- fix
query = ...
innormGeometry()
to allow full customization. <-> has been reverted as it makes sense to select layers automatically. -> improving documentation. - change
.matchOntology()
so that it's possible to match "gappy" hierarchies.
- fix example data so that github actions run properly.
- fix documentation to markdown standard.
- refine function
adb_querry()
.
- smaller fixes
- various fixes to the internal processing of matching tabular to geometric data.
- various fixes to finding matches between harmonized and external geometries.
- new functions
adb_archive()
,adb_backup()
,adb_inventory()
,adb_metadata()
,adb_ontology()
,adb_reset()
,adb_restore()
,adb_schemas()
andadb_translations()
to access the respective part of an areal database. - renamed functions
makeExmapleDB()
toadb_example()
andstart_arealDB()
toadb_init()
- some fixes to the backend as well, where now *.csv tables are not used any more, but information are stored in *.rds files, as they are not meant to be interacted with. This is now done via the
adb_*()
functions, which ensure that nothing is modified to a degree that it breaks - incl. a new feature, where a table that has a missing target column is fixed by matching the child items and then selecting the parent from that as the missing target column
- move the function
edit_matches()
, that was so far located in theontologics
package to here, as it is primarily used inarealDB
and not really a usecase forontologics
- re-enable matching with the ontology when normalising a table, by default
- introduction of the function
updateOntology()
. This updates the gazetteer after a full spatial match has been carried out, drastically taking complexity out of the functions and guaranteeing proper match between the spatial files and the gazetteer
- entirely remove matching via gazetteer for any level below the first level that serves for identifying unique territories (which resulted in a vast simplification of the
normGeometry()
function). - include a
version =
for data series. - include
query =
argument also for tables.
- this include various smaller fixes that should enable a more user-friendly and functional way of specifying geometries and running the normalisation process.
- adapt storing of stage1 files so that they are now aggregated into folders with their data series as name.
- fixing an example in
start_arealBD()
so that a gazetteer is not saved in the home-directory of linux users.
- revise
nromTable()
so that also observation in the tables are kept, if they are not successfully matched with a territory.
- fix intersection sub-algorithm to make it way faster (depending on the amount of non-intersecting features). This was problematic because st_intersection is currently still very inefficient
- fix a bug where all geometries are treated as if they overlap across the boundaries of several parents, even though only a small set does overlap. This lead to a false reconstruction of IDs.
- various fixes to enable full support for automatically matching territories with a gazetteer and concepts with an ontology
- complete overhaul of
matchOntology()
making it vastly more efficient and easier to use/debug - adapt
normGeometry()
andnormTable()
to the changes inmatchOntology()
- start making use of the spatial/topological information derived from normalizing new geometries (that were not part of the initial geometry data set) by updating the gazetteer accordingly. Only if a concept does not overlap with more than (100 - thresh), is it added as a new harmonized concept. In that case, the new geometry is one that is distinct from the harmonized geometry.
- spatial matches are now found automatically and inserted into the ontology.
- matching can be sped up by simplifying geometries.
- spatial hierarchy information of the different administrative levels are now streamlined and provided differently/more explicitly.
- various bug-fixes to handling terms (they are now more explicitly cleaned from white-space and dots in names are removed, e.g. 'St. Gallen' turns to 'St Gallen', which is important for the output column 'gazName', where the administrative levels are concatenated via a '.').
- updates on tests, to be more explicit about the explicit outcome of geometries and the ontology.
- revise the whole geometry normalization process to incorporate the new ontology system.
- fix how to derive the top-most class.
- fix internals to make the code run faster.
- rename
match_ontology()
tomatchOntology()
and revise it so that it can handle several columns sequentially - make use of getters of the ontologics package so that (large) gazetteers are not loaded but only the required info are pulled from them
- remove functions
get_variable()
andselect_path()
, the former is integrated intomatchOntology()
and the latter is not relevant anymore.
- when using
regTable
withupdate = FALSE
, the schema now has a more explicit diagnostic message.
- rename
match_gazetteer
tomatch_ontology
because it can also handle ontologies of other variables - various fixes to make matching as convenient as possible.
- instead of using id and translation tables, concepts are from now on handled via the ontologics package.
- functions
setVariables
andtranslateTables
are discontinued, instead the functionmatch_gazetteer
handles matching of concepts. - the package has been streamlined so that functions
regGeometry
andregTable
don't require anation =
argument anymore, but this can be any other type of "main" polygon. - new function
get_variable
to extract concepts from stage2 tables.
- bugfix of access-rights management
- adapt to changes in tabshiftr
- all
countries
now stored with capital letters and all geoms are stored as MULTIPOLYGON - store tables and geometries by default in rds format with csv/gpkg as option
- several minor bug-fixes
- streamline translations further by creating the file "target_terms.csv" that contains the translations that are allowed (taken from the respective translation tables or geometries).
- more strict translations by allowing only provided terms that may come from outside of a translation table (such as geometry names that were not known at the time of creating the translation table)
- streamline translations to reduce the amount of items that have to be translated (specifically, only translate nations in normTable() that have been subsetted, not the whole list of nations)
- repeat translation of terms as long as some are still (accidentally) missing.
- improve updating of tables (concerning the functions reg*(), setVariables(), translateTerms()).
- streamline overwriting of already registered items in the inventory tables.
- include 'makeExampleDB', which allows to build an example database at any stage. This is useful for testing, debugging and examplifying an areal database.
- improved documentation due to this change.
- refine package documentation
- refine workflow to adapt some minor tweaks.
- Added a
NEWS.md
file to track changes to the package. - initial submission to CRAN