Unified cache handling interface
version 0.58
use CHI;
# Choose a standard driver
#
my $cache = CHI->new( driver => 'Memory', global => 1 );
my $cache = CHI->new( driver => 'RawMemory', global => 1 );
my $cache = CHI->new( driver => 'File',
root_dir => '/path/to/root'
);
my $cache = CHI->new( driver => 'FastMmap',
root_dir => '/path/to/root',
cache_size => '1k'
);
my $cache = CHI->new( driver => 'Memcached::libmemcached',
servers => [ "10.0.0.15:11211", "10.0.0.15:11212" ],
l1_cache => { driver => 'FastMmap', root_dir => '/path/to/root' }
);
my $cache = CHI->new( driver => 'DBI',
dbh => $dbh
);
my $cache = CHI->new( driver => 'BerkeleyDB',
root_dir => '/path/to/root'
);
# Create your own driver
#
my $cache = CHI->new( driver => '+My::Special::Driver', ... );
# Cache operations
#
my $customer = $cache->get($name);
if ( !defined $customer ) {
$customer = get_customer_from_db($name);
$cache->set( $name, $customer, "10 minutes" );
}
my $customer2 = $cache->compute($name2, "10 minutes", sub {
get_customer_from_db($name2)
});
$cache->remove($name);
\s-1CHI\s0 provides a unified caching \s-1API\s0, designed to assist a developer in persisting data for a specified period of time.
The \s-1CHI\s0 interface is implemented by driver classes that support fetching, storing and clearing of data. Driver classes exist or will exist for the gamut of storage backends available to Perl, such as memory, plain files, memory mapped files, memcached, and \s-1DBI\s0.
\s-1CHI\s0 is intended as an evolution of DeWitt Clinton's Cache::Cache package, adhering to the basic Cache \s-1API\s0 but adding new features and addressing limitations in the Cache::Cache implementation.
Easy to create new drivers
Uniform support for namespaces
Automatic serialization of keys and values
Multilevel caches
Probabilistic expiration and busy locks, to reduce cache miss stampedes
Optional logging and statistics collection of cache activity
To create a new cache object, call \*(C`<CHI->new\*(C'. It takes the common options listed below. driver is required; all others are optional.
Some drivers will take additional constructor options. For example, the File driver takes \*(C`root_dir\*(C' and \*(C`depth\*(C' options.
You can configure default options for each new cache object created - see \*(L"\s-1SUBCLASSING\s0 \s-1AND\s0 \s-1CONFIGURING\s0 \s-1CHI\s0\*(R".
Note that \*(C`CHI->new\*(C' returns an instance of a subclass of CHI::Driver, not \*(C`CHI\*(C'.
A value in bytes. Automatically compress values larger than this before storing. Requires Compress::Zlib to be installed. Defaults to undef, meaning no automatic compression. Inspired by the parameter of the same name in Cache::Memcached. # Compress values larger than 1MB compress_threshold => 1024*1024
Required. The name of a cache driver, for example \*(L"Memory\*(R" or \*(L"File\*(R". \s-1CHI\s0 will prefix the string with \*(L"CHI::Driver::\*(R", unless it begins with '+'. e.g. driver => 'File'; # uses CHI::Driver::File driver => '+My::CHI::Driver::File' # uses My::CHI::Driver::File
Provide default values for the corresponding \*(L"set\*(R" options.
If set to 0 (the default), \s-1CHI\s0 alone is aware of the expiration time and does not pass it along to the backend driver. This allows you to use \*(L"get_object\*(R" to retrieve expired items. If set to 1, pass expiration times to backend driver if the driver supports it \*(-- for example, CHI::Driver::Memcached and CHI::Driver::CacheCache. This may allow the driver to better manage its space and evict items. Note that only simple expiration time will be passed along, e.g. not \*(L"expires_variance\*(R". If set to a number greater than 1 (e.g. 1.25), the time until expiration will be multiplied by that number before being passed to the backend driver. This gives you a customizable window of opportunity to retrieve expired items.
Digest algorithm to use on keys longer than \*(L"max_key_length\*(R" - e.g. \*(L"\s-1MD5\s0\*(R", \*(L"\s-1SHA-1\s0\*(R", or \*(L"\s-1SHA-256\s0\*(R". Can be a Digest object, or a string or hashref which will passed to Digest->new(). You will need to ensure Digest is installed to use these options. Default is \*(L"\s-1MD5\s0\*(R".
