/usr/share/pyshared/jedi/cache.py is in python-jedi 0.7.0-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 | """
This caching is very important for speed and memory optimizations. There's
nothing really spectacular, just some decorators. The following cache types are
available:
- module caching (`load_module` and `save_module`), which uses pickle and is
really important to assure low load times of modules like ``numpy``.
- the popular ``memoize_default`` works like a typical memoize and returns the
default otherwise.
- ``CachedMetaClass`` uses ``memoize_default`` to do the same with classes.
- ``time_cache`` can be used to cache something for just a limited time span,
which can be useful if there's user interaction and the user cannot react
faster than a certain time.
This module is one of the reasons why |jedi| is not thread-safe. As you can see
there are global variables, which are holding the cache information. Some of
these variables are being cleaned after every API usage.
"""
from __future__ import with_statement
import time
import os
import sys
import json
import hashlib
try:
import cPickle as pickle
except:
import pickle
import shutil
from jedi import settings
from jedi import common
from jedi import debug
# memoize caches will be deleted after every action
memoize_caches = []
time_caches = []
star_import_cache = {}
# for fast_parser, should not be deleted
parser_cache = {}
class ParserCacheItem(object):
def __init__(self, parser, change_time=None):
self.parser = parser
if change_time is None:
change_time = time.time()
self.change_time = change_time
def clear_caches(delete_all=False):
""" Jedi caches many things, that should be completed after each completion
finishes.
:param delete_all: Deletes also the cache that is normally not deleted,
like parser cache, which is important for faster parsing.
"""
global memoize_caches, time_caches
# memorize_caches must never be deleted, because the dicts will get lost in
# the wrappers.
for m in memoize_caches:
m.clear()
if delete_all:
time_caches = []
star_import_cache.clear()
parser_cache.clear()
else:
# normally just kill the expired entries, not all
for tc in time_caches:
# check time_cache for expired entries
for key, (t, value) in list(tc.items()):
if t < time.time():
# delete expired entries
del tc[key]
def memoize_default(default=None, cache=memoize_caches):
""" This is a typical memoization decorator, BUT there is one difference:
To prevent recursion it sets defaults.
Preventing recursion is in this case the much bigger use than speed. I
don't think, that there is a big speed difference, but there are many cases
where recursion could happen (think about a = b; b = a).
"""
def func(function):
memo = {}
cache.append(memo)
def wrapper(*args, **kwargs):
key = (args, frozenset(kwargs.items()))
if key in memo:
return memo[key]
else:
memo[key] = default
rv = function(*args, **kwargs)
memo[key] = rv
return rv
return wrapper
return func
class CachedMetaClass(type):
""" This is basically almost the same than the decorator above, it just
caches class initializations. I haven't found any other way, so I do it
with meta classes.
"""
@memoize_default()
def __call__(self, *args, **kwargs):
return super(CachedMetaClass, self).__call__(*args, **kwargs)
def time_cache(time_add_setting):
""" This decorator works as follows: Call it with a setting and after that
use the function with a callable that returns the key.
But: This function is only called if the key is not available. After a
certain amount of time (`time_add_setting`) the cache is invalid.
"""
def _temp(key_func):
dct = {}
time_caches.append(dct)
def wrapper(optional_callable, *args, **kwargs):
key = key_func(*args, **kwargs)
value = None
if key in dct:
expiry, value = dct[key]
if expiry > time.time():
return value
value = optional_callable()
time_add = getattr(settings, time_add_setting)
if key is not None:
dct[key] = time.time() + time_add, value
return value
return wrapper
return _temp
@time_cache("function_definition_validity")
def cache_function_definition(stmt):
module_path = stmt.get_parent_until().path
return None if module_path is None else (module_path, stmt.start_pos)
def cache_star_import(func):
def wrapper(scope, *args, **kwargs):
with common.ignored(KeyError):
mods = star_import_cache[scope]
if mods[0] + settings.star_import_cache_validity > time.time():
return mods[1]
# cache is too old and therefore invalid or not available
invalidate_star_import_cache(scope)
mods = func(scope, *args, **kwargs)
star_import_cache[scope] = time.time(), mods
return mods
return wrapper
def invalidate_star_import_cache(module, only_main=False):
""" Important if some new modules are being reparsed """
with common.ignored(KeyError):
t, mods = star_import_cache[module]
del star_import_cache[module]
for m in mods:
invalidate_star_import_cache(m, only_main=True)
if not only_main:
