Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

This document describes the message format for TermWeb Business Integration Message Adapter (BIMA). BIMA is used for integration of TermWeb with other enterprise systems via a message broker middleware, such as WebSphere MQ, ActiveMQ, RabbitMQ, Apache Kafka.
BIMA uses the Java Message Service (JMS) standard and Kafka Streams.

Introduction

There are two TermWeb operations that causes BIMA to generate messages: initial transfer (export of term database) and update of term data. When such an operation occurs, BIMA generates one or more messages and publishes them to the message broker's publish/subscribe topics defined in BIMA's configuration file. Each message consists of several properties and a message body, which is an XML structure. The properties are useful for applications when filtering messages to avoid unnecessary XML parsing from the body.

Installation

  • Copy the sample configuration file BimaAdapter.properties to [TERMWEB_HOME] and edit it as necessary

  • Edit [TERMWEB_HOME]/termweb.properties and add the following lines:

    Code Block
    adapter.class.1=com.termweb4.bima.BimaAdapter
    adapter.config.1=[TERMWEB_HOME]/BimaAdapter.properties
  • If you’re going to use JMS you may replace activemq libraries with another: RabbitMQ, IBM MQ. If you’re going use another implementation copy any message broker specific jar files to [TOMCAT_HOME]/webapps/ROOT/WEB-INF/lib that BIMA needs to create instances of TopicConnectionFactory and Topic. By default, TermWeb 4 comes with ActiveMQ implementation files, don’t forget to delete to avoid any conflicts:

    • activemq-client-*.jar

      • Code Block
        languagecpp
        // And its transitive dependencies
        org.apache.geronimo.specs:geronimo-jms_1.1_spec:1.1.1
        org.apache.geronimo.specs:geronimo-j2ee-management_1.1_spec:1.0.1
        org.fusesource.hawtbuf:hawtbuf:1.11

Configuration

Property

Description

adapter.enabled

Enables or disables the message publishing. Set the value to false to disable the messaging.

adapter.jms.enabled

Enables or disables the message publishing only for JMS publisher. Set the value to false to disable the messaging.

adapter.kafka.enabled

Enables or disables the message publishing only for Kafka publisher. Set the value to false to disable the messaging.

adapter.maxMessageSize

Defines the max message size in bytes for publishing to the initial topic. A value of 0 indicates no max size for the messages.

topic.update.section.namePattern

Defines the name pattern for the section update topic. See below for description of the pattern.

topic.update.concept.namePattern

Defines the name pattern for the concept update topic. See below for description of the pattern.

topic.update.term.namePattern

Defines the name pattern for the term update topic. See below for description of the pattern.

jms.tcf.type

Defines the class name for the TopicConnectionFactory

jms.tcf.*

Defines several properties for the TopicConnectionFactory class. The names of the properties depend on the properties available in the class defined by jms.tcf.type. BIMA attempts to call a setter method for each defined property, e.g., the property jms.tcf.queueManager will call setQueueManager() with the value from the property.

jms.topic.init.type

Defines the class name for the Topic used for initial export messages.

jms.topic.init.nameMethod

The name of the method in the Topic (sub)class used for setting the baseTopicName. Normally this is setBaseTopicName.

jms.topic.init.*

Other properties for the class defined by jms.topic.init.type. BIMA attempts to call the corresponding setter method for each property.

jms.topic.update.section.type

Defines the class name for the Topic used for section update messages

jms.topic.update.section.nameMethod

The name of the method in the Topic (sub)class used for setting the baseTopicName. Normally this is setBaseTopicName.

jms.topic.update.section.*

Other properties for the class defined by jms.topic.update.section.type. BIMA attempts to call the corresponding setter method for each property.

jms.topic.update.concept.type

Defines the class name for the Topic used for concept update messages

jms.topic.update.concept.nameMethod

The name of the method in the Topic (sub)class used for setting the baseTopicName. Normally this is setBaseTopicName.

jms.topic.update.concept.*

Other properties for the class defined by jms.topic.update.concept.type. BIMA attempts to call the corresponding setter method for each property.

jms.topic.update.term.type

Defines the class name for the Topic used for term update messages

jms.topic.update.term.nameMethod

The name of the method in the Topic (sub)class used for setting the baseTopicName. Normally this is setBaseTopicName.

jms.topic.update.term.*

Other properties for the class defined by jms.topic.update.term.type. BIMA attempts to call the corresponding setter method for each property.

kafka.bootstrap.servers

A list of host/port pairs to use for establishing the initial connection to the Kafka cluster. The client will make use of all servers irrespective of which servers are specified here for bootstrapping – this list only impacts the initial hosts used to discover the full set of servers. This list should be in the form host1:port1,host2:port2,...

kafka.client.id

An id string to pass to the server when making requests. The purpose of this is to be able to track the source of requests beyond just ip/port by allowing a logical application name to be included in server-side request logging. Default value is TermWebProducer.