An object to use for serializing keys that are references. See \*(L"serializer\*(R" above for the different ways this can be passed in. The default is to use \s-1JSON\s0 in canonical mode (sorted hash keys).
A label for the cache as a whole, independent of namespace - e.g. \*(L"web-file-cache\*(R". Used when referring to the cache in logs, statistics, and error messages. By default, set to \*(L"short_driver_name\*(R".
Add an L1 cache as a subcache. See \*(L"\s-1SUBCACHES\s0\*(R".
Keys over this size will be digested. The default is driver-specific; CHI::Driver::File, for example, defaults this to 240 due to file system limits. For most drivers there is no maximum.
Add an mirror cache as a subcache. See \*(L"\s-1SUBCACHES\s0\*(R".
Identifies a namespace that all cache entries for this object will be in. This allows easy separation of multiple, distinct caches without worrying about key collision. Suggestions for easy namespace selection:
In a class, use the class name: my $cache = CHI->new(namespace => _\|_PACKAGE_\|_, ...);
In a script, use the script's absolute path name: use Cwd qw(realpath); my $cache = CHI->new(namespace => realpath($0), ...);
In a web template, use the template name. For example, in Mason, $m->cache will set the namespace to the current component path.
Defaults to 'Default' if not specified.
How to handle runtime errors occurring during cache gets and cache sets, which may or may not be considered fatal in your application. Options are:
log (the default) - log an error, or ignore if no logger is set - see \*(L"\s-1LOGGING\s0\*(R"
ignore - do nothing
warn - call warn() with an appropriate message
die - call die() with an appropriate message
coderef - call this code reference with three arguments: an appropriate message, the key, and the original raw error message
An object to use for serializing data before storing it in the cache, and deserializing data after retrieving it from the cache. Only references will be serialized; plain scalars will be placed in the cache as-is. If this is a string, a Data::Serializer object will be created, with the string passed as the 'serializer' option and raw=1. Common options include 'Storable', 'Data::Dumper', and '\s-1YAML\s0'. If this is a hashref, Data::Serializer will be called with the hash. You will need to ensure Data::Serializer is installed to use these options. Otherwise, this must be a Data::Serializer object or another object that implements serialize() and deserialize(). e.g. # Serialize using raw Data::Dumper my $cache = CHI->new(serializer => 'Data::Dumper');
# Serialize using Data::Dumper, compressed and (per Data::Serializer defaults) hex-encoded my $cache = CHI->new(serializer => { serializer => 'Data::Dumper', compress => 1 });
# Serialize using custom object my $cache = CHI->new(serializer => My::Custom::Serializer->new()) The default is to use raw Storable.
List of one or more roles to apply to the \*(C`CHI::Driver\*(C' class that is constructed. The roles will automatically be prefixed with \*(C`CHI::Driver::Role::\*(C' unless preceded with a '+'. e.g. traits => ['StoresAccessedAt', '+My::CHI::Driver::Role']
The following methods can be called on any cache handle returned from \s-1CHI-\s0>new(). They are implemented in the CHI::Driver package.
Returns the data associated with $key. If $key does not exist or has expired, returns undef. Expired items are not automatically removed and may be examined with \*(L"get_object\*(R" or \*(L"get_expires_at\*(R". $key may be followed by one or more name/value parameters:
If $key exists and has not expired, call code reference with the CHI::CacheObject as a single parameter. If code returns a true value, \*(C`get\*(C' returns undef as if the item were expired. For example, to treat the cache as expired if $file has changed since the value was computed: $cache->get('foo', expire_if => sub { $_[0]->created_at < (stat($file))[9] });
If the value has expired, the get will still return undef, but the expiration time of the cache entry will be set to the current time plus the specified duration. This is used to prevent multiple processes from recomputing the same expensive value simultaneously. The problem with this technique is that it doubles the number of writes performed - see \*(L"expires_variance\*(R" for another technique.
If the item exists in cache (even if expired), place the CHI::CacheObject object in the provided \s-1SCALARREF\s0.
