata, autoload=True) |
| 207 |
self.versions = Table('versions', metadata, autoload=True) |
| 208 |
self.attributes = Table('attributes', metadata, autoload=True) |
| 209 |
except NoSuchTableError:
|
| 210 |
tables = create_tables(self.engine)
|
| 211 |
map(lambda t: self.__setattr__(t.name, t), tables) |
| 212 |
|
| 213 |
s = self.nodes.select().where(and_(self.nodes.c.node == ROOTNODE, |
| 214 |
self.nodes.c.parent == ROOTNODE))
|
| 215 |
wrapper = self.wrapper
|
| 216 |
wrapper.execute() |
| 217 |
try:
|
| 218 |
rp = self.conn.execute(s)
|
| 219 |
r = rp.fetchone() |
| 220 |
rp.close() |
| 221 |
if not r: |
| 222 |
s = self.nodes.insert(
|
| 223 |
).values(node=ROOTNODE, parent=ROOTNODE, path='')
|
| 224 |
self.conn.execute(s)
|
| 225 |
finally:
|
| 226 |
wrapper.commit() |
| 227 |
|
| 228 |
def node_create(self, parent, path): |
| 229 |
"""Create a new node from the given properties.
|
| 230 |
Return the node identifier of the new node.
|
| 231 |
"""
|
| 232 |
#TODO catch IntegrityError?
|
| 233 |
s = self.nodes.insert().values(parent=parent, path=path)
|
| 234 |
r = self.conn.execute(s)
|
| 235 |
inserted_primary_key = r.inserted_primary_key[0]
|
| 236 |
r.close() |
| 237 |
return inserted_primary_key
|
| 238 |
|
| 239 |
def node_lookup(self, path, for_update=False): |
| 240 |
"""Lookup the current node of the given path.
|
| 241 |
Return None if the path is not found.
|
| 242 |
"""
|
| 243 |
|
| 244 |
# Use LIKE for comparison to avoid MySQL problems with trailing spaces.
|
| 245 |
s = select([self.nodes.c.node],
|
| 246 |
self.nodes.c.path.like(self.escape_like(path), |
| 247 |
escape=ESCAPE_CHAR), |
| 248 |
for_update=for_update) |
| 249 |
r = self.conn.execute(s)
|
| 250 |
row = r.fetchone() |
| 251 |
r.close() |
| 252 |
if row:
|
| 253 |
return row[0] |
| 254 |
return None |
| 255 |
|
| 256 |
def node_lookup_bulk(self, paths): |
| 257 |
"""Lookup the current nodes for the given paths.
|
| 258 |
Return {} if the path is not found.
|
| 259 |
"""
|
| 260 |
|
| 261 |
if not paths: |
| 262 |
return {}
|
| 263 |
# Use LIKE for comparison to avoid MySQL problems with trailing spaces.
|
| 264 |
s = select([self.nodes.c.path, self.nodes.c.node], |
| 265 |
self.nodes.c.path.in_(paths))
|
| 266 |
r = self.conn.execute(s)
|
| 267 |
rows = r.fetchall() |
| 268 |
r.close() |
| 269 |
return dict([(row.node, row.path) for row in rows]) |
| 270 |
|
| 271 |
def node_get_properties(self, node): |
| 272 |
"""Return the node's (parent, path).
|
| 273 |
Return None if the node is not found.
|
| 274 |
"""
|
| 275 |
|
| 276 |
s = select([self.nodes.c.parent, self.nodes.c.path]) |
| 277 |
s = s.where(self.nodes.c.node == node)
|
| 278 |
r = self.conn.execute(s)
|
| 279 |
l = r.fetchone() |
| 280 |
r.close() |
| 281 |
return l
|
| 282 |
|
| 283 |
def node_get_properties_bulk(self, nodes): |
| 284 |
"""Return the (parent, path) for the specific nodes.
|
| 285 |
"""
|
| 286 |
|
| 287 |
s = select([self.nodes.c.node, self.nodes.c.parent, self.nodes.c.path]) |
| 288 |
s = s.where(self.nodes.c.node.in_(nodes))
|
| 289 |
r = self.conn.execute(s)
|
| 290 |
l = r.fetchall() |
| 291 |
r.close() |
| 292 |
return l
|
| 293 |
|
| 294 |
def node_get_versions(self, node, keys=(), propnames=_propnames): |
| 295 |
"""Return the properties of all versions at node.
|
| 296 |
If keys is empty, return all properties in the order
|
| 297 |
(serial, node, hash, size, type, source, mtime, muser, uuid,
|
| 298 |
checksum, cluster).
|
| 299 |
"""
|
| 300 |
|
| 301 |
s = select([self.versions.c.serial,
|
| 302 |
self.versions.c.node,
|
| 303 |
self.versions.c.hash,
|
| 304 |
self.versions.c.size,
|
| 305 |
self.versions.c.type,
|
| 306 |
self.versions.c.source,
|
| 307 |
self.versions.c.mtime,
|
| 308 |
self.versions.c.muser,
|
| 309 |
self.versions.c.uuid,
|
| 310 |
self.versions.c.checksum,
|
| 311 |
self.versions.c.cluster], self.versions.c.node == node) |
| 312 |
s = s.order_by(self.versions.c.serial)
|
| 313 |
r = self.conn.execute(s)
|
| 314 |
rows = r.fetchall() |
| 315 |
r.close() |
| 316 |
if not rows: |
| 317 |
return rows
|
| 318 |
|
| 319 |
if not keys: |
| 320 |
return rows
|
| 321 |
|
| 322 |
return [[p[propnames[k]] for k in keys if k in propnames] for |
| 323 |
p in rows]
|
| 324 |
|
| 325 |
def node_count_children(self, node): |
| 326 |
"""Return node's child count."""
|
| 327 |
|
| 328 |
s = select([func.count(self.nodes.c.node)])
|
| 329 |
s = s.where(and_(self.nodes.c.parent == node,
|
| 330 |
self.nodes.c.node != ROOTNODE))
|
| 331 |
r = self.conn.execute(s)
|
| 332 |
row = r.fetchone() |
| 333 |
r.close() |
| 334 |
return row[0] |
| 335 |
|
| 336 |
def node_purge_children(self, parent, before=inf, cluster=0, |
| 337 |
update_statistics_ancestors_depth=None):
|
| 338 |
"""Delete all versions with the specified
|
| 339 |
parent and cluster, and return
|
| 340 |
the hashes, the total size and the serials of versions deleted.
|
| 341 |
Clears out nodes with no remaining versions.
