404 lines
15 KiB
Python
404 lines
15 KiB
Python
# Copyright (C) 2015-2023 RhodeCode GmbH
|
||
#
|
||
# This program is free software: you can redistribute it and/or modify
|
||
# it under the terms of the GNU Affero General Public License, version 3
|
||
# (only), as published by the Free Software Foundation.
|
||
#
|
||
# This program is distributed in the hope that it will be useful,
|
||
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
||
# GNU General Public License for more details.
|
||
#
|
||
# You should have received a copy of the GNU Affero General Public License
|
||
# along with this program. If not, see <http://www.gnu.org/licenses/>.
|
||
#
|
||
# This program is dual-licensed. If you wish to learn more about the
|
||
# RhodeCode Enterprise Edition, including its added features, Support services,
|
||
# and proprietary license terms, please see https://rhodecode.com/licenses/
|
||
|
||
import functools
|
||
import logging
|
||
import os
|
||
import threading
|
||
import time
|
||
|
||
import decorator
|
||
from dogpile.cache import CacheRegion
|
||
|
||
import rhodecode
|
||
from rhodecode.lib.hash_utils import sha1
|
||
from rhodecode.lib.str_utils import safe_bytes
|
||
from rhodecode.lib.type_utils import str2bool # noqa :required by imports from .utils
|
||
|
||
from . import region_meta, cache_key_meta
|
||
|
||
log = logging.getLogger(__name__)
|
||
|
||
|
||
def isCython(func):
|
||
"""
|
||
Private helper that checks if a function is a cython function.
|
||
"""
|
||
return func.__class__.__name__ == 'cython_function_or_method'
|
||
|
||
|
||
class RhodeCodeCacheRegion(CacheRegion):
|
||
|
||
def __repr__(self):
|
||
return f'{self.__class__}(name={self.name})'
|
||
|
||
def conditional_cache_on_arguments(
|
||
self, namespace=None,
|
||
expiration_time=None,
|
||
should_cache_fn=None,
|
||
to_str=str,
|
||
function_key_generator=None,
|
||
condition=True):
|
||
"""
|
||
Custom conditional decorator, that will not touch any dogpile internals if
|
||
condition isn't meet. This works a bit different from should_cache_fn
|
||
And it's faster in cases we don't ever want to compute cached values
|
||
"""
|
||
expiration_time_is_callable = callable(expiration_time)
|
||
if not namespace:
|
||
namespace = getattr(self, '_default_namespace', None)
|
||
|
||
if function_key_generator is None:
|
||
function_key_generator = self.function_key_generator
|
||
|
||
def get_or_create_for_user_func(func_key_generator, user_func, *arg, **kw):
|
||
|
||
if not condition:
|
||
log.debug('Calling un-cached method:%s', user_func.__name__)
|
||
start = time.time()
|
||
result = user_func(*arg, **kw)
|
||
total = time.time() - start
|
||
log.debug('un-cached method:%s took %.4fs', user_func.__name__, total)
|
||
return result
|
||
|
||
key = func_key_generator(*arg, **kw)
|
||
|
||
timeout = expiration_time() if expiration_time_is_callable \
|
||
else expiration_time
|
||
|
||
log.debug('Calling cached method:`%s`', user_func.__name__)
|
||
return self.get_or_create(key, user_func, timeout, should_cache_fn, (arg, kw))
|
||
|
||
def cache_decorator(user_func):
|
||
if to_str is str:
|
||
# backwards compatible
|
||
key_generator = function_key_generator(namespace, user_func)
|
||
else:
|
||
key_generator = function_key_generator(namespace, user_func, to_str=to_str)
|
||
|
||
def refresh(*arg, **kw):
|
||
"""
|
||
Like invalidate, but regenerates the value instead
|
||
"""
|
||
key = key_generator(*arg, **kw)
|
||
value = user_func(*arg, **kw)
|
||
self.set(key, value)
|
||
return value
|
||
|
||
def invalidate(*arg, **kw):
|
||
key = key_generator(*arg, **kw)
|
||
self.delete(key)
|
||
|
||
def set_(value, *arg, **kw):
|
||
key = key_generator(*arg, **kw)
|
||
self.set(key, value)
|
||
|
||
def get(*arg, **kw):
|
||
key = key_generator(*arg, **kw)
|
||
return self.get(key)
