fhircrackr 2.0.0

Bugs fixed

• Erroneous warnings when using fhir_search() with token authentication have been fixed.
• fhir_authenticate() doesn't cache a .httr-oauth file anymore.
• fhir_url() now url encodes string type parameter values too.
• fhir_search() can now log errors from FHIR servers regardless of whether they are valid xml.
• fhir_search() can now handle relative next links when paging through bundles

New functions

• fhir_build_bundle(): Build a xml FHIR Bundle from a wide table.
• fhir_build_resource(): Build a xml FHIR resource from a wide table.
• fhir_bundle_list(): Constructor function to create fhir_bundle_list objects is now exported.
• fhir_cast(): Similar to fhir_melt() but spreads multiple entries across columns instead of rows.
• fhir_count_resource(): Count resources matching a specific search request on a server.
• fhir_get_resource_ids(): Extract the logical IDs of all FHIR resources matching a specific search request from a server.
• fhir_get_resources_by_ids(): Extract a set of resources based on their logical IDs.
• fhir_post(): POST an object to a FHIR server.
• fhir_put(): PUT an object to a FHIR server.
• fhir_recent_http_error(): Returns the most recent http error to the console.
• fhir_request(): Wrapper for fhir_url()
• fhir_resource_xml(): Constructor function to create fhirresource_xml() objects is now exported.
• fhir_rm_div(): Remove html elements from FHIR bundles.
• fhir_rm_tag(): Remove elements enclosed by a specified xml-tag from FHIR bundles.
• fhir_sample_resources_by_ids(): Randomly sample resources from a given logical ID vector and download them from a FHIR server.
• fhir_sample_resources(): Randomly sample from a set of resources matching a specific search request in a FHIR server.
• fhir_tree(): Returns a string representing the tree structure implicit in the column names of a wide table.
• pastep(): Concatenates two or more strings to a path string correctly.

New behaviour

• fhir_melt() has been rewritten to be faster and more memory efficient.

• fhir_style: This class is now deprecated, all the information contained in the fhir_style object has been moved to the fhir_table_description.

• fhir_table_description:

  • The structure of a fhir_table_description has changed insofar that the elements that used to be part of fhir_style are now direct elements of fhir_table description.
  • There are two new elements to describe the created table: format and keep_attr. For more information see ?fhir_table_description

• fhir_crack():

  • Rewritten to be faster and more efficient, can now crack bundles in parallel (only on linux) when argument ncores > 1.
  • Can now optionally return tables in wide format where multiple entries are distributed over several columns instead of being pasted together in one column (argument format).
  • Will now extract elements regardless of their xml-attribute (@value, @id, @url), attributes can be attached to the column names when argument keep_attr = TRUE.
  • Argument remove_empty_columns renamed to rm_empty_cols.

• fhir_search():

  • New argument rm_tag="div" per default removes all html bits in the bundle immediately after download. Can be disabled by setting rm_tag=NULL.

  • delay_between_attempts can now be a vector with different waiting times between retries of reaching the server. The length of delay_between_attempts determines the number of retries.

  • max_attempts is deprecated in favor of delay_between_attempts.

• fhir_load() and fhir_save() have a new argument max_bundles that allows to restrict the number of bundles that is loaded/saved.