Knime node for KripoDB package
Clone or download
Latest commit 4606e48 Apr 4, 2018

KripoDB KNIME nodes

KRIPO stands for Key Representation of Interaction in POckets.

KNIME nodes for KripoDB (

Build Status Build status SonarCloud Gate SonarCloud Coverage DOI




  • KripoDB Python package & data files,, required when nodes which use local Kripo files are used. Nodes which talk to web service work without the KripoDB Python package.

Steps to get KripoDB nodes inside KNIME:

  1. Goto Help > Install new software ... menu
  2. Press add button
  3. Fill text fields with
  4. Select --all sites-- in work with pulldown
  5. Open KNIME 3D-e-Chem Contributions folder
  6. Select KripoDB
  7. Install software & restart


See a minimal example workflow at examples/ The workflow can be run by importing it into KNIME as an archive.

Other workflows using the KripoDB nodes can be found at

Make sure the Python used by KNIME is the same the Python with kripodb package installed.


Development requirements:

Steps to get development environment setup:

  1. Download KNIME SDK from

  2. Install/Extract/start KNIME SDK

  3. Start SDK

  4. Install m2e (Maven integration for Eclipse) + KNIME Python Integration + RDKit KNIME integration + KNIME Testing framework + 3D-e-Chem node category

    1. Goto Help > Install new software ...
    2. Make sure Update site is , , and are in the pull down list otherwise add it
    3. Select --all sites-- in work with pulldown
    4. Select m2e (Maven integration for Eclipse)
    5. Select RDKit KNIME integration
    6. Select Abstract Python wrapper KNIME node and helpers
    7. Select Test Knime workflows from a Junit test
    8. Select Splash & node category for 3D-e-Chem KNIME nodes
    9. Install software & restart
  5. Import this repo as an Existing Maven project

During import the Tycho Eclipse providers must be installed.


mvn package

Jar has been made in plugin/target directory. An Eclipse update site will be made in p2/target/repository repository.


mvn verify

Tests in tests module will have been run with results in test/target/surefire-reports directory. There are unit tests and workflow tests both are executed in the KNIME eclipse application. See for more information about workflow tests.

Create web service client

The web service client is generated using Swagger Code Generator and stored inside plugin/src/java/nl/esciencecenter/e3dchem/kripodb/ws/client/ directory.

  1. Start KripoDB webservice
kripodb serve data/similarities.frozen.h5 data/fragments.sqlite data/pharmacophores.h5
  1. Download swagger code generator
  1. Generate a client for KripoDB web service
java -jar swagger-codegen-cli-2.2.3.jar generate \
--input-spec http://localhost:8084/kripo/swagger.json \
--output client \
--lang java \
--config swagger-codegen.config.json
  1. Compile client
cd client
mvn package
  1. Populate plugin with client source code and dependencies
mkdir ../plugin/lib
cp target/lib/gson-* target/lib/logging-interceptor-* target/lib/ok* target/lib/swagger-annotations-* ../plugin/lib/
rm -r ../plugin/src/java/nl/esciencecenter/e3dchem/kripodb/ws/client
cp -r src/main/java/nl/esciencecenter/e3dchem/kripodb/ws/client ../plugin/src/java/nl/esciencecenter/e3dchem/kripodb/ws/
  1. Update plugin/META-INF/MANIFEST.MF, plugin/ files to reflect contents of lib/

New release

  1. Update versions in pom files with mvn org.eclipse.tycho:tycho-versions-plugin:set-version -DnewVersion=<version>-SNAPSHOT command.
  2. Commit and push changes
  3. Create package with mvn package, will create update site in p2/target/repository
  4. Test node by installing it from local update site
  5. Append new release to 3D-e-Chem update site
  6. Make clone of repo
  7. Append release to 3D-e-Chem update site with mvn install<3D-e-Chem repo/updates>
  8. Commit and push changes in this repo and repo
  9. Create a Github release
  10. Update Zenodo entry
  11. Fix authors
  12. Fix license
  13. To Related/alternate identifiers section add as is cited by this upload entry.
  14. Make nodes available to 3D-e-Chem KNIME feature by following steps at

Create stub recordings for integration tests

The test workflow are tested against a mocked web server and not the actual site. The mock server is called WireMock and normally gives empty responses. To have WireMock server return filled responses, stubs stored in tests/src/test/resources/ directory must be provided. The stubs can be recorded by starting a WireMock server in recording mode by:

java -jar tests/lib/wiremock-standalone-2.5.0.jar --proxy-all="" \
--port=8089 --record-mappings --verbose --root-dir=tests/src/test/resources/
java -jar tests/lib/wiremock-standalone-2.5.0.jar --proxy-all="http://localhost:8084/" \
--port=8089 --record-mappings --verbose --root-dir=tests/src/test/resources/

Then in a KNIME workflow in the KripoDB nodes set the base path to http://localhost:8089. Executing the workflow will fetch data from via the WireMock server and cause new stubs to be recorded in the tests/src/test/resources/ directory.

To run the test workflows from inside KNIME desktop enviroment start the WireMock server in mock mode by:

java -jar tests/lib/wiremock-standalone-2.5.0.jar --port=8089 --verbose --root-dir=tests/src/test/resources/

Then import the test workflows in tests/src/knime/ directory, select the workflow in the KNIME explorer and in the context menu (right-click) select Run as workflow test.