- Make
get_field_index()more efficient by utilizing vectors instead oflapply(). - Adding liesWithinLayer in
get_field_index()andsimplify_idaifield()is now an option. This way, inget_field_index()the "shortDescription" can be added without the more time intensive recursive layer search that comes with the now available optionfind_layers = TRUE. For info on what this does, see documentation offind_layer(). simplify_idaifield()is now slightly faster as it usesfind_layer()differently.find_layer()can now handle a vector of multiple ids.- The
project/projectname-argument will be removed soon, and project needs to be specified inconnect_idaifield(). Relevant functions currently issue a warning to that effect.
- Add
idf_last_changed(connection, n = n): Returns a vector with the last n changed resources in the database. - Add
idf_get_changes(connection, ids = c(...)): Returns a data.frame in which each change to one of the resources listed in ids (can be either their 'identifier' or 'UUID') corresponds to one row separated by creation or modification. remove_config_names()issues a message about duplicate field or category names if it is appropriate. Functions may override data when multiple columns with the same name would otherwise exist, as e.g. 'diameter' and 'projectName:diameter'. Attaches an attribute that lists the duplicate field/category names.
- Notify of new / different version on GitHub on attaching the package (
check_idaifieldr_version()).
- Fix problem in
reformat_geometry()(MultiPolygons have to be unnested before processing). (Imported Polygons may be formatted improperly, unnest if necessary (reformat_geometry()). Geometry is still a work in progress.) - Fix a bug in
get_field_index(), where it would return an empty data.frame if there was no configuration-resource.
- add option to use or not use exact dates as min/max values for dating if present
- add
idf_json_query()which lets users construct their own queries to the CouchDB-API - improve and export
find_layer() - fix error in
get_language_lookup() - fix bug in
idf_query()-function where it would not get values from fields with possibility of multiple entries - remove config as attribute to save space
- rename
select_by()toidf_select_by() - rename
show_categories()toidf_show_categories() - update
idf_select_by()andidf_show_categories() - update vignettes
- update/enhanced documentation
- add a sticker
- of course, I missed half of it.
- fix
idf_query()/idf_index_query(): rename "type" to "category" in resource lists - type column in
idaifield_as_matrix() - fix renaming of type/category in
simplify_idaifield() - fix missed type/category issue
convert_to_onehot() - "solve" minor problem with spreading checkbox fields (#TODO)
- remove sofa as dependency and add custom interface with reduced functionality
- restructure unnesting (
unnest_docs()&find_resource()) - add
get_field_index()- getting the uidlist/index directly from the database - exchange type with category everywhere (Attention: This may very well break previous scripts)
- speed up
find_layer()significantly - check if project exists before attempting to query the database
- "type" does not refer to the type of resource (Find, Layer, Place etc.) any more and is renamed to category, as this is the current structure of the database. The uidlist-column "type" is renamed to "category" along with any reference to this information anywhere else (I hope ;) ).
- handle / reformat dating fields by getting min and max date
- get type OR category if type is empty in for
get_uid_list() - multiple queries for
idf_query()with type & category - in
simplify_idaifield(), category is currently switched to type, but this will be changed later -> #TODO: rename type everywhere to category to reflect actual db structure
- add language management for multi-language input fields when project has more than one project language
- add language list lookup preparation (
get_language_lookup()) for custom config fields - add
download_language_list()to get current translations from GitHub - add ping-checks for all database-related functions to supply custom error messages
get_idaifield_docs()will now remove the Configuration as a doc and attach it as attribute to all custom classes, thussimplify_idaifield()will use the config-attribute directly and works without a connection.simplify_idaifield()now allows to keep checkbox-fields etc. as-is- change
idf_query()andidf_index_query()to return docs instead of simplified list to allow choice - all
idaifield_...-lists are now named (with the identifier of each resource) - new demodata for future use (
data("idaifieldr_demodata")) - rewritten vignettes and generally updated documentation
- Notice: The changes may have created problems for iDAI.field 2, and I did not test using the package with iDAI.field 2, as I assume it is not used anymore.
get_idaifield_docs()should still work.
- config specific names now removed everywhere
- fixed problem with if-condition when config exists
- fixed problem with the calculation of means for ranged measurements
- comments in the code in some places
- speed up
find_layer(),replace_uid()andcheck_if_uid() - more tests
- Fixed
find_layer()(internal), seems to work as intended now. - With the editor for the project configuration in iDAI.field 3, new fields and objects started to be named according to their configurations, e.g. "milet:temperType" etc., if the fields have been newly created. From now on those config specific names are removed from all fields, so that only the part after the double dot remains (e.g. "milet:temperType" becomes "temperType").
check_if_uid()now handles vectors- Speed up
get_uid_list(),replace_uid(),convert_to_onehot(), andfix_relations()(a lot) - split up import with
get_idaifield_docs()and processing to be more usable withsimplify_idaifield(), with advice to use the latter only on subsets so the time is more manageable. - add
check_and_unnest()to export
- Version number changed to 0.2, to reflect the rather dramatic changes.
- Faster
get_uid_list()usinglapply() - Added
idf_query()to specifically query the db for groups without first downloading everything - Added
idf_index_query()to specifically query the db for things that are only available in the uidlist (needs a uidlist to do that) simplified = TRUE(simplify_idaifield()) will now convert checkbox field to multiple columns. I should probably make this an option in the future, but I need to clean up that function first (TODO).- make it possible to import the configuration file (needed for some new functions; will only work with idaifield3 and up, hopefully); connection and projectname are now attributes of the "idaifield_docs" and "idaifield_resources" objects.
idaifield_as_matrix()now returns a matrix with character values, not lists- Dimension lists will be imported as a single value - this is still stupid, actually, but I can't currently think of a better way.
- Two Vignettes: Demo.Rmd with essential workflow, about.Rmd explaining a bit more. Demo.Rmd updated to begin working on that.
I stated that I don't intend any structural changes, but that - apparently - was impossible. Old scripts will not work without interference with the new version.
- add version to connect_idaifield()-function since options have changed with the switch to Field Desktop (iDAI.field 3)
- update documentation accordingly
- Updated docs
- Changes geometry reformatting to accommodate 3D-data (if not available, set to 0)
First release.
I will try to limit future changes so that they will not affect the structural outcome of the core functions as they are now.
TODOs:
- Find a way to handle the remaining lists better, esp in combination with the non-list columns.
- Try to assign column type automatically (numeric for numbers etc.)
- How to deal with the dimension...-fields (measurements)? Currently ignored, can be flattened with dplyr and friends.
- Added a
NEWS.mdfile to track changes to the package. - Basic functions work but are not thoroughly tested