|
||
|
||
user_func.set = set_
|
||
user_func.invalidate = invalidate
|
||
user_func.get = get
|
||
user_func.refresh = refresh
|
||
user_func.key_generator = key_generator
|
||
user_func.original = user_func
|
||
|
||
# Use `decorate` to preserve the signature of :param:`user_func`.
|
||
return decorator.decorate(user_func, functools.partial(
|
||
get_or_create_for_user_func, key_generator))
|
||
|
||
return cache_decorator
|
||
|
||
|
||
def make_region(*arg, **kw):
|
||
return RhodeCodeCacheRegion(*arg, **kw)
|
||
|
||
|
||
def get_default_cache_settings(settings, prefixes=None):
|
||
prefixes = prefixes or []
|
||
cache_settings = {}
|
||
for key in settings.keys():
|
||
for prefix in prefixes:
|
||
if key.startswith(prefix):
|
||
name = key.split(prefix)[1].strip()
|
||
val = settings[key]
|
||
if isinstance(val, str):
|
||
val = val.strip()
|
||
cache_settings[name] = val
|
||
return cache_settings
|
||
|
||
|
||
def compute_key_from_params(*args):
|
||
"""
|
||
Helper to compute key from given params to be used in cache manager
|
||
"""
|
||
return sha1(safe_bytes("_".join(map(str, args))))
|
||
|
||
|
||
def custom_key_generator(backend, namespace, fn):
|
||
func_name = fn.__name__
|
||
|
||
def generate_key(*args):
|
||
backend_pref = getattr(backend, 'key_prefix', None) or 'backend_prefix'
|
||
namespace_pref = namespace or 'default_namespace'
|
||
arg_key = compute_key_from_params(*args)
|
||
final_key = f"{backend_pref}:{namespace_pref}:{func_name}_{arg_key}"
|
||
|
||
return final_key
|
||
|
||
return generate_key
|
||
|
||
|
||
def backend_key_generator(backend):
|
||
"""
|
||
Special wrapper that also sends over the backend to the key generator
|
||
"""
|
||
def wrapper(namespace, fn):
|
||
return custom_key_generator(backend, namespace, fn)
|
||
return wrapper
|
||
|
||
|
||
def get_or_create_region(region_name, region_namespace: str = None, use_async_runner=False):
|
||
from .backends import FileNamespaceBackend
|
||
from . import async_creation_runner
|
||
|
||
region_obj = region_meta.dogpile_cache_regions.get(region_name)
|
||
if not region_obj:
|
||
reg_keys = list(region_meta.dogpile_cache_regions.keys())
|
||
raise OSError(f'Region `{region_name}` not in configured: {reg_keys}.')
|
||
|
||
region_uid_name = f'{region_name}:{region_namespace}'
|
||
|
||
# Special case for ONLY the FileNamespaceBackend backend. We register one-file-per-region
|
||
if isinstance(region_obj.actual_backend, FileNamespaceBackend):
|
||
if not region_namespace:
|
||
raise ValueError(f'{FileNamespaceBackend} used requires to specify region_namespace param')
|
||
|
||
region_exist = region_meta.dogpile_cache_regions.get(region_namespace)
|
||
if region_exist:
|
||
log.debug('Using already configured region: %s', region_namespace)
|
||
return region_exist
|
||
|
||
expiration_time = region_obj.expiration_time
|
||
|
||
cache_dir = region_meta.dogpile_config_defaults['cache_dir']