|
| 342 |
"""
|
| 343 |
#update statistics
|
| 344 |
c1 = select([self.nodes.c.node],
|
| 345 |
self.nodes.c.parent == parent)
|
| 346 |
where_clause = and_(self.versions.c.node.in_(c1),
|
| 347 |
self.versions.c.cluster == cluster)
|
| 348 |
if before != inf:
|
| 349 |
where_clause = and_(where_clause, |
| 350 |
self.versions.c.mtime <= before)
|
| 351 |
s = select([func.count(self.versions.c.serial),
|
| 352 |
func.sum(self.versions.c.size)])
|
| 353 |
s = s.where(where_clause) |
| 354 |
r = self.conn.execute(s)
|
| 355 |
row = r.fetchone() |
| 356 |
r.close() |
| 357 |
if not row: |
| 358 |
return (), 0, () |
| 359 |
nr, size = row[0], row[1] if row[1] else 0 |
| 360 |
mtime = time() |
| 361 |
self.statistics_update(parent, -nr, -size, mtime, cluster)
|
| 362 |
self.statistics_update_ancestors(parent, -nr, -size, mtime, cluster,
|
| 363 |
update_statistics_ancestors_depth) |
| 364 |
|
| 365 |
s = select([self.versions.c.hash, self.versions.c.serial]) |
| 366 |
s = s.where(where_clause) |
| 367 |
r = self.conn.execute(s)
|
| 368 |
hashes = [] |
| 369 |
serials = [] |
| 370 |
for row in r.fetchall(): |
| 371 |
hashes += [row[0]]
|
| 372 |
serials += [row[1]]
|
| 373 |
r.close() |
| 374 |
|
| 375 |
#delete versions
|
| 376 |
s = self.versions.delete().where(where_clause)
|
| 377 |
r = self.conn.execute(s)
|
| 378 |
r.close() |
| 379 |
|
| 380 |
#delete nodes
|
| 381 |
s = select([self.nodes.c.node],
|
| 382 |
and_(self.nodes.c.parent == parent,
|
| 383 |
select([func.count(self.versions.c.serial)],
|
| 384 |
self.versions.c.node == self.nodes.c.node). |
| 385 |
as_scalar() == 0))
|
| 386 |
rp = self.conn.execute(s)
|
| 387 |
nodes = [row[0] for row in rp.fetchall()] |
| 388 |
rp.close() |
| 389 |
if nodes:
|
| 390 |
s = self.nodes.delete().where(self.nodes.c.node.in_(nodes)) |
| 391 |
self.conn.execute(s).close()
|
| 392 |
|
| 393 |
return hashes, size, serials
|
| 394 |
|
| 395 |
def node_purge(self, node, before=inf, cluster=0, |
| 396 |
update_statistics_ancestors_depth=None):
|
| 397 |
"""Delete all versions with the specified
|
| 398 |
node and cluster, and return
|
| 399 |
the hashes and size of versions deleted.
|
| 400 |
Clears out the node if it has no remaining versions.
|
| 401 |
"""
|
| 402 |
|
| 403 |
#update statistics
|
| 404 |
s = select([func.count(self.versions.c.serial),
|
| 405 |
func.sum(self.versions.c.size)])
|
| 406 |
where_clause = and_(self.versions.c.node == node,
|
| 407 |
self.versions.c.cluster == cluster)
|
| 408 |
if before != inf:
|
| 409 |
where_clause = and_(where_clause, |
| 410 |
self.versions.c.mtime <= before)
|
| 411 |
s = s.where(where_clause) |
| 412 |
r = self.conn.execute(s)
|
| 413 |
row = r.fetchone() |
| 414 |
nr, size = row[0], row[1] |
| 415 |
r.close() |
| 416 |
if not nr: |
| 417 |
return (), 0, () |
| 418 |
mtime = time() |
| 419 |
self.statistics_update_ancestors(node, -nr, -size, mtime, cluster,
|
| 420 |
update_statistics_ancestors_depth) |
| 421 |
|
| 422 |
s = select([self.versions.c.hash, self.versions.c.serial]) |
| 423 |
s = s.where(where_clause) |
| 424 |
r = self.conn.execute(s)
|
| 425 |
hashes = [] |
| 426 |
serials = [] |
| 427 |
for row in r.fetchall(): |
| 428 |
hashes += [row[0]]
|
| 429 |
serials += [row[1]]
|
| 430 |
r.close() |
| 431 |
|
| 432 |
#delete versions
|
| 433 |
s = self.versions.delete().where(where_clause)
|
| 434 |
r = self.conn.execute(s)
|
| 435 |
r.close() |
| 436 |
|
| 437 |
#delete nodes
|
| 438 |
s = select([self.nodes.c.node],
|
| 439 |
and_(self.nodes.c.node == node,
|
| 440 |
select([func.count(self.versions.c.serial)],
|
| 441 |
self.versions.c.node == self.nodes.c.node). |
| 442 |
as_scalar() == 0))
|
| 443 |
rp = self.conn.execute(s)
|
| 444 |
nodes = [row[0] for row in rp.fetchall()] |
| 445 |
rp.close() |
| 446 |
if nodes:
|
| 447 |
s = self.nodes.delete().where(self.nodes.c.node.in_(nodes)) |
| 448 |
self.conn.execute(s).close()
|
| 449 |
|
| 450 |
return hashes, size, serials
|
| 451 |
|
| 452 |
def node_remove(self, node, update_statistics_ancestors_depth=None): |
| 453 |
"""Remove the node specified.
|
| 454 |
Return false if the node has children or is not found.
|
| 455 |
"""
|
| 456 |
|
| 457 |
if self.node_count_children(node): |
| 458 |
return False |
| 459 |
|
| 460 |
mtime = time() |
| 461 |
s = select([func.count(self.versions.c.serial),
|
| 462 |
func.sum(self.versions.c.size),
|
| 463 |
self.versions.c.cluster])
|
| 464 |
s = s.where(self.versions.c.node == node)
|
| 465 |
s = s.group_by(self.versions.c.cluster)
|
| 466 |
r = self.conn.execute(s)
|
| 467 |
for population, size, cluster in r.fetchall(): |
| 468 |
self.statistics_update_ancestors(
|
| 469 |
node, -population, -size, mtime, cluster, |
| 470 |
update_statistics_ancestors_depth) |
| 471 |
r.close() |
| 472 |
|
| 473 |
s = self.nodes.delete().where(self.nodes.c.node == node) |
| 474 |
self.conn.execute(s).close()
|
| 475 |
return True |
| 476 |
|
| 477 |
def node_accounts(self, accounts=()): |
| 478 |
s = select([self.nodes.c.path, self.nodes.c.node]) |
| 479 |
s = s.where(and_(self.nodes.c.node != 0, |
| 480 |
self.nodes.c.parent == 0)) |
| 481 |
if accounts:
|
| 482 |
s = s.where(self.nodes.c.path.in_(accounts))
|
| 483 |
r = self.conn.execute(s)
|
| 484 |
rows = r.fetchall() |
| 485 |
r.close() |
| 486 |
return rows
|
| 487 |
|
| 488 |
def node_account_quotas(self): |
| 489 |
s = select([self.nodes.c.path, self.policy.c.value]) |
| 490 |
s = s.where(and_(self.nodes.c.node != 0, |
| 491 |
self.nodes.c.parent == 0)) |
| 492 |
s = s.where(self.nodes.c.node == self.policy.c.node) |
| 493 |
s = s.where(self.policy.c.key == 'quota') |
| 494 |
r = self.conn.execute(s)
|
| 495 |
rows = r.fetchall() |
| 496 |
r.close() |
| 497 |
return dict(rows) |
| 498 |
|
| 499 |
def node_account_usage(self, account=None, cluster=0): |
| 500 |
"""Return usage for a specific account.
