INSTRUCTIONS

    $  bundle install

    $  gem build

    $  gem install yarrrml_template_builder-{VERSION}.gem

VERSION is some version that will appear in the filename of the gem.

Execute a CDE Transformation from CSV

0) Default folder structure, relative to where you will run the transformation script:

    .
    ./data/   (this folder is mounted into sdmrdfizer - see step 1 below)
    ./data/mydataX.csv  (input csv files)
    ./data/mydataY.csv...
    ./data/triples/  (output FAIR data ends up here)
    ./config/
    ./config/***_yarrrml_template.yaml (*** is a one-word tag of the "type" of data, e.g. "height")

1) Need to have sdmrdfizer and yarrrml-parser services running ./data mounted as /data and ./config as /config. You can use docker-compose to run both services:

```yaml
version: "2.0"
services:


    yarrrml_transform:
        image: markw/yarrrml-parser-ejp:latest
        container_name: yarrrml_transform
        ports:
            - "3000:3000"
        volumes:
            - ./data:/data


    rdfizer:
        image: markw/sdmrdfizer_ejp:0.3.0
        container_name: rdfizer
        ports:
            - "4000:4000"
        volumes:
            - ./data:/data
            - ./config:/config
```

2) Create your template by running build_xxx_template or select the appropriate YARRRML template (in the ./config folder, e.g. "height_yarrrml_template.yaml"). See "example of creating a template" below if you need to create one from-scratch.

3) In the ./data folder, create a CSV file with the necessary headings for your desired transform.

4) Identify the "tag" of the YARRML template you want to use (e.g. 'height' for "height_yarrrml_template.yaml"). This tag is used to coordinate between many of the components during the automation steps, so it must match exactly with the "tag" portion of the template name.

5) Execute the transformation: You can use run_me_to_test.rb by changing the tag described as datatype_tag parameter at YARRRML_Transform:

```ruby
    y = YARRRML_Transform.new(datafile: "./data/myHeightData.csv", datatype_tag: "height")
    y.yarrrml_transform
    y.make_fair_data   # output goes to ./data/triples
```

Example of creating a template:

General Discussion:

The objective of the library is to make it easy to generate a YARRRML template that matches your situation. There is a fair amount of flexibility in what parameters are used, which are optional, and which have defaults (SIO-compliant). NOTE: What comes out of the yarrrml_template_builder is NOT a complete YARRRML file! It has place-holders for the source filename and reference formulation. you add those when calling the yarrrml_transform library.

personid_column and unique_id column are absolutely mandatory, and do not have a default.

There are three ways of representing a data column (or its default, if that column doesn't exist):

You can specify the default value (these parameters are designated "xxxx" for example "process_type".
You can specify the column header. These parameters are designated "xxxxxx_column" - for example 'process_type_column'. Specifying "process_type_column" will override any value you provided in "process_type"
You can specify nothing, in which case the internal defaults will be used (based on the "base" type for that node in SIO. Some data is allowed to be nil, and those nodes will not be created.

Almost all cases will allow any of those three methods (see detailed documentation for more)

"tags" are used for creating human-readable section names in the output YARRRML. They may not contain spaces or other odd characters... letters and underscores are fine. This is not currently sanity-checked, so...

With that said, here's what a template-building script looks like:

require "yarrrml-template-builder"

# this is the header of the CSV we are going to transform
# "pid,uniqid,qualityURI,qualityLabel,measurementURI,measurementLabel,processURI,processLabel,height,unitURI,unitLabel,date,comments"

b = YARRRML_Template_Builder.new({
  baseURI: "http://marks.test/this/", # this should resolve to wherever you are going to serve the data from.
                                      # RDF URLs become e.g. http://marks.test/this/individual_X_Y#patientRole
  source_name: "patient_height"}
  )

b.person_identifier_role_mappings({
                                    personid_column: "pid",
                                    uniqueid_column: "uniqid",
                                    identifier_type: "https://ejp-rd.eu/vocab/identifier",
                                    person_type: "https://ejp-rd.eu/vocab/Person",
                                    person_role_tag: "patientRole",
                                    role_type: 'obo:OBI_0000093'
                                    role_label: "Patient"
                                    })
b.role_in_process({
    process_type_column: "processURI",  
    process_tag:  "some_height_measuring_process",
    process_label_column: "processLabel", 
    process_start_column: "date", 
    process_end_column: nil,
    })

b.person_has_quality({
    quality_type_column: "qualityURI",  
    quality_tag:  "height_quality",
    quality_label_column: "qualityLabel", 
    })

b.process_hasoutput_output({
    output_nature: "quantitative",
    output_type_column: "measurementURI",
    output_type_label_column: "measurementLabel",
    output_value_column: "height",
    output_value_datatype: "xsd:float",
    output_comments_column: "comments",
    })

b.output_has_unit({
    output_unit_column: "unitURI",
    output_unit_label: "unitLabel"  
})

puts b.generate

Note that this will output the template to STDOUT, so capture it to a file. For everything else to work "correctly" that file should be named XXXXXX_yarrrml_template.yaml, and should live in the ./config folder. XXXXXX is the 'tag' of the data type (e.g. 'height')

The "run_me_to_test.rb" script will run as a demo - using the demo height data in the ./data folder, and the template that is created by the code above. Output will appear in the ./data/triples folder.