Name pattern

The name pattern for the different topics is a text string with variables that will be replaced with values depending on the dictionary, section and language that are affected by the update message. The variables are:

  • <DICTIONARYID>

Is replaced by the id of the dictionary.

  • <SECTIONID>

Is replaced by the id of the section.

  • <LANGUAGE>

Is replaced by the language iso code. Only applicable for term update messages.

  • <TAG>

Is replaced by the tag the user can enter when selecting an initial transmission. Thus, only applicable for initial transmission messages.

Adapter administration

When the adapter is installed in TermWeb an BIMA administration panel is available in TermWeb. Log in as administrator and go to the AdminView → System → BIMA. The administration panel opens.

General panel

The general panel allows the adapter to be enabled and disabled. It also allows temporary disable one of the parts JMS or Kafka. Please note that the setting in BimaAdapter.properties is not affected by this control. When restarting the adapter will be enabled or disabled according to the value in the properties file.

Information about current configuration is also displayed here.

Export panel

The export panel allows the administrator to do an initial transfer of all or a part of the term data in a dictionary. The selection of concept, terms and fields are based on regular Export settings used for normal exports.

Select settings and topic

Start by select one of the available export settings. If no setting is appropriate for your needs, you can click the “Create export setting” link which opens the Export settings panel and lets you define a new setting. Save the setting and then select the BIMA administration icon again in the AdminView → System and select the newly created setting.

Continue by entering a name of the topic to which the data should be published. Choose type of message broker configured in adapter (JMS or Kafka)

Click on the “Next >” button.

Confirm your selection

Your selected export setting is now displayed in detail with languages, fields and total number of concepts that will be published when using this setting. By moving your mouse pointer over the language and field summary you will see a list of all languages and fields included.

If this is the data you expected, click the “Start export” button to start publishing the messages. Otherwise, click the “< Back” button to go back and select another export setting and/or topic name.

Export progress

The selected concepts are now retrieved from the database, converted to XML segments, and published to the selected topic. For details about the message format, see section Message structure below.

A progress bar indicates how much of the export is completed. The elapsed time and an estimation of the total time is also given.

You can stop the export by pressing the Stop button. Please note that this may cause receiving systems to only get parts of the data intended.

Export summary

When the export is done, or if the export was stopped by the user, a summary is displayed with the total number of concepts published, together with elapsed time.

Message structure

Text

Message properties

The following specific properties are defined in a JMS message:

Property

Type

Description

Possible values

DictionaryId

String

Contains the id of the dictionary that contains the object described in this message

Any string

SectionId

String

Contains the id of the section that contains the object described in this message.

Any string

NewSectionId

String

Contains the new section id when the section was changed, otherwise null.

Any string

OldSectionId

String

Contains the old section id when the section was changed, otherwise null.

Any string

DataType

String

Identifies the type of data entity that is contained in this message.

Is one of:

  • SECTION (only section data)

  • CONCEPT (a full concept including terms)

  • TERM (a single term)

DataAction

String

Identifies the type of action that was performed on the data.

Is one of:

  • CREATED (the data entity was created)

  • UPDATED (some field in the entity was modified)

  • DELETED (the entity was deleted)

  • UNMODIFIED (no field in the entity was modified. Only happens for terms in a concept that has generated an message with DataAction UPDATED.)

TransmissionType

String

Identifies the type of transmission.

Is one of:

  • INITIAL (transfer of (a subset of) all concepts invoked by the user)

  • UPDATE (transfer of one or more concepts due to an update of the TermWeb database)

Restore

Boolean

Identifies if the message was generated in a restore operation in TermWeb.

True if it is a restore operation, otherwise false

Initial transmission

During an initial transfer (described in Adapter administration above) BIMA will create a message (or several messages if the message is larger than the maximum message size specified in the config file) for each section in the dictionary being exported. The TransmissionType in the message is set to INITIAL. DictionaryId and SectionId are set to their respective values. All other properties are set to null.

The body of the message is an XML structure containing all the concepts being exported from the section, which could be none if the export setting's filter does not include any concept from this section.