|
| 501 |
|
| 502 |
Keyword arguments:
|
| 503 |
account -- (default None: list usage for all the accounts)
|
| 504 |
cluster -- list current, history or deleted usage (default 0: normal)
|
| 505 |
"""
|
| 506 |
|
| 507 |
n1 = self.nodes.alias('n1') |
| 508 |
n2 = self.nodes.alias('n2') |
| 509 |
n3 = self.nodes.alias('n3') |
| 510 |
|
| 511 |
s = select([n3.c.path, func.sum(self.versions.c.size)])
|
| 512 |
s = s.where(n1.c.node == self.versions.c.node)
|
| 513 |
s = s.where(self.versions.c.cluster == cluster)
|
| 514 |
s = s.where(n1.c.parent == n2.c.node) |
| 515 |
s = s.where(n2.c.parent == n3.c.node) |
| 516 |
s = s.where(n3.c.parent == 0)
|
| 517 |
s = s.where(n3.c.node != 0)
|
| 518 |
if account:
|
| 519 |
s = s.where(n3.c.path == account) |
| 520 |
s = s.group_by(n3.c.path) |
| 521 |
r = self.conn.execute(s)
|
| 522 |
usage = r.fetchall() |
| 523 |
r.close() |
| 524 |
return dict(usage) |
| 525 |
|
| 526 |
def policy_get(self, node): |
| 527 |
s = select([self.policy.c.key, self.policy.c.value], |
| 528 |
self.policy.c.node == node)
|
| 529 |
r = self.conn.execute(s)
|
| 530 |
d = dict(r.fetchall())
|
| 531 |
r.close() |
| 532 |
return d
|
| 533 |
|
| 534 |
def policy_set(self, node, policy): |
| 535 |
#insert or replace
|
| 536 |
for k, v in policy.iteritems(): |
| 537 |
s = self.policy.update().where(and_(self.policy.c.node == node, |
| 538 |
self.policy.c.key == k))
|
| 539 |
s = s.values(value=v) |
| 540 |
rp = self.conn.execute(s)
|
| 541 |
rp.close() |
| 542 |
if rp.rowcount == 0: |
| 543 |
s = self.policy.insert()
|
| 544 |
values = {'node': node, 'key': k, 'value': v}
|
| 545 |
r = self.conn.execute(s, values)
|
| 546 |
r.close() |
| 547 |
|
| 548 |
def statistics_get(self, node, cluster=0): |
| 549 |
"""Return population, total size and last mtime
|
| 550 |
for all versions under node that belong to the cluster.
|
| 551 |
"""
|
| 552 |
|
| 553 |
s = select([self.statistics.c.population,
|
| 554 |
self.statistics.c.size,
|
| 555 |
self.statistics.c.mtime])
|
| 556 |
s = s.where(and_(self.statistics.c.node == node,
|
| 557 |
self.statistics.c.cluster == cluster))
|
| 558 |
r = self.conn.execute(s)
|
| 559 |
row = r.fetchone() |
| 560 |
r.close() |
| 561 |
return row
|
| 562 |
|
| 563 |
def statistics_update(self, node, population, size, mtime, cluster=0): |
| 564 |
"""Update the statistics of the given node.
|
| 565 |
Statistics keep track the population, total
|
| 566 |
size of objects and mtime in the node's namespace.
|
| 567 |
May be zero or positive or negative numbers.
|
| 568 |
"""
|
| 569 |
s = select([self.statistics.c.population, self.statistics.c.size], |
| 570 |
and_(self.statistics.c.node == node,
|
| 571 |
self.statistics.c.cluster == cluster))
|
| 572 |
rp = self.conn.execute(s)
|
| 573 |
r = rp.fetchone() |
| 574 |
rp.close() |
| 575 |
if not r: |
| 576 |
prepopulation, presize = (0, 0) |
| 577 |
else:
|
| 578 |
prepopulation, presize = r |
| 579 |
population += prepopulation |
| 580 |
population = max(population, 0) |
| 581 |
size += presize |
| 582 |
|
| 583 |
#insert or replace
|
| 584 |
#TODO better upsert
|
| 585 |
u = self.statistics.update().where(and_(
|
| 586 |
self.statistics.c.node == node,
|
| 587 |
self.statistics.c.cluster == cluster))
|
| 588 |
u = u.values(population=population, size=size, mtime=mtime) |
| 589 |
rp = self.conn.execute(u)
|
| 590 |
rp.close() |
| 591 |
if rp.rowcount == 0: |
| 592 |
ins = self.statistics.insert()
|
| 593 |
ins = ins.values(node=node, population=population, size=size, |
| 594 |
mtime=mtime, cluster=cluster) |
| 595 |
self.conn.execute(ins).close()
|
| 596 |
|
| 597 |
def statistics_update_ancestors(self, node, population, size, mtime, |
| 598 |
cluster=0, recursion_depth=None): |
| 599 |
"""Update the statistics of the given node's parent.
|
| 600 |
Then recursively update all parents up to the root
|
| 601 |
or up to the ``recursion_depth`` (if not None).
|
| 602 |
Population is not recursive.
|
| 603 |
"""
|
| 604 |
|
| 605 |
i = 0
|
| 606 |
while True: |
| 607 |
if node == ROOTNODE:
|
| 608 |
break
|
| 609 |
if recursion_depth and recursion_depth <= i: |
| 610 |
break
|
| 611 |
props = self.node_get_properties(node)
|
| 612 |
if props is None: |
| 613 |
break
|
| 614 |
parent, path = props |
| 615 |
self.statistics_update(parent, population, size, mtime, cluster)
|
| 616 |
node = parent |
| 617 |
population = 0 # Population isn't recursive |
| 618 |
i += 1
|
| 619 |
|
| 620 |
def statistics_latest(self, node, before=inf, except_cluster=0): |
| 621 |
"""Compute population, total size and last mtime
|
| 622 |
for all latest versions under node that
|
| 623 |
do not belong to the cluster.
|
| 624 |
"""
|
| 625 |
|
| 626 |
# The node.
|
| 627 |
props = self.node_get_properties(node)
|
| 628 |
if props is None: |
| 629 |
return None |
| 630 |
parent, path = props |
| 631 |
|
| 632 |
# The latest version.
|
| 633 |
s = select([self.versions.c.size,
|
| 634 |
self.versions.c.mtime])
|
| 635 |
if before != inf:
|
| 636 |
filtered = select([func.max(self.versions.c.serial)],
|
| 637 |
self.versions.c.node == node)
|
| 638 |
filtered = filtered.where(self.versions.c.mtime < before)
|
| 639 |
else:
|
| 640 |
filtered = select([self.nodes.c.latest_version],
|
| 641 |
self.nodes.c.node == node)
|
| 642 |
s = s.where(and_(self.versions.c.cluster != except_cluster,
|
| 643 |
self.versions.c.serial == filtered))
|
| 644 |
r = self.conn.execute(s)
|
| 645 |
props = r.fetchone() |
| 646 |
r.close() |
| 647 |
if not props: |
| 648 |
return None |
| 649 |
mtime = props.mtime |
| 650 |
|
| 651 |
# First level, just under node (get population).
|
| 652 |
v = self.versions.alias('v') |
| 653 |
s = select([func.count(v.c.serial), |
| 654 |
func.sum(v.c.size), |
| 655 |
func.max(v.c.mtime)]) |
| 656 |
if before != inf:
|
| 657 |
c1 = select([func.max(self.versions.c.serial)],
|
| 658 |
and_(self.versions.c.mtime < before,
|
| 659 |
self.versions.c.node == v.c.node))
|
| 660 |
else:
|
| 661 |
c1 = select([self.nodes.c.latest_version])
|
| 662 |
c1 = c1.where(self.nodes.c.node == v.c.node)
|
| 663 |
c2 = select([self.nodes.c.node], self.nodes.c.parent == node) |
| 664 |
s = s.where(and_(v.c.serial == c1, |
| 665 |
v.c.cluster != except_cluster, |
| 666 |
v.c.node.in_(c2))) |
| 667 |
rp = self.conn.execute(s)
|
| 668 |
r = rp.fetchone() |
| 669 |
rp.close() |
| 670 |
if not r: |
| 671 |
return None |
| 672 |
count = r[0]
