James' extensions for Rspamd

This module is for developing and delivering extensions to James for the Rspamd (the spam filtering system) and ClamAV (the antivirus engine).

How to run

The Rspamd extension requires an extra configuration file rspamd.properties to configure Rspamd connection Configuration parameters:
- rSpamdUrl : URL defining the Rspamd's server. Eg: http://rspamd:11334
- rSpamdPassword : Password for pass authentication when request to Rspamd's server. Eg: admin
Declare the extensions.properties for this module.

guice.extension.module=org.apache.james.rspamd.module.RspamdModule
guice.extension.task=org.apache.james.rspamd.module.RspamdTaskExtensionModule

Declare the Rspamd mailbox listeners in listeners.xml. Eg:

<listener>
    <class>org.apache.james.rspamd.RspamdListener</class>
</listener>

Declare the Rspamd mailet for custom mail processing.
You can specify the virusProcessor if you want to enable virus scanning for mail. Upon configurable virusProcessor you can specify how James process mail virus. We provide a sample Rspamd mailet and virusProcessor configuration:
You can specify the rejectSpamProcessor. Emails marked as rejected by Rspamd will be redirected to this processor. This corresponds to emails with the highest spam score, thus delivering them to users as marked as spam might not even be desirable.
The rewriteSubject option allows to rewritte subjects when asked by Rspamd.

<processor state="local-delivery" enableJmx="true">
    <mailet match="All" class="org.apache.james.rspamd.RspamdScanner">
        <rewriteSubject>true</rewriteSubject>
        <virusProcessor>virus</virusProcessor>
        <rejectSpamProcessor>spam</rejectSpamProcessor>
        <onMailetException>ignore</onMailetException>
    </mailet>
    <mailet match="IsMarkedAsSpam=org.apache.james.rspamd.status" class="WithStorageDirective">
        <targetFolderName>Spam</targetFolderName>
    </mailet>
    <mailet match="All" class="LocalDelivery"/>
</processor>

<!--Choose one between these two following virus processor, or configure a custom one if you want-->
<!--Hard reject virus mail-->
<processor state="virus" enableJmx="false">
    <mailet match="All" class="ToRepository">
        <repositoryPath>file://var/mail/virus/</repositoryPath>
    </mailet>
</processor>

<!--Soft reject virus mail-->
<processor state="virus" enableJmx="false">
    <mailet match="All" class="StripAttachment">
        <remove>all</remove>
        <pattern>.*</pattern>
    </mailet>
    <mailet match="All" class="AddSubjectPrefix">
        <subjectPrefix>[VIRUS]</subjectPrefix>
    </mailet>
</processor>

<!--Store rejected spam emails (with a very high score) -->
<processor state="spam" enableJmx="false">
    <mailet match="All" class="ToRepository">
        <repositoryPath>cassandra://var/mail/spam</repositoryPath>
    </mailet>
</processor>

Declare the webadmin for Rspamd in webadmin.properties

extensions.routes=org.apache.james.rspamd.route.FeedMessageRoute

How to use admin endpoint, see more at Additional webadmin endpoints

Docker compose file example: docker-compose.yml or docker-compose-distributed.yml.
Please configure ClamAV integration into Rspamd if you want to enable virus scanning.
The sample-configuration: sample-configuration
For running docker-compose, first compile this project

mvn clean install -DskipTests

then run it: docker-compose up

Additional webadmin endpoints

Report spam messages to Rspamd

One can use this route to schedule a task that reports spam messages to Rspamd for its spam classify learning.

curl -XPOST 'http://ip:port/rspamd?action=reportSpam

This endpoint has the following param:

action (required): need to be reportSpam
messagesPerSecond (optional): Concurrent learns performed for Rspamd, default to 10
period (optional): duration (support many time units, default in seconds), only messages between now and now - duration are reported. By default, all messages are reported. These inputs represent the same duration: 1d, 1day, 86400 seconds, 86400...
samplingProbability (optional): float between 0 and 1, represent the chance to report each given message to Rspamd. By default, all messages are reported.
classifiedAsSpam (optional): Boolean, true to only include messages tagged as Spam by Rspamd, false for only messages tagged as ham by Rspamd. If omitted all messages are included.
rspamdTimeout (optional): duration, Default is 15 seconds. Provide configuration timeout when HTTP request to rspamd for learning. Will return the task id. E.g:

{
    "taskId": "70c12761-ab86-4321-bb6f-fde99e2f74b0"
}

Response codes:

201: Task generation succeeded. Corresponding task id is returned.
400: Invalid arguments supplied in the user request.

More details about endpoints returning a task.

The scheduled task will have the following type FeedSpamToRspamdTask and the following additionalInformation:

{
  "errorCount": 1,
  "reportedSpamMessageCount": 2,
  "runningOptions": {
    "messagesPerSecond": 10,
    "rspamdTimeoutInSeconds": 15,
    "periodInSecond": 3600,
    "samplingProbability": 1.0
  },
  "spamMessageCount": 4,
  "timestamp": "2007-12-03T10:15:30Z",
  "type": "FeedSpamToRspamdTask"
}

Report ham messages to Rspamd

One can use this route to schedule a task that reports ham messages to Rspamd for its spam classify learning.

curl -XPOST 'http://ip:port/rspamd?action=reportHam

This endpoint has the following param:

action (required): need to be reportHam
messagesPerSecond (optional): Concurrent learns performed for Rspamd, default to 10
period (optional): duration (support many time units, default in seconds), only messages between now and now - duration are reported. By default, all messages are reported. These inputs represent the same duration: 1d, 1day, 86400 seconds, 86400...
samplingProbability (optional): float between 0 and 1, represent the chance to report each given message to Rspamd. By default, all messages are reported.
classifiedAsSpam (optional): Boolean, true to only include messages tagged as Spam by Rspamd, false for only messages tagged as ham by Rspamd. If omitted all messages are included.
rspamdTimeout (optional): duration, Default is 15 seconds. Provide configuration timeout when HTTP request to rspamd for learning. Will return the task id. E.g:

{
    "taskId": "70c12761-ab86-4321-bb6f-fde99e2f74b0"
}

Response codes:

201: Task generation succeeded. Corresponding task id is returned.
400: Invalid arguments supplied in the user request.

More details about endpoints returning a task.

The scheduled task will have the following type FeedHamToRspamdTask and the following additionalInformation:

{
  "errorCount": 1,
  "reportedHamMessageCount": 2,
  "runningOptions": {
    "messagesPerSecond": 10,
    "rspamdTimeoutInSeconds": 15,
    "periodInSecond": 3600,
    "samplingProbability": 1.0
  },
  "hamMessageCount": 4,
  "timestamp": "2007-12-03T10:15:30Z",
  "type": "FeedHamToRspamdTask"
}