Associates $data with $key in the cache, overwriting any existing entry. Returns $data. The third argument to \*(C`set\*(C' is optional, and may be either a scalar or a hash reference. If it is a scalar, it may be the string \*(L"now\*(R", the string \*(L"never\*(R", or else a duration treated as an expires_in value described below. If it is a hash reference, it may contain one or more of the following options. Most of these options can be provided with defaults in the cache constructor.
Amount of time from now until this data expires. \s-1DURATION\s0 may be an integer number of seconds or a duration expression.
The epoch time at which the data expires.
Controls the variable expiration feature, which allows items to expire a little earlier than the stated expiration time to help prevent cache miss stampedes. Value is between 0.0 and 1.0, with 0.0 meaning that items expire exactly when specified (feature is disabled), and 1.0 meaning that items might expire anytime from now until the stated expiration time. The default is 0.0. A setting of 0.10 to 0.25 would introduce a small amount of variation without interfering too much with intended expiration times. The probability of expiration increases as a function of how far along we are in the potential expiration window, with the probability being near 0 at the beginning of the window and approaching 1 at the end. For example, in all of the following cases, an item might be considered expired any time between 15 and 20 minutes, with about a 20% chance at 16 minutes, a 40% chance at 17 minutes, and a 100% chance at 20 minutes. my $cache = CHI->new ( ..., expires_variance => 0.25, ... ); $cache->set($key, $value, '20 min'); $cache->set($key, $value, { expires_at => time() + 20*60 });
my $cache = CHI->new ( ... ); $cache->set($key, $value, { expires_in => '20 min', expires_variance => 0.25 }); \s-1CHI\s0 will make a new probabilistic choice every time it needs to know whether an item has expired (i.e. it does not save the results of its determination), so you can get situations like this: my $value = $cache->get($key); # returns undef (indicating expired) my $value = $cache->get($key); # returns valid value this time!
if ($cache->is_valid($key)) # returns undef (indicating expired) if ($cache->is_valid($key)) # returns true this time! Typical applications won't be affected by this, since the object is recomputed as soon as it is determined to be expired. But it's something to be aware of.
Combines the \*(C`get\*(C' and \*(C`set\*(C' operations in a single call. Attempts to get $key; if successful, returns the value. Otherwise, calls $code and uses the return value as the new value for $key, which is then returned. Caller context (scalar or list) is respected. $options can be undef, a scalar, or a hash reference. If it is undef, it has no effect. If it is a scalar, it is treated as the \*(C`expires_in\*(C' duration and passed as the third argument to \*(C`set\*(C'. If it is a hash reference, it may contain name/value pairs for both \*(C`get\*(C' and \*(C`set\*(C'. e.g. # No expiration my $value = $cache->compute($key, undef, sub { # compute and return value for $key here });
# Expire in 5 minutes my $value = $cache->compute($key, '5min', sub { # compute and return value for $key here });
# Expire in 5 minutes or when a particular condition occurs my $value = $cache->compute($key, { expires_in => '5min', expire_if => sub { ... } }, sub { # compute and return value for $key here });
# List context my @value = $cache->compute($key, '5min', sub { ... return @some_list; }); This method will eventually support the ability to recompute a value in the background just before it actually expires, so that users are not impacted by recompute time. Note: Prior to version 0.40, the last two arguments were in reverse order; both will be accepted for backward compatibility. We think the coderef looks better at the end.
Remove the data associated with the $key from the cache. If $key exists, expire it by setting its expiration time into the past. Does not necessarily remove the data. Since this involves essentially setting the value again, \*(C`remove\*(C' may be more efficient for some drivers.
Returns a boolean indicating whether $key exists in the cache and has not expired. Note: Expiration may be determined probabilistically if \*(L"expires_variance\*(R" was used. Returns a boolean indicating whether $key exists in the cache and has expired. Note: Expiration may be determined probabilistically if \*(L"expires_variance\*(R" was used. Returns the epoch time at which $key definitively expires. Returns undef if the key does not exist or it has no expiration time. Returns a CHI::CacheObject object containing data about the entry associated with $key, or undef if no such key exists. The object will be returned even if the entry has expired, as long as it has not been removed.
These methods combine both reading and writing of a cache entry in a single operation. The names and behaviors were adapted from memcached <http://memcached.org/>.