|
| 673 |
mtime = max(mtime, r[2]) |
| 674 |
if count == 0: |
| 675 |
return (0, 0, mtime) |
| 676 |
|
| 677 |
# All children (get size and mtime).
|
| 678 |
# This is why the full path is stored.
|
| 679 |
if before != inf:
|
| 680 |
s = select([func.count(v.c.serial), |
| 681 |
func.sum(v.c.size), |
| 682 |
func.max(v.c.mtime)]) |
| 683 |
c1 = select([func.max(self.versions.c.serial)],
|
| 684 |
and_(self.versions.c.mtime < before,
|
| 685 |
self.versions.c.node == v.c.node))
|
| 686 |
else:
|
| 687 |
inner_join = \ |
| 688 |
self.versions.join(self.nodes, onclause=\ |
| 689 |
self.versions.c.serial == self.nodes.c.latest_version) |
| 690 |
s = select([func.count(self.versions.c.serial),
|
| 691 |
func.sum(self.versions.c.size),
|
| 692 |
func.max(self.versions.c.mtime)], from_obj=[inner_join])
|
| 693 |
|
| 694 |
c2 = select([self.nodes.c.node],
|
| 695 |
self.nodes.c.path.like(self.escape_like(path) + '%', |
| 696 |
escape=ESCAPE_CHAR)) |
| 697 |
if before != inf:
|
| 698 |
s = s.where(and_(v.c.serial == c1, |
| 699 |
v.c.cluster != except_cluster, |
| 700 |
v.c.node.in_(c2))) |
| 701 |
else:
|
| 702 |
s = s.where(and_(self.versions.c.cluster != except_cluster,
|
| 703 |
self.versions.c.node.in_(c2)))
|
| 704 |
|
| 705 |
rp = self.conn.execute(s)
|
| 706 |
r = rp.fetchone() |
| 707 |
rp.close() |
| 708 |
if not r: |
| 709 |
return None |
| 710 |
size = long(r[1] - props.size) |
| 711 |
mtime = max(mtime, r[2]) |
| 712 |
return (count, size, mtime)
|
| 713 |
|
| 714 |
def statistics_latest_bulk(self, nodes, before=inf, except_cluster=0): |
| 715 |
"""Compute population, total size and last mtime
|
| 716 |
for all latest versions under node that
|
| 717 |
do not belong to the cluster.
|
| 718 |
"""
|
| 719 |
|
| 720 |
# The node.
|
| 721 |
props = self.node_get_properties_bulk(nodes)
|
| 722 |
if not props: # empty list |
| 723 |
return ()
|
| 724 |
paths = [p.path for p in props] |
| 725 |
|
| 726 |
# The latest version.
|
| 727 |
s = select([self.versions.c.node,
|
| 728 |
self.versions.c.size,
|
| 729 |
self.versions.c.mtime])
|
| 730 |
if before != inf:
|
| 731 |
filtered = select([func.max(self.versions.c.serial)],
|
| 732 |
self.versions.c.node.in_(nodes))
|
| 733 |
filtered = filtered.where(self.versions.c.mtime < before)
|
| 734 |
else:
|
| 735 |
filtered = select([self.nodes.c.latest_version],
|
| 736 |
self.nodes.c.node.in_(nodes))
|
| 737 |
s = s.where(and_(self.versions.c.cluster != except_cluster,
|
| 738 |
self.versions.c.serial.in_(filtered)))
|
| 739 |
r = self.conn.execute(s)
|
| 740 |
props = r.fetchall() |
| 741 |
r.close() |
| 742 |
if not props: |
| 743 |
return ()
|
| 744 |
mtimes = {}
|
| 745 |
sizes = {}
|
| 746 |
for p in props: |
| 747 |
mtimes[p.node] = p.mtime |
| 748 |
sizes[p.node] = p.size |
| 749 |
|
| 750 |
# First level, just under node (get population).
|
| 751 |
n = self.nodes.alias('n') |
| 752 |
v = self.versions.alias('v') |
| 753 |
s = select([n.c.parent, |
| 754 |
func.count(v.c.serial).label('count'),
|
| 755 |
func.sum(v.c.size).label('total_size'),
|
| 756 |
func.max(v.c.mtime).label('max_timestamp')])
|
| 757 |
if before != inf:
|
| 758 |
c1 = select([func.max(self.versions.c.serial)],
|
| 759 |
and_(self.versions.c.mtime < before,
|
| 760 |
self.versions.c.node == v.c.node))
|
| 761 |
else:
|
| 762 |
c1 = select([self.nodes.c.latest_version])
|
| 763 |
c1 = c1.where(self.nodes.c.node == v.c.node)
|
| 764 |
c2 = select([self.nodes.c.node], self.nodes.c.parent.in_(nodes)) |
| 765 |
s = s.where(and_(v.c.serial == c1, |
| 766 |
v.c.cluster != except_cluster, |
| 767 |
v.c.node.in_(c2), |
| 768 |
v.c.node == n.c.node)) |
| 769 |
s = s.group_by(n.c.parent) |
| 770 |
rp = self.conn.execute(s)
|
| 771 |
r = rp.fetchall() |
| 772 |
rp.close() |
| 773 |
if not r: |
| 774 |
return None |
| 775 |
counts = {}
|
| 776 |
for stats in r: |
| 777 |
counts[stats.parent] = stats.total_size |
| 778 |
mtimes[stats.parent] = max(mtimes[stats.parent],
|
| 779 |
stats.max_timestamp) |
| 780 |
|
| 781 |
# All children (get size and mtime).
|
| 782 |
# This is why the full path is stored.
|
| 783 |
n1 = self.nodes.alias('n1') |
| 784 |
n2 = self.nodes.alias('n2') |
| 785 |
if before != inf:
|
| 786 |
s = select([n2.c.node, |
| 787 |
func.count(v.c.serial).label('count'),
|
| 788 |
func.sum(v.c.size).label('total_size'),
|
| 789 |
func.max(v.c.mtime).label('max_timestamp')])
|
| 790 |
c1 = select([func.max(self.versions.c.serial)],
|
| 791 |
and_(self.versions.c.mtime < before,
|
| 792 |
self.versions.c.node == v.c.node))
|
| 793 |
else:
|
| 794 |
inner_join = \ |
| 795 |
self.versions.join(self.nodes, onclause=\ |
| 796 |
self.versions.c.serial == self.nodes.c.latest_version) |
| 797 |
s = select([n2.c.node, |
| 798 |
func.count(self.versions.c.serial).label('count'), |
| 799 |
func.sum(self.versions.c.size).label('total_size'), |
| 800 |
func.max(self.versions.c.mtime).label('max_timestamp')], |
| 801 |
from_obj=[inner_join]) |
| 802 |
|
| 803 |
c2 = select([self.nodes.c.node])
|
| 804 |
like = lambda p: self.nodes.c.path.like(self.escape_like(p) + '%', |
| 805 |
escape=ESCAPE_CHAR) |
| 806 |
c2 = c2.where(or_(*map(like, paths)))
|
| 807 |
if before != inf:
|
| 808 |
s = s.where(and_(v.c.serial == c1, |
| 809 |
v.c.cluster != except_cluster, |
| 810 |
v.c.node.in_(c2))) |
| 811 |
else:
|
| 812 |
s = s.where(and_(self.versions.c.cluster != except_cluster,
|
| 813 |
self.versions.c.node.in_(c2)))
|
| 814 |
s = s.where(self.nodes.c.parent == n1.c.node)
|
| 815 |
s = s.where(n1.c.parent == n2.c.node) |
| 816 |
s = s.where(n2.c.node != ROOTNODE) |
| 817 |
s = s.group_by(n2.c.node) |
| 818 |
|
| 819 |
rp = self.conn.execute(s)
|
| 820 |
r = rp.fetchall() |
| 821 |
rp.close() |
| 822 |
if not r: |
| 823 |
return None |
| 824 |
for stats in r: |
| 825 |
sizes[stats.node] = long(stats.total_size - sizes[stats.node])
|
| 826 |
mtimes[stats.node] = max(mtimes[stats.node], stats.max_timestamp)
|
| 827 |
l = [] |
| 828 |
append = l.append |
| 829 |
for node in nodes: |
| 830 |
append((node, counts[node], sizes[node], mtimes[node])) |
| 831 |
return l
|
| 832 |
|
| 833 |
def nodes_set_latest_version(self, node, serial): |
| 834 |
s = self.nodes.update().where(self.nodes.c.node == node) |
| 835 |
s = s.values(latest_version=serial) |
| 836 |
self.conn.execute(s).close()
|
| 837 |
|
| 838 |
def version_create(self, node, hash, size, type, source, muser, uuid, |
| 839 |
checksum, cluster=0,
|
| 840 |
update_statistics_ancestors_depth=None):
|
| 841 |
"""Create a new version from the given properties.
