Skip to contents

Query Wikipedia using the MediaWiki Action API

Source: R/wiki-action-request.R
wiki_action_request.Rd

Wikipedia exposes a To build up a query, you first call wiki_action_request() to create the basic request object, then use the helper functions query_page_properties(), query_list_pages() and query_generate_pages() to modify the request, before calling next_batch() or retrieve_all() to perform the query and download results from the server.

Usage

wiki_action_request(..., action = "query", language = "en")

Arguments

...

<dynamic-dots> Parameters for the request

action

The action to perform, typically 'query'

language

The language edition of Wikipedia to request, e.g. 'en' or 'fr'

Value

An action_api object, an S3 list that subclasses httr2::request. The dependencies between different aspects of the Action API are complex. At the time of writing, there are five major subclasses of action_api/httr2_request:

  • generator/action_api/httr2_request, returned (sometimes) by query_generate_pages

  • list/action_api/httr2_request, returned by query_list_pages

  • titles, pageids and revids/action_api/httr2_request, returned by the various query_by_ functions

    You can use query_page_properties to modify any kind of query except for list queries: indeed, the central limitation of the list queries is that you cannot choose what properties to return for the pages the meet the given criterion. The concept of a generator is complex. If the generator is based on a property module, then it must be combined with a query_by_ function to produce a valid query. If the generator is based on a list module, then it cannot be combined with a query_by_ query.

Details

wikkitidy provides an ergonomic API for the Action API's Query modules. These modules are most useful for researchers, because they allow you to explore the structure of Wikipedia and its back pages. You can obtain a list of available modules in your R console using list_all_property_modules(), list_all_list_modules() and list_all_generators(),

See also

gracefully()

Examples

# List the first 10 pages in the category 'Australian historians'
historians <- wiki_action_request() %>%
  query_list_pages(
    "categorymembers",
    cmtitle = "Category:Australian_historians",
    cmlimit = 10
  ) %>%
  gracefully(next_batch)
historians
#> <complete/query_tbl>
#> ℹ There are more results on the server. Retrieve them with `next_batch()` or `retrieve_all()`
#> ✔ Data complete for all records
#> # A tibble: 10 × 3
#>      pageid    ns title                    
#>       <int> <int> <chr>                    
#>  1 72000612     0 Michelle Arrow           
#>  2 74445832     0 Alan Atkinson (historian)
#>  3 46828642     0 Craig Benjamin           
#>  4 53702558     0 Frank Murcott Bladen     
#>  5 59403076     0 Frank Bongiorno          
#>  6 31593145     0 R. J. B. Bosworth        
#>  7 23093698     0 Tim Bowden               
#>  8 22906640     0 Phillip Bradley          
#>  9 33224431     0 Richard Broome           
#> 10 68287945     0 David Brophy (historian)