v0.8.0

• removed a lot of the functions that we don't use (eg get_orcid_*) or have overlapping functionality (eg. set_abstract)
• deprecated the functions (eg. eml_contact) previously mentioned in the datateam training manual for a gentler transition in the case some interns still used them
• switching over to Github Actions

Fixes

• eml_otherEntity_to_dataTable some of the unboxing issues fixed
• eml_nsf_to_project handles different last names better now

New functions

• title_to_file_name
• get_coord_list
  -eml_get_raster_metadata

EML2 Update

This is a major release, refactoring the code to be compatible with the major changes in EML.

All of the metadata helper functions, and some publishing functions, were updated to be compatible with the new version.

Additionally, three new functions were added

• get_orcid_email
• eml_get_simple
• reorder_pids

v0.6.4

This is a significant update, but not crucial. It enhances usability and remedies some nagging issues.

NEW

• check for "https" in sysmeta update functions
• helper for setting project data
• logic for multiple roles
• check to see if sysmeta fileName is already equal to user input
• Travis CI support
• eml_abstract helper
• if statement in set_public_read
• EML object as an alternative option for metadata_path in publish_update
• error checking logic for personnel passed to eml_project
• arguments for studyAreaDescription and designDescription
• example for resource map parsing

FIXED

• updated error catches to include messages
• updated eml@packageId element in update_object
• replaced length() with nchar() in set_abstract since we are evaluating if a String is >0 characters
• missing docs for update_object
• modified publish_update to only remove eml@access if metadata_path is not null
• updated eml_party to include case for personnel
• removed validation logic for incorrect test (role is not required to be "originator")
• removed resource_map_pid from parent_child_pids arg
• tests and built .Rds for projects
• various whitespace and line length issues
• roxygen doc issues with util.R
• changed ::: to :: in call to getTriples (was a mistake)
• roxygen tag issue (was causing a check failure)
• bug in call to as (introduced in an earlier commit)
• broken eml_project example
• example code for set_abstract
• tests so they run on Travis
• removed export from a few functions that are obsolete, added examples, fixed param and return values
• various documentation errors
• RD warnings and errors
• regression from removing remove_public_access
• variety of R CMD CHECK issues
• bug in test code that prevents environment from being loaded correctly
• pid_to_eml_datatable to accept attributes and validateAttributes as optional arguments
• whitesace issues in pid_to_eml_datatable
• reduced dependencies on ncdf4 and yaml for Travis CI, thanks to @isteves for this one

v0.6.3

This is a minor patch release and is a recommended update for all users.

CHANGED

• publish_update and update_resource_map no longer preserve provenance information across resource map updates as they had been doing for the last few months. The consequence for users of these two functions is that any provenance information will be wiped out and have to be re-added between updates. If you want to preserve provenance information, please use the relevant methods in dataone/datapack to update your packages and avoid publish_update except in cases where the package has no provenance information in it.

v0.6.2

This is a minor update and it is not critical for users to update right away unless they want the new functionality.

NEW

• The package no longer stores a list of DataONE format IDs and instead loads them from the CN when the package is loaded. This should prevent the package's list getting outdated as the CN's list is changed.
• The package (once again) checks to make sure it's the latest installable version and warns the user if their version of out of dated

FIXED

• Usages of class(x) == y have been changed to use is instead thanks @ThomasThelen

v0.6.1

This is a very minor release, mainly just to get the new userId functionality out there for others to use.

NEW

• The EML party helpers (eml_party, and its helperseml_creator, eml_contact, eml_associated_party now take an optional userId argument which takes an ORICD identifier. Thanks @jeanetteclark 🌮 🎉 🌮

  Use it like eml_creator("Bryce", "Mecum", userId = "https://orcid.org/0000-0002-0381-3766")

CHANGED

• The package should produce little or no arcticdata.log files. These were helpful in the past but are not necessary any longer and really just cluttered things up for everyone

v0.6

This release adds some handy new functions (see NEW) but, importantly, we've changed the default replication policy to FALSE, numReplicas to 0, and the set of blocked MNs to KNB and UCSB1. As such, this is a strongly recommended update for all users.

FIXED

• get_all_versions should be much faster now
• A number of documentation fixes have been made
• eml_project should now set funding numbers better
• eml_creator, eml_contact, and other EML party functions have been re-written and should work better now

NEW

• publish_update, update_object, create_object, create_resource_map, and update_resource map now all set a default replication policy of allowed = FALSE, numReplicas = 0, and sets KNB and mnUCSB1 as blocked replication targets
• New function mdq_run: You can now run your metadata documents through the Quality engine from R, see mdq_run
• New function show_indexing_status: You can now query the indexing status of all PIDs in an updated package with show_indexing_status

IMPROVED

• update_object and publish_object now do sanity checks on the format_id they are passed
• publish_update now does even more sanity checking of its arguments. These extra checks shouldn't slow anything down in a noticeable way

v0.5.13

NEW

• Added a new helper function pid_to_eml_physical to streamline producing physical objects from PIDs

CHANGED

• The {sysmeta|pid}_to_eml_{other_entity|physical} functions are now vectorized.
• Deprecation: sysmeta_to_other_entity has been renamed to sysmeta_to_eml_other_entity to keep names more consistent across the helper functions
• Deprecation: pid_to_other_entity has been renamed to pid_to_eml_other_entity to keep names more consistent across the helper functions

FIXED

• publish_object no longer defaults to a format ID of application/xml for files that end with .xml. This was causing issues in the past where EML XML documents were getting the wrong format ID and the uploader wasn't noticing until later on.
• get_package should now produce more helpful error messages in some cases where it didn't previously

v0.5.11

CHANGED

• get_package now uses slightly different logic than before. Previously, it would first look for resource maps and then filter out any obsoleted ones. This caused packages in the middle of a version chain to be not findable by get_package (You would've had to call get_package on the most recent version). Now this is fixed and you should be able to call get_package on any package in a version chain.
• eml_add_other_entities has been renamed to eml_add_entities and now supports adding all entity types in one run, i.e., you can add a mix of dataTable and otherEntity elements to your document in one call.
• profile renamed to view_profile to avoid conflicting with built-in functions and make the name a little more specific

FIXED

• Routines creating EML physical sections now add the unit attribute (with value 'bytes') to object sizes. This isn't strictly required by EML but it's better to be explicit.

v0.5.10

NEW

• A new helper function eml_add_other_entities has been added. Use this to quickly add EML otherEntity elements to an EML document from files you have on disk

FIXED

• A few documentation and code bugs were fixed