Some drivers (e.g. CHI::Driver::Memcached::libmemcached, CHI::Driver::DBI) may implement these as truly atomic operations, and will be documented thusly. The default implementations are not atomic: the get and set occur discretely and another process could potentially modify the cache in between them.
These operations are labeled \s-1ALPHA\s0 because we haven't yet figured out how they integrate with other \s-1CHI\s0 features, in particular \*(L"\s-1SUBCACHES\s0\*(R". APIs and behavior may change. Do a set, but only if $key is not valid in the cache. Do a set, but only if $key is valid in the cache. Append $new_data to whatever value is currently associated with $key. Has no effect if $key does not exist in the cache. Returns true if $key was in the cache, false otherwise. This is intended for simple string values only. For efficiency's sake, \s-1CHI\s0 won't attempt to check for, or handle, the case where data is serialized or compressed; the new data will simply be appended, and an error will most probably occur when you try to retrieve the value. Does not modify expiration or other metadata. If $key exists but is expired, it will remain expired. If you use a driver with the non-atomic (default) implementation, some appends may be lost due to race conditions.
Remove all entries from the namespace.
Returns a list of keys in the namespace. This may or may not include expired keys, depending on the driver. The keys may not look the same as they did when passed into \*(L"set\*(R"; they may have been serialized, utf8 encoded, and/or digested (see \*(L"\s-1KEY\s0 \s-1AND\s0 \s-1VALUE\s0 \s-1TRANSFORMATIONS\s0\*(R"). However, they may still be passed back into \*(L"get\*(R", \*(L"set\*(R", etc. to access the same underlying objects. i.e. the following code is guaranteed to produce all key/value pairs from the cache: map { ($_, $c->get($_)) } $c->get_keys()
Remove all entries that have expired from the namespace associated with this cache instance. Warning: May be very inefficient, depending on the number of keys and the driver.
Returns a list of namespaces associated with the cache. This may or may not include empty namespaces, depending on the driver.
The methods in this section process multiple keys and/or values at once. By default these are implemented with the obvious map operations, but some cache drivers (e.g. Cache::Memcached) can override them with more efficient implementations. Get the keys in list reference $keys, and return a list reference of the same length with corresponding values or undefs. Like \*(L"get_multi_arrayref\*(R", but returns a hash reference with each key in $keys mapping to its corresponding value or undef. Will only work with scalar keys. Set the multiple keys and values provided in hash reference $key_values. $set_options is a scalar or hash reference, used as the third argument to set. Will only work with scalar keys. Removes the keys in list reference $keys.
Returns a hash reference containing all the non-expired keys and values in the cache.
Returns the name of the root class under which this object was created, e.g. \*(C`CHI\*(C' or \*(C`My::CHI\*(C'. See \*(L"\s-1SUBCLASSING\s0 \s-1AND\s0 \s-1CONFIGURING\s0 \s-1CHI\s0\*(R".
Returns the full name of the driver class. e.g. CHI->new(driver=>'File')->driver_class => CHI::Driver::File CHI->new(driver=>'+CHI::Driver::File')->driver_class => CHI::Driver::File CHI->new(driver=>'+My::Driver::File')->driver_class => My::Driver::File You should use this rather than \*(C`ref()\*(C'. Due to some subclassing tricks \s-1CHI\s0 employs, the actual class of the object is neither guaranteed nor likely to be the driver class.
Returns the name of the driver class, minus the CHI::Driver:: prefix, if any. e.g. CHI->new(driver=>'File')->short_driver_name => File CHI->new(driver_class=>'CHI::Driver::File')->short_driver_name => File CHI->new(driver_class=>'My::Driver::File')->short_driver_name => My::Driver::File
expires_in expires_at expires_variance label on_get_error on_set_error
namespace serializer
The following methods are deprecated and will be removed in a later version:
is_empty
Duration expressions, which appear in the \*(L"set\*(R" command and various other parts of the \s-1API\s0, are parsed by Time::Duration::Parse. A duration is either a plain number, which is treated like a number of seconds, or a number and a string representing time units where the string is one of:
s second seconds sec secs m minute minutes min mins h hr hour hours d day days w week weeks M month months y year years
e.g. the following are all valid duration expressions:
25 3s 5 seconds 1 minute and ten seconds 1 hour
\s-1CHI\s0 strives to accept arbitrary keys and values for caching regardless of the limitations of the underlying driver.
