Google Cloud Search Connector SDK

Google Cloud Search connectors are software programs allowing Google Cloud Search to operate on third-party data. The Content Connector SDK is a wrapper around the REST API allowing one to quickly create connectors.

There are two types of connectors: content connectors and identity connectors. Content connectors are used to traverse a repository and index the data so that Google Cloud Search can effectively search that data. For more information on building content connectors using Google Cloud Search connector SDK refer to

Identity connectors are used to map your enterprise's identities and group rosters to the Google accounts and groups used by Google Cloud Search. For more information on building identity connectors using Google Cloud Search connector SDK, refer to

Build instructions

Building the Google Cloud Search Connector SDK requires the following development tools:

  1. Clone the SDK repository from GitHub:

    git clone
    cd connector-sdk
  2. Checkout the desired version of the SDK:

    git checkout tags/v1-0.0.5
  3. Install the SDK components:

    mvn install

Maven commands:

Running integration tests

The integration tests check the correct interoperability between the connector SDK and the Cloud Search APIs. They are run using the Failsafe plug-in and follow its naming conventions (i.e., they are suffixed with "IT").

To run them, build and install the SDK, then cd into the it/ subdirectory, and issue the command

mvn verify \
    -DskipITs=false \
    -Dapi.test.serviceAccountPrivateKeyFile=<PATH_TO_SERVICE_ACCOUNT_KEY> \
    -Dapi.test.identitySourceId=<IDENTITY_SOURCE_ID> \
    -Dapi.test.customerId=<CUSTOMER_ID> \


It's necessary to set a specific schema for the tests to run, otherwise a check during the set-up of the tests will fail. The schema can be found in it/schema.json (or it/schema.protoascii).

The following is an example of a complete command:

mvn verify \
    -DskipITs=false \
    -Dapi.test.serviceAccountPrivateKeyFile=key.json \
    -Dapi.test.sourceId=01cb7dfca117fbb591360a2b6e46912e \
    -Dapi.test.identitySourceId=01cb7dfca117fbb591360a2a5b64177c \
    -Dapi.test.customerId=C14e3bmn2 \

Adding -Dit.test=<IT class>#<method> to the command line allows to run only a specific test.

Notice also that will it's possible to run only unit tests (using -DskipITs=true) it's not obvious how to run only integration tests (-DskipTests=true skips both unit and integration tests). A work-around for this is specifying -Dtest=NONE -DfailIfNoTests=false (assuming that there's no test class called NONE) plus the appropriate flags to run integration tests. For example:

mvn verify \
    -Dtest=NONE -DfailIfNoTests=false \
    -DskipITs=false \
    -Dit.test=SomeTest \

Code coverage

A code coverage report for unit tests is generated for each module as part of the "test" goal. The output directory for the report is <module>/target/site/jacoco/.