enable both big and limited query

R/search_finna.R

@@ -1,35 +1,35 @@
-#' Finna Index Search with Pagination and Advanced Options
+#' Finna Index Search with Total Limit Option
 #'
-#' This function performs a search on the Finna index with extended options, allowing for a wide range of search types, filters, facets, and sorting methods.
-#' It retrieves all available data by paginating through the API results.
+#' This function retrieves records from the Finna index with an option to limit the total number of records returned.
+#' The function paginates through the results, fetching records until the specified total limit is reached.
 #'
 #' @name search_finna
-#' @param lookfor A string containing the search terms. Boolean operators (AND, OR, NOT) can be included.
+#' @param query description
 #' @param type A string specifying the type of search. Options include "AllFields", "Title", "Author", "Subject". Defaults to "AllFields".
 #' @param fields A vector of fields to be returned in the search results. Defaults to NULL, which returns a standard set of fields.
 #' @param filters A vector of filter queries to refine the search. Defaults to NULL.
 #' @param facets A vector specifying which facets to return in the results. Defaults to NULL.
 #' @param facetFilters A vector of regular expressions to filter facets. Defaults to NULL.
 #' @param sort A string defining the sort order of the results. Options include:
-#' \itemize{
-#'   \item "relevance,id asc" (default)
-#'   \item "main_date_str desc" (Year, newest first)
-#'   \item "main_date_str asc" (Year, oldest first)
-#'   \item "last_indexed desc" (Last modified)
-#'   \item "first_indexed desc" (Last added)
-#'   \item "callnumber,id asc" (Classmark)
-#'   \item "author,id asc" (Author)
-#'   \item "title,id asc" (Title)
-#' }
-#' @param limit An integer specifying the number of records to return per page. Defaults to 100 (maximum).
+#'   \itemize{
+#'     \item "relevance,id asc" (default)
+#'     \item "main_date_str desc" (Year, newest first)
+#'     \item "main_date_str asc" (Year, oldest first)
+#'     \item "last_indexed desc" (Last modified)
+#'     \item "first_indexed desc" (Last added)
+#'     \item "callnumber,id asc" (Classmark)
+#'     \item "author,id asc" (Author)
+#'     \item "title,id asc" (Title)
+#'   }
+#' @param limit An integer specifying the total number of records to return across multiple pages.
 #' @param lng A string for the language of returned translated strings. Options are "fi", "en-gb", "sv", "se". Defaults to "fi".
 #' @param prettyPrint A logical value indicating whether to pretty-print the JSON response. Useful for debugging. Defaults to FALSE.
-#' @return A tibble containing all search results with relevant fields extracted and provenance information.
+#' @return A tibble containing the search results with relevant fields extracted and provenance information.
 #' @examples
-#' search_results <- search_finna("sibelius", sort = "main_date_str desc")
+#' search_results <- search_finna("sibelius", sort = "main_date_str desc", limit = 100)
 #' print(search_results)
 #' @export
-search_finna <- function(lookfor,
+search_finna <- function(query = NULL,#lookfor
                          type = "AllFields",
                          fields = NULL,
                          filters = NULL,
@@ -41,30 +41,37 @@ search_finna <- function(lookfor,
                          prettyPrint = FALSE) {
 
   # Handle empty search queries
-  if (lookfor == "" || is.null(lookfor)) {
-    warning("Error: Empty search query provided.")
-    return(NULL)
-  }
+  # if (query == "" || is.null(query)) {
+  #   warning("Error: Empty search query provided.")
+  #   return(NULL)
+  # }
 
   # Define the base URL for the search API
   base_url <- "https://api.finna.fi/v1/search"
 
-  # Initialize variables for pagination
-  all_data <- list()  # Store all pages of data
-  page <- 1           # Start from the first page
+  # Initialize empty list to store all records
+  all_data <- list()
+
+  # Define the pagination parameters
+  total_fetched <- 0
+  page <- 1
+  records_per_page <- 100  # Fetch 100 records per page for efficiency
 
-  repeat {
-    # Construct the query parameters for each page
+  while (total_fetched < limit) {
+    # Calculate the remaining number of records to fetch
+    remaining_to_fetch <- min(records_per_page, limit - total_fetched)
+
+    # Construct the query parameters for the current page
     query_params <- list(
-      lookfor = lookfor,
+      lookfor = query,
       type = type,
       `field[]` = fields,
       `filter[]` = filters,
       `facet[]` = facets,
       `facetFilter[]` = facetFilters,
       sort = sort,
       page = page,
-      limit = limit,
+      limit = remaining_to_fetch,  # Set the page limit to fetch only the remaining records
       lng = lng,
       prettyPrint = prettyPrint
     )
@@ -78,87 +85,83 @@ search_finna <- function(lookfor,
       }
     )
 
-    # Check if the response is valid
-    if (is.null(response) || httr::status_code(response) != 200) {
-      error_message <- sprintf("Failed to perform the search. Status code: %d - Response: %s",
-                               httr::status_code(response), httr::content(response, "text"))
-      warning(error_message)
-      break
-    }
+    # Process the response based on the status code
+    if (httr::status_code(response) == 200) {
+      # Parse the JSON content of the response
+      search_results <- httr::content(response, "parsed")
 