Keys that are references are serialized - see \*(L"key_serializer\*(R".
Keys with wide (>255) characters are utf8 encoded.
Keys exceeding the maximum length for the underlying driver are digested - see \*(L"max_key_length\*(R" and \*(L"key_digester\*(R".
For some drivers (e.g. CHI::Driver::File), keys containing special characters or whitespace are escaped with URL-like escaping.
Note: All transformations above with the exception of escaping are one-way, meaning that \s-1CHI\s0 does not attempt to undo them when returned from \*(L"get_keys\*(R"; and idempotent, meaning that applying them a second time has no effect. So when you call \*(L"get_keys\*(R", the key you get may not be exactly what you passed in, but you'll be able to pass that key in to get the corresponding object.
Values which are references are automatically serialized before storing, and deserialized after retrieving - see \*(L"serializer\*(R".
Values with their utf8 flag on are utf8 encoded before storing, and utf8 decoded after retrieving.
It is possible to a cache to have one or more subcaches. There are currently two types of subcaches: L1 and mirror.
An L1 (or \*(L"level one\*(R") cache sits in front of the primary cache, usually to provide faster access for commonly accessed cache entries. For example, this places an in-process Memory cache in front of a Memcached cache:
my $cache = CHI->new( driver => 'Memcached', servers => [ "10.0.0.15:11211", "10.0.0.15:11212" ], l1_cache => { driver => 'Memory', global => 1, max_size => 1024*1024 } );
On a \*(C`get\*(C', the L1 cache is checked first - if a valid value exists, it is returned. Otherwise, the primary cache is checked - if a valid value exists, it is returned, and the value is placed in the L1 cache with the same expiration time. In this way, items fetched most frequently from the primary cache will tend to be in the L1 cache.
\*(C`set\*(C' operations are distributed to both the primary and L1 cache.
You can access the L1 cache with the \*(C`l1_cache\*(C' method. For example, this clears the L1 cache but leaves the primary cache intact:
$cache->l1_cache->clear();
A mirror cache is a write-only cache that, over time, mirrors the content of the primary cache. \*(C`set\*(C' operations are distributed to both the primary and mirror cache, but \*(C`get\*(C' operations go only to the primary cache.
Mirror caches are useful when you want to migrate from one cache to another. You can populate a mirror cache and switch over to it once it is sufficiently populated. For example, here we migrate from an old to a new cache directory:
my $cache = CHI->new( driver => 'File', root_dir => '/old/cache/root', mirror_cache => { driver => 'File', root_dir => '/new/cache/root' }, );
We leave this running for a few hours (or as needed), then replace it with
my $cache = CHI->new( driver => 'File', root_dir => '/new/cache/root' );
You can access the mirror cache with the \*(C`mirror_cache\*(C' method. For example, to see how many keys have made it over to the mirror cache:
my @keys = $cache->mirror_cache->get_keys();
As illustrated above, you create subcaches by passing the \*(C`l1_cache\*(C' and/or \*(C`mirror_cache\*(C' option to the \s-1CHI\s0 constructor. These options, in turn, should contain a hash of options to create the subcache with.
The cache containing the subcache is called the parent cache.
The following options are automatically inherited by the subcache from the parent cache, and may not be overridden:
expires_at expires_in expires_variance serializer
(Reason: for efficiency, we want to create a single cache object and store it in both caches. The cache object contains expiration information and is dependent on the serializer. At some point we could conceivably add code that will use a single object or separate objects as necessary, and thus allow the above to be overridden.)
The following options are automatically inherited by the subcache from the parent cache, but may be overridden:
namespace on_get_error on_set_error
All other options are initialized in the subcache as normal, irrespective of their values in the parent.
It is not currently possible to pass an existing cache in as a subcache.
These behaviors hold regardless of the type of subcache.
The following methods are distributed to both the primary cache and subcache:
clear expire purge remove
The following methods return information solely from the primary cache. However, you are free to call them explicitly on the subcache. (Trying to merge in subcache information automatically would require too much guessing about the caller's intent.)
get_keys get_namespaces get_object get_expires_at exists_and_is_expired is_valid dump_as_hash
It is valid for a cache to have one of each kind of subcache, e.g. an L1 cache and a mirror cache.
