Search the Internet Archive, retrieve metadata, and download files.
This API client for the Internet Archive is intended primarily for searching for items, retrieving metadata for items, and downloading the files associated with items. The functions can be used with the pipe operator (
%>%) from magrittr and the data manipulation verbs in dplyr to create pipelines from searching to downloading. For the full details of what is possible with the Internet Archive API, see their advanced search help.
Install this package from CRAN:
devtools::install_github("ropensci/internetarchive", build_vignettes = TRUE)
Then load the package. We will also use dplyr for manipulating the retrieved data.
library("internetarchive")library("dplyr", warn.conflicts = FALSE)
The simplest way to search the Internet Archive is to use a keyword search. The following function searches for these keywords in the most important metadata fields, and returns a list of item identifiers.
ia_keyword_search("isaac hecker")#> 19 total items found. This query requested 5 results.#>  "TheLifeOfFatherHecker" "fatherhecker01sedg"#>  "fatherhecker00sedggoog" "lifeoffatherheck01elli"#>  "lifeoffatherheck00elli"
You can pass an item identifier to the
ia_browse() function to open an item in your browser. If you pass this function multiple identifiers, it will open only the first one.
Usually it is more useful to perform an advanced search. You can construct an advanced search as a named character vector, where the names correspond to the fields. The following search, for instance, looks for items published by the American Tract Society in 1864. Run the function
ia_list_fields() to see the list of accepted metadata fields.
ats_query <- c("publisher" = "american tract society", "year" = "1864")ia_search(ats_query, num_results = 20)#> 3 total items found. This query requested 20 results.#>  "huguenotsfrance00martgoog" "vitalgodlinessa00plumgoog"#>  "sketcheseloquen00wategoog"
You can change the number of items returned by the search using the
num_results = argument, and you can request subsequent pages of results with the
page = argument.
ia_keyword_search() both return a character vector of identifiers, so both can be used in the same way at the beginning of a pipeline.
To search by a date range, use the
date field and the years (or ISO 8601 dates) separated by
TO. Here we search for publications by the American Tract Society in the 1840s.
ia_search(c("publisher" = "american tract society", date = "1840 TO 1850"))#> 88 total items found. This query requested 5 results.#>  "scripturebiogra00hookgoog" "memoirmrssarahl00hookgoog"#>  "circulationandc00socigoog" "historyreformat09aubgoog"#>  "historyreformat22aubgoog"
Once you have retrieved a list of items, you can retrieve their metadata and the list of files associated with the items.
To get a single item's metadata, you can pass its identifier to the
hecker <- ia_get_items("TheLifeOfFatherHecker")
The result is a list where the names of items in the list are the item identifiers, and the rest of the list is the metadata. This nested list can be difficult to work with, so the
ia_metadata() returns a data frame of the metadata, and
ia_files() returns a data frame of the files associated with the item.
These functions can also retrieve the information for multiple items when used in a pipeline. Here we search for all the items about Hecker, retrieve their metadata, and turn it into a data frame. We then filter the data frame to get only the titles.
ia_keyword_search("isaac hecker", num_results = 20) %>%ia_get_items() %>%ia_metadata() %>%filter(field == "title") %>%select(value)
ia_download() function will download all the files in a data frame returned from
ia_files(). This function should be used with caution, and you should first filter the data frame to download only the files that you wish. In the following example, we retrieve a list of all the files associated with items published by the American Tract Society in 1864. Then we filter the list so we get only text files, then we pick only the first text file associated with each item. Finally we download the files to a directory we specify (in this case, a temporary directory).
dir <- tempdir()ia_search(ats_query) %>%ia_get_items() %>%ia_files() %>%filter(type == "txt") %>%group_by(id) %>%slice(1) %>%ia_download(dir = dir, overwrite = FALSE) %>%glimpse()
ia_download() returns a modified version of the data frame that was passed to it, adding a column
local_file with the path to the download files.
overwrite = argument is
FALSE, then you can pass the same data frame of files to
ia_download() and it will download only the files that it has not already downloaded.
rmarkdownas a suggested package since
knitrhas changed for vignettes