|
| 842 |
Return the (serial, mtime) of the new version.
|
| 843 |
"""
|
| 844 |
|
| 845 |
mtime = time() |
| 846 |
s = self.versions.insert().values(
|
| 847 |
node=node, hash=hash, size=size, type=type, source=source, |
| 848 |
mtime=mtime, muser=muser, uuid=uuid, checksum=checksum, |
| 849 |
cluster=cluster) |
| 850 |
serial = self.conn.execute(s).inserted_primary_key[0] |
| 851 |
self.statistics_update_ancestors(node, 1, size, mtime, cluster, |
| 852 |
update_statistics_ancestors_depth) |
| 853 |
|
| 854 |
self.nodes_set_latest_version(node, serial)
|
| 855 |
|
| 856 |
return serial, mtime
|
| 857 |
|
| 858 |
def version_lookup(self, node, before=inf, cluster=0, all_props=True): |
| 859 |
"""Lookup the current version of the given node.
|
| 860 |
Return a list with its properties:
|
| 861 |
(serial, node, hash, size, type, source, mtime,
|
| 862 |
muser, uuid, checksum, cluster)
|
| 863 |
or None if the current version is not found in the given cluster.
|
| 864 |
"""
|
| 865 |
|
| 866 |
v = self.versions.alias('v') |
| 867 |
if not all_props: |
| 868 |
s = select([v.c.serial]) |
| 869 |
else:
|
| 870 |
s = select([v.c.serial, v.c.node, v.c.hash, |
| 871 |
v.c.size, v.c.type, v.c.source, |
| 872 |
v.c.mtime, v.c.muser, v.c.uuid, |
| 873 |
v.c.checksum, v.c.cluster]) |
| 874 |
if before != inf:
|
| 875 |
c = select([func.max(self.versions.c.serial)],
|
| 876 |
self.versions.c.node == node)
|
| 877 |
c = c.where(self.versions.c.mtime < before)
|
| 878 |
else:
|
| 879 |
c = select([self.nodes.c.latest_version],
|
| 880 |
self.nodes.c.node == node)
|
| 881 |
s = s.where(and_(v.c.serial == c, |
| 882 |
v.c.cluster == cluster)) |
| 883 |
r = self.conn.execute(s)
|
| 884 |
props = r.fetchone() |
| 885 |
r.close() |
| 886 |
if props:
|
| 887 |
return props
|
| 888 |
return None |
| 889 |
|
| 890 |
def version_lookup_bulk(self, nodes, before=inf, cluster=0, |
| 891 |
all_props=True, order_by_path=False): |
| 892 |
"""Lookup the current versions of the given nodes.
|
| 893 |
Return a list with their properties:
|
| 894 |
(serial, node, hash, size, type, source, mtime, muser, uuid,
|
| 895 |
checksum, cluster).
|
| 896 |
"""
|
| 897 |
if not nodes: |
| 898 |
return ()
|
| 899 |
v = self.versions.alias('v') |
| 900 |
n = self.nodes.alias('n') |
| 901 |
if not all_props: |
| 902 |
s = select([v.c.serial]) |
| 903 |
else:
|
| 904 |
s = select([v.c.serial, v.c.node, v.c.hash, |
| 905 |
v.c.size, v.c.type, v.c.source, |
| 906 |
v.c.mtime, v.c.muser, v.c.uuid, |
| 907 |
v.c.checksum, v.c.cluster]) |
| 908 |
if before != inf:
|
| 909 |
c = select([func.max(self.versions.c.serial)],
|
| 910 |
self.versions.c.node.in_(nodes))
|
| 911 |
c = c.where(self.versions.c.mtime < before)
|
| 912 |
c = c.group_by(self.versions.c.node)
|
| 913 |
else:
|
| 914 |
c = select([self.nodes.c.latest_version],
|
| 915 |
self.nodes.c.node.in_(nodes))
|
| 916 |
s = s.where(and_(v.c.serial.in_(c), |
| 917 |
v.c.cluster == cluster)) |
| 918 |
if order_by_path:
|
| 919 |
s = s.where(v.c.node == n.c.node) |
| 920 |
s = s.order_by(n.c.path) |
| 921 |
else:
|
| 922 |
s = s.order_by(v.c.node) |
| 923 |
r = self.conn.execute(s)
|
| 924 |
rows = r.fetchall() |
| 925 |
r.close() |
| 926 |
return rows
|
| 927 |
|
| 928 |
def version_get_properties(self, serial, keys=(), propnames=_propnames, |
| 929 |
node=None):
|
| 930 |
"""Return a sequence of values for the properties of
|
| 931 |
the version specified by serial and the keys, in the order given.
|
| 932 |
If keys is empty, return all properties in the order
|
| 933 |
(serial, node, hash, size, type, source, mtime, muser, uuid,
|
| 934 |
checksum, cluster).
|
| 935 |
"""
|
| 936 |
|
| 937 |
v = self.versions.alias()
|
| 938 |
s = select([v.c.serial, v.c.node, v.c.hash, |
| 939 |
v.c.size, v.c.type, v.c.source, |
| 940 |
v.c.mtime, v.c.muser, v.c.uuid, |
| 941 |
v.c.checksum, v.c.cluster], v.c.serial == serial) |
| 942 |
if node is not None: |
| 943 |
s = s.where(v.c.node == node) |
| 944 |
rp = self.conn.execute(s)
|
| 945 |
r = rp.fetchone() |
| 946 |
rp.close() |
| 947 |
if r is None: |
| 948 |
return r
|
| 949 |
|
| 950 |
if not keys: |
| 951 |
return r
|
| 952 |
return [r[propnames[k]] for k in keys if k in propnames] |
| 953 |
|
| 954 |
def version_put_property(self, serial, key, value): |
| 955 |
"""Set value for the property of version specified by key."""
|
| 956 |
|
| 957 |
if key not in _propnames: |
| 958 |
return
|
| 959 |
s = self.versions.update()
|
| 960 |
s = s.where(self.versions.c.serial == serial)
|
| 961 |
s = s.values(**{key: value})
|
| 962 |
self.conn.execute(s).close()
|
| 963 |
|
| 964 |
def version_recluster(self, serial, cluster, |
| 965 |
update_statistics_ancestors_depth=None):
|
| 966 |
"""Move the version into another cluster."""
