Commit: ba07279b6dbc2362089e2250ab2829b4ce232ad6 - myprojects-hg-reborn (hg)

Commit MetaInfo

• Revisão: ba07279b6dbc2362089e2250ab2829b4ce232ad6 (tree)
• Hora: 2025-01-23 03:29:21
• Autor: Lorenzo Isella <lorenzo.isella@gmai...>
• Commiter: Lorenzo Isella

Mensagem de Log

A simple script retrieving the data either as a json or as an rdf using a sparql query.

Mudança Sumário

• add: R-codes/retrieve_json_rdf.R

Diff

diff -r c7ac9866c1a9 -r ba07279b6dbc R-codes/retrieve_json_rdf.R
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/R-codes/retrieve_json_rdf.R	Wed Jan 22 19:29:21 2025 +0100
@@ -0,0 +1,97 @@
+rm(list=ls())
+
+library(tidyverse)
+library(janitor)
+library(httr)
+library(jsonlite)
+library(rdflib)
+
+source("/home/lorenzo/myprojects-hg/R-codes/stat_lib.R")
+
+
+fetch_sparql <- function(endpoint, query, return_format = c("json", "rdf")) {
+  # Match the return format to either "json" or "rdf"
+  return_format <- match.arg(return_format)
+  
+  # Set headers and accept type based on desired format
+  accept_header <- if (return_format == "json") {
+    "application/sparql-results+json"  # JSON format for SELECT query results
+  } else {
+    "text/turtle"  # RDF format (e.g., Turtle)
+  }
+  
+  # Perform the HTTP request
+  response <- POST(
+    url = endpoint,
+    body = list(query = query),
+    encode = "form",
+    add_headers(Accept = accept_header)
+  )
+  
+  # Check if the request was successful
+  if (http_error(response)) {
+    stop("Failed to fetch SPARQL results. HTTP status code: ", status_code(response))
+  }
+  
+  # Process response based on the requested format
+  if (return_format == "json") {
+    # Parse JSON response into a list
+    result <- content(response, as = "parsed", type = "application/json")
+  } else if (return_format == "rdf") {
+    # Parse RDF response into an rdflib object
+    rdf_data <- content(response, as = "text")  # Get the raw RDF data
+    temp_file <- tempfile(fileext = ".ttl")    # Save it temporarily as a Turtle file
+    writeLines(rdf_data, temp_file)
+    result <- rdf_parse(temp_file, format = "turtle")  # Load into rdflib
+  }
+  
+  return(result)
+}
+
+
+
+endpoint <- "https://data.europa.eu/sparql"
+query <- "
+PREFIX dcat: <http://www.w3.org/ns/dcat#>
+SELECT * WHERE { ?d a dcat:Dataset } LIMIT 10
+"
+
+
+
+# Fetch as JSON
+json_result <- fetch_sparql(endpoint, query, return_format = "json")
+
+# Check the structure of the JSON result (nested list)
+str(json_result)
+
+json_tibble <- json_result |>
+    simple_json_to_tibble()
+
+# Fetch as RDF
+rdf_result <- fetch_sparql(endpoint, query, return_format = "rdf")
+
+# Print the RDF graph
+print(rdf_result)
+
+# Query the local RDF graph with rdflib
+## local_query <- "
+## PREFIX dcat: <http://www.w3.org/ns/dcat#>
+## SELECT * WHERE { ?d a dcat:Dataset } LIMIT 5
+## "
+
+
+
+
+local_query <- "
+SELECT ?subject ?predicate ?object WHERE {
+  ?subject ?predicate ?object .
+}
+LIMIT 5
+"
+
+
+query_result <-rdf_query(rdf_result, local_query)
+
+print(query_result)
+
+print("So far so good")