Update transmissions

An update transmission occurs when some term data is updated in TermWeb. This is done by either renaming or deleting a section, creation, modification, or deletion of concepts, or by importing a file. BIMA then creates one or more messages and publishes them to the broker.

Section updates

Messages caused by section updates are published to the section update topic(s) defined by topic.update.section in the config file. All messages have DataType set to SECTION.

If a section is renamed a message is published to the topic with DataAction set to UPDATED. The XML in the body consists only of the dictionary and section. The new name is found in the “name” attribute of the section tag.

If a section is deleted a message with DataAction set to DELETED is published. Only dictionary and section in the XML.

Term data updates

For every change there will be a message for the entire concept together with a message for each affected term. There is therefore a redundancy in the messages since the data in the term messages also are included in the concept message.

Concept messages are published to the concept topic(s) defined by topic.update.concept in the config file. Term update messages are published to the topic(s) defined by topic.update.term. The subscribing client applications can thus select the most appropriate data structure.

In the group of messages, the concept message has DataType set to CONCEPT, an appropriate value set for DataAction and the body is the XML for the entire concept. Each term message has DataType set to TERM, a DataAction value as appropriate and the XML structure for the term.

When importing a file there will be a series of messages, just like if every concept was created or modified manually. The messages will be sent to the broker when the import of the file is completed.

Description of term data update transmissions

The messages generated from the different kinds of updates are as follows.

Creation of a new concept

One message for the entire concept, and one message for each term in the concept. All have DataAction set to CREATED.

Deletion of a concept

One message for the entire concept, and one message for each term in the concept. All have DataAction set to DELETED.

Modification of a concept

One message for the entire concept with DataAction set to UPDATED.

For every term that has been created, there is a message with the term data and DataAction set to CREATED.

For every term that has been deleted, there is a message with the term data and DataAction set to DELETED.

If any term field except the term name or language has been modified, there is a message with the term data and DataAction set to UPDATED.

If the term name or language has been changed, the old term is considered to have been deleted and a new one created. Therefore, this will create two messages: one DELETED and one CREATED.

If any field on the concept level has been modified, a message with DataAction UNMODIFIED is generated for every term that has not already generated a message with DataAction UPDATED, CREATED or DELETED. Subscribing systems can then determine if some action needs to be taken with the data.

If the section has been changed, the concept is considered to have been deleted and new one created. The messages for the deletion are published to the old section's topic and the creation messages are published to the new section's topic.

XML structure

The XML structure for the messages is like the structure of a TermWeb XML export file and is described in TermWeb-BIMA.dtd.
Message example:

Code Block
languagexml
<dictionary id=”8” name=”Automotive terms”>
  <section id=”16” name="Headings">
    <concept id="4116633" termno=”430”>
      <metadata>
        <createdOn>1999-05-12 03:12:54</createdOn>
        <createdBy>admin</createdBy>
        <changedOn>2004-10-04 13:53:27</changedOn>
        <changedBy>admin</changedBy>
      </metadata>
      <field name="Domain">All domains</field>
      <field name="Origin">Termlex</field>
      <field name="Term No.">430</field>
      ...
      <term>
	<metadata>
	  <createdOn />
	  <createdBy />
	  <changedOn>2004-10-04 13:53:22</changedOn>
	  <changedBy>admin</changedBy>
	</metadata>
        <field name="Term">Lubrication, injection pump</field>
        <field name="Language">en-GB</field>
        <field name="Region"/>
        ...
      </term>
      <term>
      ...
      </term>
      ...
    </concept> 
    ...
  </section>
  ....
</dictionary>

As previously stated, there are one message containing the entire concept and one for each term affected for every term data update. The XML structure in a concept message contains exactly one <concept> tag with corresponding content. For a term message, the body contains exactly one <term> tag with content. Since 2005-01-31 the Status field on the concept level is also included in every term message before the <term> tag.

Kafka integration

Integration for Kafka is using the same message structure. By default, all messages serialized as JSON strings. XML part of message will be stored in a field called body.

Keys for messages will be like topic name with additional suffix like part number or entry id.

For initial transmission key will be <topic-name>_p1-p100, where p1 and p100 sequential position number of first and last element in message.
Numbering starts with 1.

For update transmission key will be <topic-name>_C1234 or <topic-name>_T1234 depending on type of message. C1234 is and id for Concept and T1234 id for Term. In case of Section change BIMA will create two messages one with ending suffix _deleted and another one with _created.

Table of Contents