Changelog¶
0.8.4 - Minor project changes¶
- Fixed the build issue with
xunitparser
by usinggenbadge
. Fixed #28. - Changed the reference test dataset for the one advised by ODS. Fixed #23
0.8.3 - Minor fixes concerning import errors¶
-
Improved
odskeys
error message whenclick
is not installed. Now you canpip install odsclient[full]
to install all optional dependencies at once. Fixes #26 -
No
ImportError
is raised by default ifkeyring
is not installed. Instead a warning message is issued. Fixes #25
0.8.2 - Auto-closing requests.Session
¶
-
The underlying
requests.Session
is now automatically closed byODSClient
objects when they are garbaged out, to avoid leaving hanging Sockets around. This is not the case if a customSession
is passed. An advanced boolean parameterauto_close_session
can be used to turn this feature off. Fixes #27 -
New CI/CD: github actions !
0.8.1 - file_cache
on get_whole_dataframe
¶
file_cache
is now available on package-levelget_whole_dataframe
. Fixes #24
0.8.0 - New cache feature¶
- New cache functionality: a
file_cache
argument is available on most methods. The cache is by default located under.odsclient/
and containsutf-8
-encoded versions of the dataset files. New util methodsclean_cache
andget_cached_dataset_entry
. Fixes #20
0.7.0 - New streaming-related features¶
- You can now display a progress bar using
tqdm=True
. Note that this requires thetqdm
package to be installed, and since some ODS platforms do not return theContent-Length
HTTP header, only the size and download rate might be displayed. Fixed #9 - You can now stream a dataset directly to a file using
to_path
. Fixed #9 apikey_filepath
can now be apathlib.Path
.
0.6.0 - New feature: push dataset¶
- New
push_dataset_realtime
client method and shortcut, to push a CSV string or apandas
dataframe to an ODS server through the Realtime API. PR #21 by@zoltanctoth
, thanks !
0.5.1 - better packaging¶
- packaging improvements: set the "universal wheel" flag to 1, and cleaned up the
setup.py
. In particular removed dependency tosix
for setup and addedpy.typed
file. Removed tests folder from package. Fixes #19
0.5.0 - odskeys improvements¶
- New
odskeys show
command to open the OS GUI for keyring management. Fixed #18 odskeys get/set/remove
now always print the complete url used for the keyring entry. Fixes #17- Improved documentation about api keys management. Fixed #15
0.4.0 - odskeys commandline¶
- New
odskeys
commandline utility to ease registration of api keys inkeyring
. Fixes #16. Added a dedicated documentation page. - Documentation improvements, in particular concerning api key registration.
- Filled the API reference documentation section
0.3.0 - api keys¶
API Keys:
- New documentation section on api keys. Fixes #11
- New methods
get_apikey()
and<ODSClient>.get_apikey()
for debugging purposes - API keys can now be provided through OS Environment Variables. Fixes #6. New method
<ODSClient>.get_apikey_from_envvar()
for debugging. - API keys can now be provided through
keyring
. Fixes #8 - API key files:
apikeyfile_path
argument renamedapikey_filepath
. Now tolerating new lines or blanks at the end of api key files. Fixes #12.apikey_filepath
now has a default value'ods.apikey'
and file check is now optional. Fixes #13 - Now correctly raising an error (
InsufficientRightsForODSResourceError
) when rights are not sufficient for a resource. Fixed #7. enforce_apikey
controls if an api key is mandatory before calling (whatever way it is provided: explicit, file, env variable, or keyring). ANoODSAPIKeyFoundError
is raised in case none is found.
Misc:
- Added documentation on target platform customization. Fixes #14
- Now tolerating trailing slashs at the end of custom
base_url
(they are removed automatically). - Init: now exposing
ODSClient
andODSException
at the root package level. - new method
get_apikey()
onODSClient
0.2.0 - datasets as pandas dataframes + misc.¶
New method get_whole_dataframe
to directly retrieve a dataset as a pandas dataframe. It works in streaming mode so as to correctly handle large datasets. Fixes #1
Added csv_separator
option to get_whole_dataset
methods. Set default value of use_labels_for_header
to True
to mimic what is available in the ODS website.
Now parsing the ODS errors correctly to raise ODSException
. Fixes #2
0.1.0 - First public version¶
Extracted from internal sources. Ability to download a whole dataset from any of the ODS platform using the "download flat dataset" API.