|
| 967 |
|
| 968 |
props = self.version_get_properties(serial)
|
| 969 |
if not props: |
| 970 |
return
|
| 971 |
node = props[NODE] |
| 972 |
size = props[SIZE] |
| 973 |
oldcluster = props[CLUSTER] |
| 974 |
if cluster == oldcluster:
|
| 975 |
return
|
| 976 |
|
| 977 |
mtime = time() |
| 978 |
self.statistics_update_ancestors(node, -1, -size, mtime, oldcluster, |
| 979 |
update_statistics_ancestors_depth) |
| 980 |
self.statistics_update_ancestors(node, 1, size, mtime, cluster, |
| 981 |
update_statistics_ancestors_depth) |
| 982 |
|
| 983 |
s = self.versions.update()
|
| 984 |
s = s.where(self.versions.c.serial == serial)
|
| 985 |
s = s.values(cluster=cluster) |
| 986 |
self.conn.execute(s).close()
|
| 987 |
|
| 988 |
def version_remove(self, serial, update_statistics_ancestors_depth=None): |
| 989 |
"""Remove the serial specified."""
|
| 990 |
|
| 991 |
props = self.version_get_properties(serial)
|
| 992 |
if not props: |
| 993 |
return
|
| 994 |
node = props[NODE] |
| 995 |
hash = props[HASH] |
| 996 |
size = props[SIZE] |
| 997 |
cluster = props[CLUSTER] |
| 998 |
|
| 999 |
mtime = time() |
| 1000 |
self.statistics_update_ancestors(node, -1, -size, mtime, cluster, |
| 1001 |
update_statistics_ancestors_depth) |
| 1002 |
|
| 1003 |
s = self.versions.delete().where(self.versions.c.serial == serial) |
| 1004 |
self.conn.execute(s).close()
|
| 1005 |
|
| 1006 |
props = self.version_lookup(node, cluster=cluster, all_props=False) |
| 1007 |
if props:
|
| 1008 |
self.nodes_set_latest_version(node, serial)
|
| 1009 |
|
| 1010 |
return hash, size |
| 1011 |
|
| 1012 |
def attribute_get(self, serial, domain, keys=()): |
| 1013 |
"""Return a list of (key, value) pairs of the specific version.
|
| 1014 |
|
| 1015 |
If keys is empty, return all attributes.
|
| 1016 |
Othwerise, return only those specified.
|
| 1017 |
"""
|
| 1018 |
|
| 1019 |
if keys:
|
| 1020 |
attrs = self.attributes.alias()
|
| 1021 |
s = select([attrs.c.key, attrs.c.value]) |
| 1022 |
s = s.where(and_(attrs.c.key.in_(keys), |
| 1023 |
attrs.c.serial == serial, |
| 1024 |
attrs.c.domain == domain)) |
| 1025 |
else:
|
| 1026 |
attrs = self.attributes.alias()
|
| 1027 |
s = select([attrs.c.key, attrs.c.value]) |
| 1028 |
s = s.where(and_(attrs.c.serial == serial, |
| 1029 |
attrs.c.domain == domain)) |
| 1030 |
r = self.conn.execute(s)
|
| 1031 |
l = r.fetchall() |
| 1032 |
r.close() |
| 1033 |
return l
|
| 1034 |
|
| 1035 |
def attribute_set(self, serial, domain, node, items, is_latest=True): |
| 1036 |
"""Set the attributes of the version specified by serial.
|
| 1037 |
Receive attributes as an iterable of (key, value) pairs.
|
| 1038 |
"""
|
| 1039 |
#insert or replace
|
| 1040 |
#TODO better upsert
|
| 1041 |
for k, v in items: |
| 1042 |
s = self.attributes.update()
|
| 1043 |
s = s.where(and_(self.attributes.c.serial == serial,
|
| 1044 |
self.attributes.c.domain == domain,
|
| 1045 |
self.attributes.c.key == k))
|
| 1046 |
s = s.values(value=v) |
| 1047 |
rp = self.conn.execute(s)
|
| 1048 |
rp.close() |
| 1049 |
if rp.rowcount == 0: |
| 1050 |
s = self.attributes.insert()
|
| 1051 |
s = s.values(serial=serial, domain=domain, node=node, |
| 1052 |
is_latest=is_latest, key=k, value=v) |
| 1053 |
self.conn.execute(s).close()
|
| 1054 |
|
| 1055 |
def attribute_del(self, serial, domain, keys=()): |
| 1056 |
"""Delete attributes of the version specified by serial.
|
| 1057 |
If keys is empty, delete all attributes.
|
| 1058 |
Otherwise delete those specified.
|
| 1059 |
"""
|
| 1060 |
|
| 1061 |
if keys:
|
| 1062 |
#TODO more efficient way to do this?
|
| 1063 |
for key in keys: |
| 1064 |
s = self.attributes.delete()
|
| 1065 |
s = s.where(and_(self.attributes.c.serial == serial,
|
| 1066 |
self.attributes.c.domain == domain,
|
| 1067 |
self.attributes.c.key == key))
|
| 1068 |
self.conn.execute(s).close()
|
| 1069 |
else:
|
| 1070 |
s = self.attributes.delete()
|
| 1071 |
s = s.where(and_(self.attributes.c.serial == serial,
|
| 1072 |
self.attributes.c.domain == domain))
|
| 1073 |
self.conn.execute(s).close()
|
| 1074 |
|
| 1075 |
def attribute_copy(self, source, dest): |
| 1076 |
s = select( |
| 1077 |
[dest, self.attributes.c.domain, self.attributes.c.node, |
| 1078 |
self.attributes.c.key, self.attributes.c.value], |
| 1079 |
self.attributes.c.serial == source)
|
| 1080 |
rp = self.conn.execute(s)
|
| 1081 |
attributes = rp.fetchall() |
| 1082 |
rp.close() |
| 1083 |
for dest, domain, node, k, v in attributes: |
| 1084 |
select_src_node = select( |
| 1085 |
[self.versions.c.node],
|
| 1086 |
self.versions.c.serial == dest)
|
| 1087 |
# insert or replace
|
| 1088 |
s = self.attributes.update().where(and_(
|
| 1089 |
self.attributes.c.serial == dest,
|
| 1090 |
self.attributes.c.domain == domain,
|
| 1091 |
self.attributes.c.key == k))
|
| 1092 |
s = s.values(node=select_src_node, value=v) |
| 1093 |
rp = self.conn.execute(s)
|
| 1094 |
rp.close() |
| 1095 |
if rp.rowcount == 0: |
| 1096 |
s = self.attributes.insert()
|
| 1097 |
s = s.values(serial=dest, domain=domain, node=select_src_node, |
| 1098 |
is_latest=True, key=k, value=v)
|
| 1099 |
self.conn.execute(s).close()
|
| 1100 |
|
| 1101 |
def attribute_unset_is_latest(self, node, exclude): |
| 1102 |
u = self.attributes.update().where(and_(
|
| 1103 |
self.attributes.c.node == node,
|
| 1104 |
self.attributes.c.serial != exclude)).values({'is_latest': False}) |
| 1105 |
self.conn.execute(u)
|
| 1106 |
|
| 1107 |
def latest_attribute_keys(self, parent, domain, before=inf, |
| 1108 |
except_cluster=0, pathq=None): |
| 1109 |
"""Return a list with all keys pairs defined
|
| 1110 |
for all latest versions under parent that
|
| 1111 |
do not belong to the cluster.