-    # Parse the JSON content of the response
-    search_results <- httr::content(response, "parsed")
-
-    # Extract and structure relevant data from the search results
-    records <- search_results$records
-    if (length(records) == 0) {
-      message("No more records found. Stopping pagination.")
-      break
-    }
+      # Extract and structure relevant data from the search results
+      records <- search_results$records
+      if (is.null(records) || length(records) == 0) {
+        break  # No more records, stop fetching
+      }
 
-    data <- lapply(records, function(record) {
-      list(
-        Title = record$title %||% NA,
-        Author = if (!is.null(record$nonPresenterAuthors) && length(record$nonPresenterAuthors) > 0) {
-          paste(sapply(record$nonPresenterAuthors, function(author) author$name), collapse = ", ")
-        } else {
-          NA
-        },
-        Year = record$year %||% NA,
-        Language = if (!is.null(record$languages) && length(record$languages) > 0) record$languages[[1]] else NA,
-        Formats = if (!is.null(record$formats) && length(record$formats) > 0) {
-          paste(sapply(record$formats, function(format) format$translated), collapse = ", ")
-        } else {
-          NA
-        },
-        Subjects = if (!is.null(record$subjects) && length(record$subjects) > 0) {
-          paste(sapply(record$subjects, function(subject) paste(subject, collapse = ", ")), collapse = "; ")
-        } else {
-          NA
-        },
-        Library = if (!is.null(record$buildings) && length(record$buildings) > 0) {
-          paste(sapply(record$buildings, function(building) building$translated), collapse = ", ")
-        } else {
-          NA
-        },
-        Series = tryCatch({
-          if (!is.null(record$series)) {
-            if (is.list(record$series)) {
-              if (length(record$series) > 0) {
-                paste(sapply(record$series, function(series) series$name %||% NA), collapse = ", ")
+      # Add the records to the all_data list
+      data <- lapply(records, function(record) {
+        list(
+          Title = record$title %||% NA,
+          Author = if (!is.null(record$nonPresenterAuthors) && length(record$nonPresenterAuthors) > 0) {
+            paste(sapply(record$nonPresenterAuthors, function(author) author$name), collapse = ", ")
+          } else {
+            NA
+          },
+          Year = record$year %||% NA,
+          Language = if (!is.null(record$languages) && length(record$languages) > 0) record$languages[[1]] else NA,
+          Formats = if (!is.null(record$formats) && length(record$formats) > 0) {
+            paste(sapply(record$formats, function(format) format$translated), collapse = ", ")
+          } else {
+            NA
+          },
+          Subjects = if (!is.null(record$subjects) && length(record$subjects) > 0) {
+            paste(sapply(record$subjects, function(subject) paste(subject, collapse = ", ")), collapse = "; ")
+          } else {
+            NA
+          },
+          Library = if (!is.null(record$buildings) && length(record$buildings) > 0) {
+            paste(sapply(record$buildings, function(building) building$translated), collapse = ", ")
+          } else {
+            NA
+          },
+          Series = tryCatch({
+            if (!is.null(record$series)) {
+              if (is.list(record$series)) {
+                if (length(record$series) > 0) {
+                  paste(sapply(record$series, function(series) series$name %||% NA), collapse = ", ")
+                } else {
+                  NA
+                }
+              } else if (is.atomic(record$series)) {
+                as.character(record$series)
               } else {
                 NA
               }
-            } else if (is.atomic(record$series)) {
-              as.character(record$series)
             } else {
               NA
             }
-          } else {
-            NA
-          }
-        }, error = function(e) NA)
-      )
-    })
+          }, error = function(e) NA)
+        )
+      })
 
-    # Append the current page's data to the list of all data
-    all_data <- c(all_data, data)
+      all_data <- c(all_data, data)
 
-    # Check if we've reached the last page
-    if (length(records) < limit) {
-      message("Retrieved last page of results.")
-      break
-    }
+      # Update the total number of fetched records
+      total_fetched <- total_fetched + length(records)
+      page <- page + 1
 
-    # Increment the page number for the next iteration
-    page <- page + 1
+    } else {
+      # Handle API errors with detailed messages
+      error_message <- sprintf("Failed to perform the search. Status code: %d - Response: %s",
+                               httr::status_code(response), httr::content(response, "text"))
+      warning(error_message)
+      return(NULL)
+    }
   }
 
-  # Convert the collected data into a tibble for easy analysis
+  # Convert the list of extracted data into a tibble for easy analysis
   tibble_results <- tibble::as_tibble(do.call(rbind, lapply(all_data, function(x) unlist(x, recursive = FALSE))))
 
   # Attach the language attribute to the tibble
   attr(tibble_results, "language") <- lng
-  cat("Data retrieved from Finna API (https://www.finna.fi) - metadata licensed under CC0.\n")
+  #cat("Data retrieved from Finna API (https://www.finna.fi) - metadata licensed under CC0.\n")
   return(tibble_results)
 }

_pkgdown.yml → pkgdown/_pkgdown.yml

@@ -21,12 +21,15 @@ navbar:
         href: news/index.html
     right:
       - text: "GitHub"
+        icon: fa-github-code-o
         href: https://github.com/rOpenGov/finna
 
 # Optional: Define the authors
 authors:
   - name: "Akewak Jeba"
     url: https://github.com/rOpenGov/finna
+  - name: "Leo Lahti"
+    url: https://github.com/rOpenGov/finna
 
 # Optional: Specify the output directory (if different from default)
 output_dir: docs