Add a new data source: generic TSV data source

[julie-sullivan]

• accepts TSV or CSV
• column mappings come from the project XML

Here is a file snippet

Gene    Gene name       Sample  Value   Unit
ENSG00000000003 TSPAN6  adipose tissue  31.5    TPM
ENSG00000000003 TSPAN6  adrenal gland   26.4    TPM
ENSG00000000003 TSPAN6  appendix        9.2     TPM

Assumes the model is correct

[julie-sullivan]

How do we get the Gene <--> RNASeqResult relationship?

[julie-sullivan]

Can go through the model and see which relationships have a type of RNASeqResult? NO Shouldn't guess.

These should also be in the project XML somehow. How?

<property name="rel3" value="Gene.RNASeqResults"/>

[sammyjava]

Nice. I have scads and scads of special-purpose TSV sources. Although most of those have the data model constraints written into them as well, so it's hard to envision a generic source that would handle them. But nice to have one in any case.

[danielabutano]

This implementation is pretty simple:

1. it reads generic-separated-values files (TSV or CSV)
2. it loads more than one class (in the example below Gene, Organism and Protein)
3. it's able to create properly one to many, many to many relations between the entities loaded
4. it doesn't require to set the relations between classes. Given the classes configured in the properties file, it deduces the relations from the model.

Note: Unfortuanlly with the current implementation/configuration, it's not possible to load two entities with the same type (e.g. protein and its isoforms, which are Protein too) because there is no way to distinguish the protein from the isoform protein

Configuration example:

<source name="delimited" type="delimited">
  <property name="delimited.dataSourceName" value="TSV Source Name"/>
  <property name="delimited.dataSetTitle" value="TSV Data Set"/>
  <property name="delimited.licence" value="http//usemydatalicence.com"/>
  <property name="delimited.hasHeader" value="true"/>
  <property name="delimited.columns" value="Gene.primaryIdentifier, Organism.taxonId, null,Protein.primaryIdentifier,Protein.primaryAccession"/>
  <property name="delimited.includes" value="test.tsv"/>
  <property name="src.data.dir" location="/home/daniela/Projects/data/malaria/tsv/"/>
</source>

Mandatory poperties:

• delimited.dataSourceName
• delimited.dataSetTitle
• delimited.columns

Optional poperties:

• delimited.separator (tab/TAB/comma/COMMA). Default value is: tab
• delimited.hasHeader (true/false). Default value is: true