Python Client Changelog¶
Released on 2020-12-10.
This version now requires pandas==0.25.3. It allows to write/load parquet files.
get_dataset_as_dataframeto efficiently download the data from a datasource into a Pandas dataframe. The data is exchanged between the Serenytics backend by either a pandas pickle file, a parquet file (only for Redshift storage) or a CSV file. This function replaces the function
get_datasetwhich was using only CSV formats. And in some cases these CSV format where not practical.
function Data.get_as_dataframe will be marked deprecated in a future version (not removed yet because it is useful to get the result of a query with options into a dataframe, while
get_dataset_as_dataframedoes not accept query options so far).
reload_data_from_dataframehas been rewritten. It was converting the pandas dataframe to a json format and then was calling
reload_datawhich was not optimized at all. In the new version, it uploads the dataframe as a pickle file and the backend loads it into the storage.
Released on 2018-01-08.
- new method available on data sources:
Released on 2017-06-16.
source.batch()call is synchronous by default to get errors if data operations failed.
Released on 2017-05-22.
webapp.warm_up_cache()takes a new optional parameter
payloadfor template dashboards.
Released on 2017-03-16.
source.update_data_from_file()takes a new optional parameter