|
| 1112 |
"""
|
| 1113 |
|
| 1114 |
pathq = pathq or []
|
| 1115 |
|
| 1116 |
# TODO: Use another table to store before=inf results.
|
| 1117 |
v = self.versions.alias('v') |
| 1118 |
s = select([self.attributes.c.key]).distinct()
|
| 1119 |
if before != inf:
|
| 1120 |
filtered = select([func.max(v.c.serial)], |
| 1121 |
and_(v.c.mtime < before, |
| 1122 |
v.c.node == self.versions.c.node))
|
| 1123 |
else:
|
| 1124 |
filtered = select([self.nodes.c.latest_version])
|
| 1125 |
filtered = filtered.where(self.nodes.c.node == \
|
| 1126 |
self.versions.c.node).correlate(self.versions) |
| 1127 |
s = s.where(self.versions.c.serial == filtered)
|
| 1128 |
s = s.where(self.versions.c.cluster != except_cluster)
|
| 1129 |
s = s.where(self.versions.c.node.in_(select([self.nodes.c.node], |
| 1130 |
self.nodes.c.parent == parent)))
|
| 1131 |
s = s.where(self.attributes.c.serial == self.versions.c.serial) |
| 1132 |
s = s.where(self.attributes.c.domain == domain)
|
| 1133 |
s = s.where(self.nodes.c.node == self.versions.c.node) |
| 1134 |
s = s.order_by(self.attributes.c.key)
|
| 1135 |
conja = [] |
| 1136 |
conjb = [] |
| 1137 |
for path, match in pathq: |
| 1138 |
if match == MATCH_PREFIX:
|
| 1139 |
conja.append(self.nodes.c.path.like(self.escape_like(path) + '%', |
| 1140 |
escape=ESCAPE_CHAR)) |
| 1141 |
elif match == MATCH_EXACT:
|
| 1142 |
conjb.append(path) |
| 1143 |
if conja or conjb: |
| 1144 |
s = s.where(or_(self.nodes.c.path.in_(conjb),*conja))
|
| 1145 |
rp = self.conn.execute(s)
|
| 1146 |
rows = rp.fetchall() |
| 1147 |
rp.close() |
| 1148 |
return [r[0] for r in rows] |
| 1149 |
|
| 1150 |
def latest_version_list(self, parent, prefix='', delimiter=None, |
| 1151 |
start='', limit=10000, before=inf, |
| 1152 |
except_cluster=0, pathq=[], domain=None, |
| 1153 |
filterq=[], sizeq=None, all_props=False): |
| 1154 |
"""Return a (list of (path, serial) tuples, list of common prefixes)
|
| 1155 |
for the current versions of the paths with the given parent,
|
| 1156 |
matching the following criteria.
|
| 1157 |
|
| 1158 |
The property tuple for a version is returned if all
|
| 1159 |
of these conditions are true:
|
| 1160 |
|
| 1161 |
a. parent matches
|
| 1162 |
|
| 1163 |
b. path > start
|
| 1164 |
|
| 1165 |
c. path starts with prefix (and paths in pathq)
|
| 1166 |
|
| 1167 |
d. version is the max up to before
|
| 1168 |
|
| 1169 |
e. version is not in cluster
|
| 1170 |
|
| 1171 |
f. the path does not have the delimiter occuring
|
| 1172 |
after the prefix, or ends with the delimiter
|
| 1173 |
|
| 1174 |
g. serial matches the attribute filter query.
|
| 1175 |
|
| 1176 |
A filter query is a comma-separated list of
|
| 1177 |
terms in one of these three forms:
|
| 1178 |
|
| 1179 |
key
|
| 1180 |
an attribute with this key must exist
|
| 1181 |
|
| 1182 |
!key
|
| 1183 |
an attribute with this key must not exist
|
| 1184 |
|
| 1185 |
key ?op value
|
| 1186 |
the attribute with this key satisfies the value
|
| 1187 |
where ?op is one of ==, != <=, >=, <, >.
|
| 1188 |
|
| 1189 |
h. the size is in the range set by sizeq
|
| 1190 |
|
| 1191 |
The list of common prefixes includes the prefixes
|
| 1192 |
matching up to the first delimiter after prefix,
|
| 1193 |
and are reported only once, as "virtual directories".
|
| 1194 |
The delimiter is included in the prefixes.
|
| 1195 |
|
| 1196 |
If arguments are None, then the corresponding matching rule
|
| 1197 |
will always match.
|
| 1198 |
|
| 1199 |
Limit applies to the first list of tuples returned.
|
| 1200 |
|
| 1201 |
If all_props is True, return all properties after path,
|
| 1202 |
not just serial.
|
| 1203 |
"""
|
| 1204 |
|
| 1205 |
if not start or start < prefix: |
| 1206 |
start = strprevling(prefix) |
| 1207 |
nextling = strnextling(prefix) |
| 1208 |
|
| 1209 |
v = self.versions.alias('v') |
| 1210 |
n = self.nodes.alias('n') |
| 1211 |
if before != inf:
|
| 1212 |
filtered = select([func.max(v.c.serial)], |
| 1213 |
and_(v.c.mtime < before, |
| 1214 |
v.c.node == self.versions.c.node))
|
| 1215 |
inner_join = \ |
| 1216 |
self.nodes.join(self.versions, |
| 1217 |
onclause=self.versions.c.serial==filtered)
|
| 1218 |
else:
|
| 1219 |
filtered = select([self.nodes.c.latest_version])
|
| 1220 |
filtered = filtered.where(self.nodes.c.node == self.versions.c.node).correlate(self.versions) |
| 1221 |
inner_join = \ |
| 1222 |
self.nodes.join(self.versions, |
| 1223 |
onclause=\ |
| 1224 |
self.versions.c.serial==filtered)
|
| 1225 |
if not all_props: |
| 1226 |
s = select([self.nodes.c.path,
|
| 1227 |
self.versions.c.serial],from_obj=[inner_join]).distinct()
|
| 1228 |
else:
|
| 1229 |
s = select([self.nodes.c.path,
|
| 1230 |
self.versions.c.serial, self.versions.c.node, |
| 1231 |
self.versions.c.hash,
|
| 1232 |
self.versions.c.size, self.versions.c.type, |
| 1233 |
self.versions.c.source,
|
| 1234 |
self.versions.c.mtime, self.versions.c.muser, |
| 1235 |
self.versions.c.uuid,
|
| 1236 |
self.versions.c.checksum,
|
| 1237 |
self.versions.c.cluster],from_obj=[inner_join]).distinct()
|
| 1238 |
|
| 1239 |
s = s.where(self.versions.c.cluster != except_cluster)
|
| 1240 |
s = s.where(self.versions.c.node.in_(select([self.nodes.c.node], |
| 1241 |
self.nodes.c.parent == parent)))
|
| 1242 |
|
| 1243 |
s = s.where(self.versions.c.node == self.nodes.c.node) |
| 1244 |
s = s.where(and_(self.nodes.c.path > bindparam('start'), |
| 1245 |
self.nodes.c.path < nextling))
|
| 1246 |
conja = [] |
| 1247 |
conjb = [] |
| 1248 |
for path, match in pathq: |
| 1249 |
if match == MATCH_PREFIX:
|
| 1250 |
conja.append(self.nodes.c.path.like(self.escape_like(path) + '%', |
| 1251 |