A cache cannot have more than one of each kind of subcache, but a subcache can have its own subcaches, and so on. e.g.
my $cache = CHI->new( driver => 'Memcached', servers => [ "10.0.0.15:11211", "10.0.0.15:11212" ], l1_cache => { driver => 'File', root_dir => '/path/to/root', l1_cache => { driver => 'RawMemory', global => 1 } } );
Returns a boolean indicating whether this cache has subcaches.
Returns the L1 cache for this cache, if any. Can only be called if has_subcaches is true.
Returns the mirror cache for this cache, if any. Can only be called if has_subcaches is true.
Returns the subcaches for this cache, in arbitrary order. Can only be called if has_subcaches is true.
Returns a boolean indicating whether this is a subcache.
Returns the type of subcache as a string, e.g. 'l1_cache' or 'mirror_cache'. Can only be called if is_subcache is true.
Returns the parent cache (weakened to prevent circular reference). Can only be called if is_subcache is true.
At this time, subcache behavior is hardcoded into CHI::Driver, so there is no easy way to modify the behavior of existing subcache types or create new ones. We'd like to make this more flexible eventually.
If \*(L"is_size_aware\*(R" or \*(L"max_size\*(R" are passed to the constructor, the cache will be size aware - that is, it will keep track of its own size (in bytes) as items are added and removed. You can get a cache's size with \*(L"get_size\*(R".
Size aware caches generally keep track of their size in a separate meta-key, and have to do an extra store whenever the size changes (e.g. on each set and remove).
If a cache's size rises above its \*(L"max_size\*(R", items are discarded until the cache size is sufficiently below the max size. (See \*(L"max_size_reduction_factor\*(R" for how to fine-tune this.)
The order in which items are discarded is controlled with \*(L"discard_policy\*(R". The default discard policy is 'arbitrary', which discards items in an arbitrary order. The available policies and default policy can differ with each driver, e.g. the CHI::Driver::Memory driver provides and defaults to an '\s-1LRU\s0' policy.
Size awareness was chiefly designed for, and works well with, the CHI::Driver::Memory driver: one often needs to enforce a maximum size on a memory cache, and the overhead of tracking size in memory is negligible. However, the capability may be useful with other drivers.
Some drivers - for example, CHI::Driver::FastMmap and CHI::Driver::Memcached - inherently keep track of their size and enforce a maximum size, and it makes no sense to turn on \s-1CHI\s0's size awareness for these.
Also, for drivers that cannot atomically read and update a value - for example, CHI::Driver::File - there is a race condition in the updating of size that can cause the size to grow inaccurate over time.
You can subclass \s-1CHI\s0 for your own application and configure it in a variety of ways, e.g. predefining storage types and defaults for new cache objects. Your configuration will be independent of the main \s-1CHI\s0 class and other \s-1CHI\s0 subclasses.
Start with a trivial subclass:
package My::CHI; use base qw(CHI); 1;
Then, just use your subclass in place of \s-1CHI:\s0
my $cache = My::CHI->new( ... );
print $cache->chi_root_class; ==> 'My::CHI'
This obviously doesn't change any behavior by itself. Here's an example with actual config:
package My::CHI; use base qw(CHI);
_\|_PACKAGE_\|_->config({ storage => { local_file => { driver => 'File', root_dir => '/my/root' }, memcached => { driver => 'Memcached::libmemcached', servers => [ '10.0.0.15:11211', '10.0.0.15:11212' ] }, }, namespace => { 'Foo' => { storage => 'local_file' }, 'Bar' => { storage => 'local_file', depth => 3 }, 'Baz' => { storage => 'memcached' }, } defaults => { storage => 'local_file' }, memoize_cache_objects => 1, });
1;
Each of these config keys is explained in the next section.
