#11 and #3 (and also #8): provided an API to the controlled vocabulary (as long as no ontology service is in place), also wrote first pytest, but not yet working (settings have to be established in test database), prepared scripts to create pages also documenting the controlled vocabulary (pages: index can still show the source code, which is wanted)

#11: still struggling with creating pdf (but now a little bit better) -- see (pandoc) command in issue

#11: started adding descriptive text for automatic openapi documentation (especially for use documenting metadata)

#3: found solution for loading controlled vocabulary only once at start time (with raw database connection); have to put it into place (so that pytests are also running); for now the workaround is still active

#4: nested download for station inside timeseries still not working properly, but at least shows some information; upload after gitlab crash

#4: nested download was extended for timeseries; TBD: showing nested station information is not working yet; also changed programme from many-to-many to 1:n

#3: added unique constraint to controlled vocabularies; added foreign keys (controlled vocabularies) to models

#3,#4: nested uploads/downloads using controlled vocabulary; still to investigate: how to load enum_tables at beginning (this commit is a workaround)

#3: now nested human readable download is established for timeseries; TBD: do this for all other modesl

#3: toar_controlled_vocabulary is used for downloading timeseries, pytests are added for timeseries; not yet working: how to show human readable fields inside nested models (how is this 'conversion' done automatically?)

almost done: using human readable controlled vocabulary for upload; TBD: using human readable controlled vocabulary for download, adapting pytests

#7: default values for stationmeta_global were defined; #4: nested upload tested for 'stationmeta_global' (not defining all metadata -- getting the rest from default values)

#4: nested upload for stations done; TBD: more testing (esp. pytest), use controlled vocabulary (related to #3)

#4: nested download for stationmeta realized; TBD: uploading of new test database, nested download for stationmeta

#8: common base and test base is in place (but teardown of test_base has to be worked on for chained pytests)

#4: defined association tables also in models; now using one base for the whole app 'toardb'; this has now also to be adapted to test_base

#4: nested upload/download for roles (timeseries) working in production and pytest (uploading codes not human readable version)

#4: a typo made my days and nights long; after git crash do another upload; now detangling models_core and models_role again (different source code files)

#4: code completely messed up, but nested upload working; nevertheless download is NOT returning the nested view

#2: many-to-many relationship for role codes seem to work now within FastAPI; will probably again not work with pytest; has to be cleaned up again: just storing before crash

#1: added another test (timeseries not existing); tried to have more user-friendly error messages; still struggling why I get doubled quotation marks now

#1: pytest for uploading data with no_autoflush now working; added timeshift_from_utc to data insertion; added error message if trying to insert duplicates

#1: pytest for data established; renamed date_time to datetime; needed 'no_autoflush' in timeseries queries

adapted sphinx documentation to new structure (ONE app); some pretty printing (fastAPI does not provide sphinx extensions like django)

created ONE toardb app, also adapted pytests for toardb app (still very last test will throw an error for teardown; inserted example for url percent encoding