escape=ESCAPE_CHAR)) |
| 1252 |
elif match == MATCH_EXACT:
|
| 1253 |
conjb.append(path) |
| 1254 |
if conja or conjb: |
| 1255 |
s = s.where(or_(self.nodes.c.path.in_(conjb), *conja))
|
| 1256 |
|
| 1257 |
if sizeq and len(sizeq) == 2: |
| 1258 |
if sizeq[0]: |
| 1259 |
s = s.where(self.versions.c.size >= sizeq[0]) |
| 1260 |
if sizeq[1]: |
| 1261 |
s = s.where(self.versions.c.size < sizeq[1]) |
| 1262 |
|
| 1263 |
if domain and filterq: |
| 1264 |
a = self.attributes.alias('a') |
| 1265 |
included, excluded, opers = parse_filters(filterq) |
| 1266 |
if included:
|
| 1267 |
subs = select([1])
|
| 1268 |
subs = subs.where( |
| 1269 |
self.attributes.c.serial ==
|
| 1270 |
self.versions.c.serial).correlate(self.versions) |
| 1271 |
subs = subs.where(self.attributes.c.domain == domain)
|
| 1272 |
subs = subs.where(or_(*[self.attributes.c.key.op('=')(x) for |
| 1273 |
x in included]))
|
| 1274 |
s = s.where(exists(subs)) |
| 1275 |
if excluded:
|
| 1276 |
subs = select([1])
|
| 1277 |
subs = subs.where( |
| 1278 |
self.attributes.c.serial ==
|
| 1279 |
self.versions.c.serial).correlate(self.versions) |
| 1280 |
subs = subs.where(self.attributes.c.domain == domain)
|
| 1281 |
subs = subs.where(or_(*[self.attributes.c.key.op('=')(x) for |
| 1282 |
x in excluded]))
|
| 1283 |
s = s.where(not_(exists(subs))) |
| 1284 |
if opers:
|
| 1285 |
for k, o, val in opers: |
| 1286 |
subs = select([1])
|
| 1287 |
subs = subs.where(self.attributes.c.serial ==
|
| 1288 |
self.versions.c.serial)
|
| 1289 |
subs = subs.correlate(self.versions)
|
| 1290 |
subs = subs.where(self.attributes.c.domain == domain)
|
| 1291 |
subs = subs.where( |
| 1292 |
and_(self.attributes.c.key.op('=')(k), |
| 1293 |
self.attributes.c.value.op(o)(val)))
|
| 1294 |
s = s.where(exists(subs)) |
| 1295 |
|
| 1296 |
s = s.order_by(self.nodes.c.path)
|
| 1297 |
|
| 1298 |
if not delimiter: |
| 1299 |
s = s.limit(limit) |
| 1300 |
rp = self.conn.execute(s, start=start)
|
| 1301 |
r = rp.fetchall() |
| 1302 |
rp.close() |
| 1303 |
return r, ()
|
| 1304 |
|
| 1305 |
pfz = len(prefix)
|
| 1306 |
dz = len(delimiter)
|
| 1307 |
count = 0
|
| 1308 |
prefixes = [] |
| 1309 |
pappend = prefixes.append |
| 1310 |
matches = [] |
| 1311 |
mappend = matches.append |
| 1312 |
|
| 1313 |
rp = self.conn.execute(s, start=start)
|
| 1314 |
while True: |
| 1315 |
props = rp.fetchone() |
| 1316 |
if props is None: |
| 1317 |
break
|
| 1318 |
path = props[0]
|
| 1319 |
idx = path.find(delimiter, pfz) |
| 1320 |
|
| 1321 |
if idx < 0: |
| 1322 |
mappend(props) |
| 1323 |
count += 1
|
| 1324 |
if count >= limit:
|
| 1325 |
break
|
| 1326 |
continue
|
| 1327 |
|
| 1328 |
if idx + dz == len(path): |
| 1329 |
mappend(props) |
| 1330 |
count += 1
|
| 1331 |
continue # Get one more, in case there is a path. |
| 1332 |
pf = path[:idx + dz] |
| 1333 |
pappend(pf) |
| 1334 |
if count >= limit:
|
| 1335 |
break
|
| 1336 |
|
| 1337 |
rp = self.conn.execute(s, start=strnextling(pf)) # New start. |
| 1338 |
rp.close() |
| 1339 |
|
| 1340 |
return matches, prefixes
|
| 1341 |
|
| 1342 |
def latest_uuid(self, uuid, cluster): |
| 1343 |
"""Return the latest version of the given uuid and cluster.
|
| 1344 |
|
| 1345 |
Return a (path, serial) tuple.
|
| 1346 |
If cluster is None, all clusters are considered.
|
| 1347 |
|
| 1348 |
"""
|
| 1349 |
|
| 1350 |
v = self.versions.alias('v') |
| 1351 |
n = self.nodes.alias('n') |
| 1352 |
s = select([n.c.path, v.c.serial]) |
| 1353 |
filtered = select([func.max(self.versions.c.serial)])
|
| 1354 |
filtered = filtered.where(self.versions.c.uuid == uuid)
|
| 1355 |
if cluster is not None: |
| 1356 |
filtered = filtered.where(self.versions.c.cluster == cluster)
|
| 1357 |
s = s.where(v.c.serial == filtered) |
| 1358 |
s = s.where(n.c.node == v.c.node) |
| 1359 |
|
| 1360 |
r = self.conn.execute(s)
|
| 1361 |
l = r.fetchone() |
| 1362 |
r.close() |
| 1363 |
return l
|
| 1364 |
|
| 1365 |
def domain_object_list(self, domain, paths, cluster=None): |
| 1366 |
"""Return a list of (path, property list, attribute dictionary)
|
| 1367 |
for the objects in the specific domain and cluster.
|
| 1368 |
"""
|
| 1369 |
|
| 1370 |
v = self.versions.alias('v') |
| 1371 |
n = self.nodes.alias('n') |
| 1372 |
a = self.attributes.alias('a') |
| 1373 |
|
| 1374 |
s = select([n.c.path, v.c.serial, v.c.node, v.c.hash, v.c.size, |
| 1375 |
v.c.type, v.c.source, v.c.mtime, v.c.muser, v.c.uuid, |
| 1376 |
v.c.checksum, v.c.cluster, a.c.key, a.c.value]) |
| 1377 |
if cluster:
|
| 1378 |
s = s.where(v.c.cluster == cluster) |
| 1379 |
s = s.where(v.c.serial == a.c.serial) |
| 1380 |
s = s.where(a.c.domain == domain) |
| 1381 |
s = s.where(a.c.node == n.c.node) |
| 1382 |
s = s.where(a.c.is_latest == true()) |
| 1383 |
if paths:
|
| 1384 |
s = s.where(n.c.path.in_(paths)) |
| 1385 |
|
| 1386 |
r = self.conn.execute(s)
|
| 1387 |
rows = r.fetchall() |
| 1388 |
r.close() |
| 1389 |
|
| 1390 |
group_by = itemgetter(slice(12)) |
| 1391 |
rows.sort(key=group_by) |
| 1392 |
groups = groupby(rows, group_by) |
| 1393 |
return [(k[0], k[1:], dict([i[12:] for i in data])) for |
| 1394 |
(k, data) in groups]
|
| 1395 |
|
| 1396 |
def get_props(self, paths): |
| 1397 |
inner_join = \ |
| 1398 |
self.nodes.join(self.versions, |
| 1399 |
onclause=self.versions.c.serial ==
|
| 1400 |
self.nodes.c.latest_version)
|
| 1401 |
cc = self.nodes.c.path.in_(paths)
|
| 1402 |
s = select([self.nodes.c.path, self.versions.c.type], |
| 1403 |
from_obj=[inner_join]).where(cc).distinct() |
| 1404 |
r = self.conn.execute(s)
|
| 1405 |
rows = r.fetchall() |
| 1406 |
r.close() |
| 1407 |
if rows:
|
| 1408 |
return rows
|
| 1409 |
return None |