A map of names to parameter hashrefs. This provides a way to encapsulate common sets of parameters that might be used in many caches. e.g. if you define storage => { local_file => { driver => 'File', root_dir => '/my/root' }, ... } then my $cache = My::CHI->new (namespace => 'Foo', storage => 'local_file'); is equivalent to my $cache = My::CHI->new (namespace => 'Foo', driver => 'File', root_dir => '/my/root');
A map of namespace names to parameter hashrefs. When you create a cache object with the specified namespace, the hashref of parameters will be applied as defaults. e.g. if you define namespace => { 'Foo' => { driver => 'File', root_dir => '/my/root' }, 'Bar' => { storage => 'database' }, ... } then my $cache1 = My::CHI->new (namespace => 'Foo'); my $cache2 = My::CHI->new (namespace => 'Bar'); is equivalent to my $cache1 = My::CHI->new (namespace => 'Foo', driver => 'File', root_dir => '/my/root'); my $cache2 = My::CHI->new (namespace => 'Bar', storage => 'database');
A hash of parameters that will be used as core defaults for all cache objects created under this root class. e.g. defaults => { on_get_error => 'die', expires_variance => 0.2, } These can be overridden by namespace defaults, storage settings, or \*(C`new\*(C' parameters.
True or false, indicates whether \*(C`My::CHI->new\*(C' should memoize and return the same cache object if given the same parameters. This can speed things up if you create cache objects frequently. Will currently only work for 0- or 1- key parameter hashes. e.g. My::CHI->config({ memoize_cache_objects => 1, }); then # $cache1 and $cache2 will be the same object, regardless of what # namespace and storage defaults are associated with 'Foo' # my $cache1 = My::CHI->new(namespace => 'Foo'); my $cache2 = My::CHI->new(namespace => 'Foo');
# $cache3 and $cache4 will be different objects # my $cache3 = My::CHI->new (namespace => 'Bar', driver => 'File', root_dir => '/my/root'); my $cache4 = My::CHI->new (namespace => 'Bar', driver => 'File', root_dir => '/my/root'); To clear the memoized cache objects, call My::CHI->clear_memoized_cache_objects;
Defaults are applied in the following order, from highest to lowest precedence:
Parameters passed in \*(C`new\*(C'
Namespace defaults, if any
Storage settings, if any
Core defaults defined under 'defaults'
A subclass will automatically inherit the configuration of its parent if it does not call \*(C`config\*(C' itself (ala Class::Data::Inheritable).
use YAML::XS qw(LoadFile);
_\|_PACKAGE_\|_->config(LoadFile("/path/to/cache.yml"));
The following drivers are currently available as part of this distribution:
CHI::Driver::Memory - In-process memory based cache
CHI::Driver::RawMemory - In-process memory based cache that stores references directly instead of serializing/deep-copying
CHI::Driver::File - File-based cache using one file per entry in a multi-level directory structure
CHI::Driver::FastMmap - Shared memory interprocess cache via mmap'ed files
CHI::Driver::Null - Dummy cache in which nothing is stored
CHI::Driver::CacheCache - \s-1CHI\s0 wrapper for Cache::Cache
The following drivers are currently available as separate \s-1CPAN\s0 distributions:
CHI::Driver::Memcached - Distributed memory-based cache (works with Cache::Memcached, Cache::Memcached::Fast, and Cache::Memcached::libmemcached)
CHI::Driver::DBI - Cache in any DBI-supported database
CHI::Driver::BerkeleyDB - Cache in BerkeleyDB files
CHI::Driver::Redis - Cache in Redis <http://redis.io/>
CHI::Driver::SharedMem - Cache in shared memory
This list is likely incomplete. A complete set of drivers can be found on \s-1CPAN\s0 by searching for \*(L"CHI::Driver\*(R".
See CHI::Benchmarks for a comparison of read/write times of both \s-1CHI\s0 and non-CHI cache implementations.
\*(C`etc/bench/bench.pl\*(C' in the \*(C`CHI\*(C' distribution contains a script to run these types of benchmarks on your own system.
See CHI::Driver::Development for information on developing new drivers.
\*(C`CHI\*(C' uses Log::Any for logging events. For example, a debug log message is sent for every cache get and set.
See Log::Any documentation for how to control where logs get sent, if anywhere.
\s-1CHI\s0 can record statistics, such as number of hits, misses and sets, on a per-namespace basis and log the results to your Log::Any logger. You can then use utilities included with this distribution to read stats back from the logs and report a summary. See CHI::Stats for details.
