@wbg-mde/r-factory/README.MD

Metadata editor R integration module

# Metadata Editor Import(r-factory)

R integration module of the Metadata Editor application. This module contains R scripts, Node based R utility methods and test cases. This module passing data from NodeJS to R by using [PanApps](http://services.panapps.co/) customized version of [r-script](https://www.npmjs.com/package/@wbg-mde/r-script). This module includes various data import export features as well as data analytic features such as resequence, spread metadata etc. Here are the list features..

- import dataset from file formats `SPSS / STATA / CSV`
- export dataset to different file formats
- destring straing variables
- resequence variables
- spread metadata
- export to dictionary
- update variable status
- calculate variable statistics

### Prerequisites

Install Node and R if not installed. Set environment variable for windows. 

- Node 10.15.1
- R version 3.3.3

Check whether R packages are installed and the version. If not please install using the command `install.packages("package_name")`

##### R packages 
- jsonlite (version: 1.3)
- haven (version: 1.1.0)
- plyr (version: 1.8.4)
- stringr (version: 1.2.0)
- labelled (version 1.0.0)
- readr (version 1.1.1)

### Installation

Install the dependencies and devDependencies.

```sh
npm install
```

Build the application
```sh
npm run build
```

Test the application
```sh
npm run test
```

Publish the application to npm
```sh
npm publish --access public
```

### Running the tests

Unit test are written for each features. You can copy input files to `test-data/input` directory. Please see the commands to run unit test below.

Note:- 
Please start the editor before run the tests. Editor start the OpenCPU API server and it will be used in the unit test.

#### `npm run test` 

unit test to check the dataset import/export functionalities. Keep only dataset files to be tested in the `test-data/input/dataset` folder, remove other files.

known issues - some datasets may fail the unit tests due to labelled integer validation while exporting to STATA dataset format(eg: cs1_pupil.dta)

flow of test execution :-

- import datasets from `test-data/input/datasets` directory
- export the imported files to `test-data/output/datasets`
- import the exported datasets


#### `npm run test:resequence`

unit test to import dataset and perform resequence on the imported datasets. Drop the dataset files to be tested in the `test-data/input/dataset` folder and run command

flow of test execution :-

- import datasets from `test-data/input/datasets` directory
- perform resequence and write updated varable json file to `test-data/output/json` directory


#### `npm run test:destring`

unit test to check destring functionality in the imported file. Since we have to mention the variables to be destringed, the test is limited for a particular dataset "ghs_2015_person_v1.1_20160608.dta". Keep this file in the input folder and remove others while run the test.

flow of test execution :-

- import datasets from `test-data/input/datasets` directory
- perform destring to the selected variables and write the updated csv file to `test_data/output/csv/` directory

#### `npm run test:dictionary`
#### `npm run test:dictionary:stata`
#### `npm run test:dictionary:spss`

unit test for export to dictionary format.

flow of test execution :-

- import datasets from `test-data/input/datasets` directory
- export dataset to `test_data/data-dictionary/` directory

### `npm run test:validateKey`

unit test to check the unique key constraint for the given key variable of a dataset.

Steps :-

- Copy the data file to `test-data/input/datasets` directory. 
- Set the data `datasetname` and `keyVariables` in `dist/test/validation.unit.test.js` constructor method
- run the command

flow of test execution :-

- import dataset from `test-data/input/datasets` directory
- validate the key variables

### Contributors
- Navin VI (navin.v.i@panapps.co)
- Anoop Xaviour (anoopx@panapps.co)
- Libin Thomas (libint@panapps.co)

License
----
MIT