|
||
namespace_cache_dir = cache_dir
|
||
|
||
# we default the namespace_cache_dir to our default cache dir.
|
||
# however, if this backend is configured with filename= param, we prioritize that
|
||
# so all caches within that particular region, even those namespaced end up in the same path
|
||
if region_obj.actual_backend.filename:
|
||
namespace_cache_dir = os.path.dirname(region_obj.actual_backend.filename)
|
||
|
||
if not os.path.isdir(namespace_cache_dir):
|
||
os.makedirs(namespace_cache_dir)
|
||
new_region = make_region(
|
||
name=region_uid_name,
|
||
function_key_generator=backend_key_generator(region_obj.actual_backend)
|
||
)
|
||
|
||
namespace_filename = os.path.join(
|
||
namespace_cache_dir, f"{region_name}_{region_namespace}.cache_db")
|
||
# special type that allows 1db per namespace
|
||
new_region.configure(
|
||
backend='dogpile.cache.rc.file_namespace',
|
||
expiration_time=expiration_time,
|
||
arguments={"filename": namespace_filename}
|
||
)
|
||
|
||
# create and save in region caches
|
||
log.debug('configuring new region: %s', region_uid_name)
|
||
region_obj = region_meta.dogpile_cache_regions[region_namespace] = new_region
|
||
|
||
region_obj._default_namespace = region_namespace
|
||
if use_async_runner:
|
||
region_obj.async_creation_runner = async_creation_runner
|
||
return region_obj
|
||
|
||
|
||
def clear_cache_namespace(cache_region: str | RhodeCodeCacheRegion, cache_namespace_uid: str, method: str) -> int:
|
||
from . import CLEAR_DELETE, CLEAR_INVALIDATE
|
||
|
||
if not isinstance(cache_region, RhodeCodeCacheRegion):
|
||
cache_region = get_or_create_region(cache_region, cache_namespace_uid)
|
||
log.debug('clearing cache region: %s [prefix:%s] with method=%s',
|
||
cache_region, cache_namespace_uid, method)
|
||
|
||
num_affected_keys = 0
|
||
|
||
if method == CLEAR_INVALIDATE:
|
||
# NOTE: The CacheRegion.invalidate() method’s default mode of
|
||
# operation is to set a timestamp local to this CacheRegion in this Python process only.
|
||
# It does not impact other Python processes or regions as the timestamp is only stored locally in memory.
|
||
cache_region.invalidate(hard=True)
|
||
|
||
if method == CLEAR_DELETE:
|
||
num_affected_keys = cache_region.backend.delete_multi_by_prefix(prefix=cache_namespace_uid)
|
||
|
||
return num_affected_keys
|
||
|
||
|
||
class ActiveRegionCache(object):
|
||
def __init__(self, context, cache_data):
|
||
self.context = context
|
||
self.cache_data = cache_data
|
||
|
||
def should_invalidate(self):
|
||
return False
|
||
|
||
|
||
class FreshRegionCache(object):
|
||
def __init__(self, context, cache_data):
|
||
self.context = context
|
||
self.cache_data = cache_data
|
||
|
||
def should_invalidate(self):
|
||
return True
|
||
|
||
|
||
class InvalidationContext(object):
|
||
"""
|
||
usage::
|
||
|
||
from rhodecode.lib import rc_cache
|
||
|
||
cache_namespace_uid = CacheKey.SOME_NAMESPACE.format(1)
|
||
region = rc_cache.get_or_create_region('some_region', cache_namespace_uid)
|
||
|
||
@region.conditional_cache_on_arguments(namespace=cache_namespace_uid, condition=True)
|
||
def heavy_compute(cache_name, param1, param2):
|
||
print('COMPUTE {}, {}, {}'.format(cache_name, param1, param2))
|
||
|
||
# invalidation namespace is shared namespace key for all process caches
|
||
# we use it to send a global signal
|
||
invalidation_namespace = 'repo_cache:1'
|
||
|
||
inv_context_manager = rc_cache.InvalidationContext(
|
||
uid=cache_namespace_uid, invalidation_namespace=invalidation_namespace)
|
||
with inv_context_manager as invalidation_context:
|
||
args = ('one', 'two')
|
||
# re-compute and store cache if we get invalidate signal
|
||
if invalidation_context.should_invalidate():
|
||
result = heavy_compute.refresh(*args)
|
||
else:
|
||
result = heavy_compute(*args)
|
||
|
||
compute_time = inv_context_manager.compute_time
|
||
log.debug('result computed in %.4fs', compute_time)
|
||
|
||
# To send global invalidation signal, simply run
|
||
CacheKey.set_invalidate(invalidation_namespace)
|
||
|
||
"""
|
||
|
||
def __repr__(self):
|
||
return f'<InvalidationContext:{self.cache_key}[{self.uid}]>'
|
||
|
||
def __init__(self, uid, invalidation_namespace='',
|
||
raise_exception=False, thread_scoped=None):
|
||
self.uid = uid
|
||
self.invalidation_namespace = invalidation_namespace
|
||
self.raise_exception = raise_exception
|
||
self.proc_id = rhodecode.CONFIG.get('instance_id') or 'DEFAULT'
|
||
self.thread_id = 'global'
|
||
|
||
if thread_scoped is None:
|
||
# if we set "default" we can override this via .ini settings
|
||
thread_scoped = rhodecode.ConfigGet().get_bool('cache_thread_scoped')