\s-1CHI\s0 is intended as an evolution of DeWitt Clinton's Cache::Cache package. It starts with the same basic \s-1API\s0 (which has proven durable over time) but addresses some implementation shortcomings that cannot be fixed in Cache::Cache due to backward compatibility concerns. In particular:
Some of Cache::Cache's subclasses (e.g. Cache::FileCache) have been justifiably criticized as inefficient. \s-1CHI\s0 has been designed from the ground up with performance in mind, both in terms of general overhead and in the built-in driver classes. Method calls are kept to a minimum, data is only serialized when necessary, and metadata such as expiration time is stored in packed binary format alongside the data.
New Cache::Cache subclasses can be tedious to create, due to a lack of code refactoring, the use of non-OO package subroutines, and the separation of \*(L"cache\*(R" and \*(L"backend\*(R" classes. With \s-1CHI\s0, the goal is to make the creation of new drivers as easy as possible, roughly the same as writing a \s-1TIE\s0 interface to your data store. Concerns like serialization and expiration options are handled by the driver base class so that individual drivers don't have to worry about them.
Probably because of the reasons above, Cache::Cache subclasses were never created for some of the most popular caches available on \s-1CPAN\s0, e.g. Cache::FastMmap and Cache::Memcached. \s-1CHI\s0's goal is to be able to support these and other caches with a minimum performance overhead and minimum of glue code required.
The Cache distribution is another redesign and implementation of Cache, created by Chris Leishman in 2003. Like \s-1CHI\s0, it improves performance and reduces the barrier to implementing new cache drivers. It breaks with the Cache::Cache interface in a few ways that I considered non-negotiable - for example, get/set do not serialize data, and namespaces are an optional feature that drivers may decide not to implement.
\s-1CPAN\s0 sports a variety of full-featured standalone cache modules representing particular backends. \s-1CHI\s0 does not reinvent these but simply wraps them with an appropriate driver. For example, CHI::Driver::Memcached and CHI::Driver::FastMmap are thin layers around Cache::Memcached and Cache::FastMmap.
Of course, because these modules already work on their own, there will be some overlap. Cache::FastMmap, for example, already has code to serialize data and handle expiration times. Here's how \s-1CHI\s0 resolves these overlaps.
\s-1CHI\s0 handles its own serialization, passing a flat binary string to the underlying cache backend. The notable exception is CHI::Driver::RawMemory which does no serialization.
\s-1CHI\s0 packs expiration times (as well as other metadata) inside the binary string passed to the underlying cache backend. The backend is unaware of these values; from its point of view the item has no expiration time. Among other things, this means that you can use \s-1CHI\s0 to examine expired items (e.g. with $cache->get_object) even if this is not supported natively by the backend. At some point \s-1CHI\s0 will provide the option of explicitly notifying the backend of the expiration time as well. This might allow the backend to do better storage management, etc., but would prevent \s-1CHI\s0 from examining expired items.
Naturally, using \s-1CHI\s0's FastMmap or Memcached driver will never be as time or storage efficient as simply using Cache::FastMmap or Cache::Memcached. In terms of performance, we've attempted to make the overhead as small as possible, on the order of 5% per get or set (benchmarks coming soon). In terms of storage size, \s-1CHI\s0 adds about 16 bytes of metadata overhead to each item. How much this matters obviously depends on the typical size of items in your cache.
Questions and feedback are welcome, and should be directed to the perl-cache mailing list:
http://groups.google.com/group/perl-cache-discuss
Bugs and feature requests will be tracked at \s-1RT:\s0
http://rt.cpan.org/NoAuth/Bugs.html?Dist=CHI [email protected]
The latest source code can be browsed and fetched at:
http://github.com/jonswar/perl-chi/tree/master git clone git://github.com/jonswar/perl-chi.git
Thanks to Dewitt Clinton for the original Cache::Cache, to Rob Mueller for the Perl cache benchmarks, and to Perrin Harkins for the discussions that got this going.
\s-1CHI\s0 was originally designed and developed for the Digital Media group of the Hearst Corporation, a diversified media company based in New York City. Many thanks to Hearst management for agreeing to this open source release.
Cache::Cache
Jonathan Swartz <[email protected]>
This software is copyright (c) 2012 by Jonathan Swartz.
This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.