Periscope Result Ranker module
The Periscope Result Ranker module uses neural networks to store Find Bar search results and rank them by relevance. It extends the Periscope modules.
The module learns user preferences to offer better result ranking for subsequent searches. By default, search result rankings are stored per user. You can change the configuration as necessary. For example, you can enable an individual ranking for each user or for selected users only. Additionally, you can configure the memory size of networks to mitigate possible memory consumption issues in large setups.
info.magnolia.periscope.ResultRankerConfiguration allows you to
entirely disable the ranking of Find Bar search results. To do so, set
true in the
resultRankerConfiguration: disabled: false
You may want to disable this configuration to reduce memory usage or resolve potential compatibility issues with DL4J libraries.
Maven is the easiest way to install the module. Add the following to your bundle:
<dependency> <groupId>info.magnolia.periscope</groupId> <artifactId>magnolia-periscope-result-ranker</artifactId> <version>1.2.4</version> </dependency>
The module comes with the following default configuration:
outputUnits: 10000 rankingNetworkStorageStrategy: class: info.magnolia.periscope.rank.ml.jcr.JcrUsernameNetworkStorageStrategy
required, default is
The memory size of neural networks.
The result-ranking system requires memory (heap space) and disk space per unit for each user (local ranking) or instance (global ranking). You can adjust the size of the memory used per unit to mitigate possible memory consumption issues (see Result Ranker memory size).
The result-ranking memory strategy.
The default strategy stores result rankings per user. Other strategies are possible (see Result Ranker strategy).
To adjust the strategy, set the
required, default is
Other possible values must be a subtype of
The Periscope Result Ranker module creates a certain number of memory units. The total number of memory units depends on the Result Ranker strategy. The size of a single memory unit is based on the Result Ranker memory size.
You can set the Result Ranker strategy via the
class property of the
This default configuration enables result ranking for each user operating on the author instance. With this strategy, the module creates one neural network for each user. The memory footprint grows with every new user working on an author instance.
With this strategy, only users with the role
have local (per-user) ranking memory. Any other users work with the
global (per-instance) ranking memory.
To keep the memory footprint for the Result Ranker on a minimum level,
You can develop your own custom result-ranking strategy. To do this,
create a custom class that implements
and set the
class property in the configuration accordingly.
You can globally configure the memory size of neural networks. To do so,
outputUnits value. The default memory size is
value represents the maximum number of Find Bar search results that can
be stored and ranked in networks.
Large networks can store more results but use up more memory, while small networks consume less memory but might lose stored results to free up additional memory. Results are removed from networks based on a least-recently-used policy. This ensures that frequent results remain in memory irrespective of when they were added to the networks. You may want to configure the memory size of networks depending on the Result Ranker strategy.
When you change the
outputUnits value (e.g. from
you need to clean up the JCR
rankings workspace. Otherwise, an error
will appear when you select a search result. This is because the
networks loaded into memory were created using a configuration that no
longer exists, rendering any stored results obsolete. When you change
outputUnits value, make sure that you delete any stored networks
and log into the Magnolia instance again to regenerate networks using
the new configuration (see Clearing
Result Ranker memory).
The Periscope Result Ranker module configuration resides in
The module is deployed as a JAR file, but you can change the
configuration by one of the following means:
The configuration is read by the Resources module. Magnolia scans the following for a resource (in this particular order):
Classpath (JAR file)
The Java bean representation of the selected resource is then stored in the registry. The object in the registry may get decorated if a decorator exists for that configuration.
Storing configuration in the JCR
The configuration data is read on startup and after it has been changed. The actual data is stored in the module’s configuration registry. You can look it up using the Definitions app in modules > periscope-result-ranker.
If you change the
Open the Resource Files app.
Browse to and select periscope-result-ranker > config.yaml.
In the action bar, click Edit file. The Resource Files app creates a copy of the currently used configuration and stores it in the JCR
Edit the file as necessary.
Click Save changes.
With decoration, you can adapt the currently used configuration (whether it is from a light module, hotfix or JAR file). To learn more about decoration, see Definition decoration concept.
A decorator file can reside in any Magnolia Maven module or any light
decorator file location). In the example below, we will create a
decorator file in a light module named
Within the light module, create the file
outputUnits: 1000 rankingNetworkStorageStrategy: class: info.magnolia.periscope.rank.ml.jcr.JcrUserRoleNetworkStorageStrategy
The Periscope Result Ranker module stores all user-based, role-based,
and custom rankings in the JCR
rankings workspace. The module creates
one node for each memory unit.
Nodes for local (per-user) rankings are named after user names. Nodes
for global (per-instance) rankings are named
To clear the Result Ranker memory:
Open the JCR app.
Switch to the
Select the nodes you want to delete.
In the action bar, click Delete item.
The Periscope Result Ranker module supports the 64-bit versions of Linux, Mac OS and Windows by default. If you use Magnolia in one of the environments below, you must add the corresponding dependencies manually.
<dependency> <groupId>org.bytedeco.javacpp-presets</groupId> <artifactId>openblas</artifactId> <classifier>linux-armhf</classifier> </dependency>
<dependency> <groupId>org.bytedeco.javacpp-presets</groupId> <artifactId>openblas</artifactId> <classifier>linux-ppc64le</classifier> </dependency> <dependency> <groupId>org.nd4j</groupId> <artifactId>nd4j-native</artifactId> <classifier>linux-ppc64le</classifier> </dependency>