|
||
|
||
# Append the thread id to the cache key if this invalidation context
|
||
# should be scoped to the current thread.
|
||
if thread_scoped is True:
|
||
self.thread_id = threading.current_thread().ident
|
||
|
||
self.cache_key = compute_key_from_params(uid)
|
||
self.cache_key = 'proc:{}|thread:{}|params:{}'.format(
|
||
self.proc_id, self.thread_id, self.cache_key)
|
||
self.proc_key = f'proc:{self.proc_id}'
|
||
self.compute_time = 0
|
||
|
||
def get_or_create_cache_obj(self, cache_type, invalidation_namespace=''):
|
||
from rhodecode.model.db import CacheKey
|
||
|
||
invalidation_namespace = invalidation_namespace or self.invalidation_namespace
|
||
# fetch all cache keys for this namespace and convert them to a map to find if we
|
||
# have specific cache_key object registered. We do this because we want to have
|
||
# all consistent cache_state_uid for newly registered objects
|
||
cache_obj_map = CacheKey.get_namespace_map(invalidation_namespace)
|
||
cache_obj = cache_obj_map.get(self.cache_key)
|
||
log.debug('Fetched cache obj %s using %s cache key.', cache_obj, self.cache_key)
|
||
|
||
if not cache_obj:
|
||
new_cache_args = invalidation_namespace
|
||
first_cache_obj = next(iter(cache_obj_map.values())) if cache_obj_map else None
|
||
cache_state_uid = None
|
||
if first_cache_obj:
|
||
cache_state_uid = first_cache_obj.cache_state_uid
|
||
cache_obj = CacheKey(self.cache_key, cache_args=new_cache_args,
|
||
cache_state_uid=cache_state_uid)
|
||
cache_key_meta.cache_keys_by_pid.add(self.proc_key)
|
||
|
||
return cache_obj
|
||
|
||
def __enter__(self):
|
||
"""
|
||
Test if current object is valid, and return CacheRegion function
|
||
that does invalidation and calculation
|
||
"""
|
||
log.debug('Entering cache invalidation check context: %s', self.invalidation_namespace)
|
||
# register or get a new key based on uid
|
||
self.cache_obj = self.get_or_create_cache_obj(cache_type=self.uid)
|
||
cache_data = self.cache_obj.get_dict()
|
||
self._start_time = time.time()
|
||
if self.cache_obj.cache_active:
|
||
# means our cache obj is existing and marked as it's
|
||
# cache is not outdated, we return ActiveRegionCache
|
||
self.skip_cache_active_change = True
|
||
|
||
return ActiveRegionCache(context=self, cache_data=cache_data)
|
||
|
||
# the key is either not existing or set to False, we return
|
||
# the real invalidator which re-computes value. We additionally set
|
||
# the flag to actually update the Database objects
|
||
self.skip_cache_active_change = False
|
||
return FreshRegionCache(context=self, cache_data=cache_data)
|
||
|
||
def __exit__(self, exc_type, exc_val, exc_tb):
|
||
from rhodecode.model.db import IntegrityError, Session
|
||
|
||
# save compute time
|
||
self.compute_time = time.time() - self._start_time
|
||
|
||
if self.skip_cache_active_change:
|
||
return
|
||
|
||
try:
|
||
self.cache_obj.cache_active = True
|
||
Session().add(self.cache_obj)
|
||
Session().commit()
|
||
except IntegrityError:
|
||
# if we catch integrity error, it means we inserted this object
|
||
# assumption is that's really an edge race-condition case and
|
||
# it's safe is to skip it
|
||
Session().rollback()
|
||
except Exception:
|
||
log.exception('Failed to commit on cache key update')
|
||
Session().rollback()
|
||
if self.raise_exception:
|
||
raise
|