TET is platform independent and can be deployed in a verity of environments. The preferred environment for TET is Ubuntu Linux 12.04 64bit which has been used for development, deployment and testing of the TET. TET user interface for CKAN requires CKAN instance to be installed on the same or on a remote machine. CKAN requires a machine with python, java virtual machine, apache solar and instance of Postgres or MySQL. Details related CKAN installation can be found on CKAN website. Apart from standard CKAN installation TET requires DataStore and some TET extension plugins to be installed on CKAN.
- CKAN instance with DataStore and DataStore API enabled
- (Optional) SPOD instance
- The following plugins should be installed and enabled on the CKAN instance:
- Optionally you can install the following CKAN plugins:
- Install TET user interface using instructions available on: https://github.com/routetopa/tet
- In tet/settings.py make sure that CKAN_URL and SPOD_URL points to appropriate CKAN and SPOD instances
- Run the server:
python manage.py runserver
Migration, software versions and metadata schema mismatch are common issues that can be encountered during the deployment process. Migration data could be tricky and depends on the source platform. Most of the open data platforms have implemented the harvesting feature that enables transfer of the data from one platform to another. In some cases, custom scripts development may be required in order to transfer the data. TET has been built on CKAN in version 2.6, and some features may not be compatible with its previous releases. Therefore, in some cases, CKAN version needs to be upgraded in order to maintain compatibility with developed extensions. Each organization have they own way of describing metadata associated with the data. TET defines a metadata schema that is compliant with the W3C recommendations for data on the web. Changes may be required in order to adopt the TET schema ideally. However, it is also possible to configure the TET schema in such a way, that it could be easily adopted with exiting metadata schema. Categories and roles are essential metadata fields for enabling personalized search and recommendations. These fields could be missing, or list of categories and roles may be different than provided. While the categories issue can be easily fixed by changing the settings in the configuration file, the roles issue would either require manual data entry or the development of a custom script.