##Introduction Repository for DPUs (Data processing units) for UnifiedViews ETL tool (https://github.com/UnifiedViews/Core).
##For Users
###DPU List
####Domain specific Prepared for 1.3.0
| Name | Type | Location | Description |
|---|---|---|---|
| e-sukl | Extractor | dpu-domain-specific/e-sukl | |
| e-isvav | Extractor | dpu-domain-specific/e-isvav |
#####General
| Name | Type | Location | Description |
|---|---|---|---|
| t-tabular | Transformer | dpu/t-tabular | |
| l-solr | Loader | dpu/l-solr | |
| t-rdfStatementParser | Transformer | dpu/t-rdfStatementParser |
This manual assumes that you have already downloaded and installed libraries from https://github.com/UnifiedViews.
- download this repository
- execute tools/init.sh
- create new DPU form maven archeotype dpu-template-advanced
- if the target instance does not contain libs as osgi-bundles you must update dependency scope from provided to compile
##For developers (Please READ before contributing) Purpose of specific directories.
###Libs The Libs directory contains useful project that can be used by DPUs:
- boost-dpu
- service-external - provides classes for manipulation with external services (query http sparql).
- service-serializaiton-xml - provides service for xml serialization
- utils-dataunit
- utils-dataunit-rdf - wrap for RDFDataUnit, that makes work with rdf easier
- utils-test - load/store methods for RDFDataUnit.
###Dependencies Contains DPU's dependencies (jar libraries - osgi bundles). If DPU needs some osgi-library then the library should be located here and there should be lib/libs.txt file (line oriented) that contains list of required dependencies in the DPU folder, e.g., as here: https://github.com/mff-uk/DPUs/tree/master/dpu/loader-scp.
###New DPUs There are three directories, where the DPUs are stored:
- dpu-deprecated - this DPUs should not be used and should be replaced by other DPUs as they are no longer maintained.
- dpu-domain-specific - domain specific DPUs, mostly extractors.
- dpu - general DPUs. If DPU is located here then is should have potential to be used in multiple pipelines across different domains.