ODAM: Find a Dataverse dataset via Google Search¶

The purpose of this R script is to test if a dataset can be "findable" by a search engine such as Google. For this test, we have chosen our example dataset 'frim1'. The searched keywords are thus 'frim1' associated with 'dataset'. See the code and its results.

Step 1: Start by searching for the keywords in Google Search
Step 2 - If found links in step 1, select those that correspond to a Dataverse dataset
Step 3 - If found some in step 2, then retrieve the metadata

See below the code and its results.

Code

packages <- c('RCurl', 'XML', 'dataverse' )
if (length(setdiff(packages, rownames(installed.packages()))) > 0) {
  install.packages(setdiff(packages, rownames(installed.packages())), 
                    repos='http://cran.rstudio.com')
}

library(RCurl)
library(XML)
library(dataverse)

# Search terms
searchTerms <- "frim1 dataset"
language <- "en"

# Collapse search terms.
entry <- paste(searchTerms, collapse="+")
# Do the search 
siteHTML <- RCurl::getForm("http://www.google.com/search",
                    hl="en", lr="", q=entry,
                    btnG="Search")
# Extract nodes with the links
html <- XML::htmlTreeParse(siteHTML, useInternalNodes = TRUE, error=function (...){})
nodes <- XML::getNodeSet(html, "//div[@class='kCrYT']//a")

# Get all the links
links <- sapply(nodes, function(x) x <- XML::xmlAttrs(x)[["href"]])
links <- links[grep("google.com/search", links, invert = TRUE)]
links <- gsub('/url.q=', '', gsub('&sa.+$', '', links))

# harmonizes the writing of URLs
links <- gsub("%3D", "=", gsub( "%3F", "?", links ) )

# Compile into a data.frame
df <- data.frame(
   label=sapply( XML::getNodeSet(html, "//div[@class='BNeawe vvjwJb AP7Wnd']"), function(x) { XML::xmlValue(x) } ),
   links=links,
   stringsAsFactors=FALSE
)
print(df,right=F)

Output

  label                                                                    
1 FRIM - Fruit Integrative Modelling - Experimental ... - Data Inra        
2 [PDF] pe-pa – tomato frim1 - CBIB                                        
3 Home · inrae/ODAM Wiki · GitHub                                          
4 [PDF] Data Capture ? Daniel Jacob INRA UMR 1332 BFP                      
5 [PDF] Modeling the growth of tomato fruits based on enzyme ... - HAL-CNRS
6 FRIM - Fruit Integrative Modelling | Zenodo                              
7 Rodam package - CRAN                                                     
8 Odam: Open Data, Access and Mining - SlideShare                          
9 Make your data great now - SlideShare                                    
  links                                                                                     
1 https://data.inra.fr/dataset.xhtml%3FpersistentId%3Ddoi:10.15454/95JUTK                   
2 http://services.cbib.u-bordeaux.fr/MERYB/DATA/protocols_upload/1_5_Analytical_177.pdf     
3 https://github.com/INRA/ODAM/wiki                                                         
4 https://hal.archives-ouvertes.fr/hal-02070883/file/Howto_FAIR_DataLifecycle_Aprill2019.pdf
5 https://hal-cnrs.archives-ouvertes.fr/hal-02611223/document                               
6 https://zenodo.org/record/154041                                                          
7 https://cran.r-project.org/web/packages/Rodam/vignettes/Rodam.html                        
8 https://www.slideshare.net/danieljacob771282/odam-open-data-access-and-mining             
9 https://www.slideshare.net/danieljacob771282/make-your-data-great-now

Code

# is there a link to a dataverse site?
ret <- grep( "dataset.xhtml.persistentId=doi", links )
if ( length(ret) ) {
   # Bingo !
   doi <- paste0( 'doi:', gsub("^http.+doi:", "", links[ret[1]]) )
   dataverse <- gsub( "/dataset.+", "", links[ret[1]] )

   # Retrieving Dataset and File Metadata
   Sys.setenv("DATAVERSE_SERVER" = dataverse)
   dataset <- dataverse::get_dataset(doi)

   # Display the dataset information
   dataset

}

Output

Dataset (261843): 
Version: 4.2, RELEASED
Release Date: 2020-06-08T14:57:17Z
License: NONE
17 Files:
             label version    id      contentType
1 datapackage.json       1 99623 application/json
2        frim1.zip       3 96878  application/zip

Code

# Get more metadata
if ( length(ret) ) {
   metadata <- dataverse::dataset_metadata(doi)
   metadata$fields[  ! metadata$fields$multiple | 
                      metadata$fields$typeClass=="controlledVocabulary",
                      c("typeName", "value") ]
}

Output

          typeName                                                                       value
1            title                                          FRIM - Fruit Integrative Modelling
2   alternativeURL                              https://pmb-bordeaux.fr/dataexplorer/?ds=frim1
6       kindOfData                                                                     Dataset
7       dataOrigin                                                           experimental data
8          subject Computer science, Information management, Omics, Plant Health and Pathology
9    lifeCycleStep                                               Study design, Data collection
12        language                                                                     English
14  productionDate                                                                        2010
15 productionPlace                                                INVENIO, Ste Livrade, France
17       depositor                                                               Jacob, Daniel
18   dateOfDeposit                                                                  2018-10-08