# We need a list here because otherwise the list is being changed
# during the iteration in py3k: iteritems -> items.
for key, (t, mods) in list(star_import_cache.items()):
if module in mods:
invalidate_star_import_cache(key)
def load_module(path, name):
"""
Returns the module or None, if it fails.
"""
if path is None and name is None:
return None
tim = os.path.getmtime(path) if path else None
n = name if path is None else path
try:
parser_cache_item = parser_cache[n]
if not path or tim <= parser_cache_item.change_time:
return parser_cache_item.parser
else:
# In case there is already a module cached and this module
# has to be reparsed, we also need to invalidate the import
# caches.
invalidate_star_import_cache(parser_cache_item.parser.module)
except KeyError:
if settings.use_filesystem_cache:
return ModulePickling.load_module(n, tim)
def save_module(path, name, parser, pickling=True):
try:
p_time = None if not path else os.path.getmtime(path)
except OSError:
p_time = None
pickling = False
n = name if path is None else path
item = ParserCacheItem(parser, p_time)
parser_cache[n] = item
if settings.use_filesystem_cache and pickling:
ModulePickling.save_module(n, item)
class _ModulePickling(object):
version = 3
"""
Version number (integer) for file system cache.
Increment this number when there are any incompatible changes in
parser representation classes. For example, the following changes
are regarded as incompatible.
- Class name is changed.
- Class is moved to another module.
- Defined slot of the class is changed.
"""
def __init__(self):
self.__index = None
self.py_tag = 'cpython-%s%s' % sys.version_info[:2]
"""
Short name for distinguish Python implementations and versions.
It's like `sys.implementation.cache_tag` but for Python < 3.3
we generate something similar. See:
http://docs.python.org/3/library/sys.html#sys.implementation
.. todo:: Detect interpreter (e.g., PyPy).
"""
def load_module(self, path, original_changed_time):
try:
pickle_changed_time = self._index[path]
except KeyError:
return None
if original_changed_time is not None \
and pickle_changed_time < original_changed_time:
# the pickle file is outdated
return None
with open(self._get_hashed_path(path), 'rb') as f:
parser_cache_item = pickle.load(f)
debug.dbg('pickle loaded', path)
parser_cache[path] = parser_cache_item
return parser_cache_item.parser
def save_module(self, path, parser_cache_item):
self.__index = None
try:
files = self._index
except KeyError:
files = {}
self._index = files
with open(self._get_hashed_path(path), 'wb') as f:
pickle.dump(parser_cache_item, f, pickle.HIGHEST_PROTOCOL)
files[path] = parser_cache_item.change_time
self._flush_index()
@property
def _index(self):
if self.__index is None:
try:
with open(self._get_path('index.json')) as f:
data = json.load(f)
except (IOError, ValueError):
self.__index = {}
else:
# 0 means version is not defined (= always delete cache):
if data.get('version', 0) != self.version:
self.delete_cache()
self.__index = {}
else:
self.__index = data['index']
return self.__index
def _remove_old_modules(self):
# TODO use
change = False
if change:
self._flush_index(self)
self._index # reload index
def _flush_index(self):
data = {'version': self.version, 'index': self._index}
with open(self._get_path('index.json'), 'w') as f:
json.dump(data, f)
self.__index = None
def delete_cache(self):
shutil.rmtree(self._cache_directory())
def _get_hashed_path(self, path):
return self._get_path('%s.pkl' % hashlib.md5(path.encode("utf-8")).hexdigest())
def _get_path(self, file):
dir = self._cache_directory()
if not os.path.exists(dir):
os.makedirs(dir)
return os.path.join(dir, file)
def _cache_directory(self):
return os.path.join(settings.cache_directory, self.py_tag)
# is a singleton
ModulePickling = _ModulePickling()
|