Skip to content
This repository has been archived by the owner on Apr 15, 2022. It is now read-only.

Releases: splicemachine/pysplice

Rev1 DB Model Deployment

03 Feb 14:20
dfb446c
Compare
Choose a tag to compare
  • Initial code for in DB Model deployment
  • Log_model_params logs only params (was metrics)
  • Support for FloatType and DecimalType in context types
  • More generalized way of determining the model stage of a Pipeline
  • Default schema added for artifact storage
  • insert_artifact() extended for mleap storage support
  • Default schema for create_table uses current schema instead of SPLICE

1.0.6-k8

16 Jan 17:05
eb62999
Compare
Choose a tag to compare

Deploy AzureML code fixes

Bug fix in mlmanager.start_run

07 Nov 15:36
8885093
Compare
Choose a tag to compare
1.0.5-k8

Update management.py (#29)

K8s Release 4

24 Oct 14:54
Compare
Choose a tag to compare

Added support for e-e MLManager in testing mode

K8s Release 3

18 Oct 20:38
3b535fd
Compare
Choose a tag to compare

Updated authentication using retrying and added support for in code model deployment

K8s Release 2

03 Oct 20:11
68e63b2
Compare
Choose a tag to compare

Bug fix in MLManager

First release of PySplice on K8s

25 Sep 21:40
ea6779e
Compare
Choose a tag to compare
Dbaas 2571 (#23)

* DBAAS-2387: Added createTable function to create a new schema.table from a dataframe

* DBAAS-2387: Added an optional upper param to insert, upsert and update in case of mixed case columns in SQL tables

* Added a replaceDataframeSchema function to properly set the case of each column name of the dataframe

* DBAAS-2387: fixed schema_table_name in params

* DBAAS-2387: Added 26 Unit Tests; Refactored Create Table Function; Refactored Tests

* removed pyc files and __pycache__ via gitignore

* removed pytest cache

* DBAAS-2387: removed changes from .gitignore

* DBAAS-2571: Updating MLManager to have new features released in MLFlow 1.0

* DBAAS-2571: More features for MLFlow 1.0 Upgrade; deprecated SpliceMLContext; Fixed specificity in Binary Classification Evaluator

* DBAAS-2571: Added Deploymenet to Sagemaker programmatically

* DBAAS-2571: Fixed minor bugs and renamed functions

* DBAAS-2571: Fixed typo in if-statement checking validity of MLFlow REST tracking endpoint URL

* DBAAS-2571: Added Governance for new runs as well as support for overriding username

* DBAAS-2571: Fixed get user function to be less verbose

* DBAAS-2571: Added more checks

* DBAAS-2571: Removed useless dependencies

* DBAAS-2571: Removed mutable initial argument

* DBAAS-2571: Fixed user governance

* DBAAS-2571: Added database store support via py4j for models & artifacts

* DBAAS-2571: Fixed typo in reset run

* DBAAS-2571: Updated JVM

* DBAAS-2571: Added Basic Auth to Job Initiation and added Azure Deployment

* DBAAS-2571: Added support for OneVsRest for Model Hyperparameter extraction

* Missing comma in dependencies

* Update setup.py

* Update setup.py

* Update context.py

Sandline DCOS release

20 Sep 16:12
4711379
Compare
Choose a tag to compare

This is a hard line for DCOS work. This will be functional for DCOS clusters and you can pull from this on new cluster creation. Moving forward, all work will be for kubernetes with the assumption of DB:2.8.0.1920+

First Production Release

25 Apr 00:20
7a3be05
Compare
Choose a tag to compare
Pre-release
Merge pull request #11 from splicemachine/DBAAS-2476/main

DBAAS-2476: missing import, print exceptions

MLManager Beta Updates

16 Apr 22:01
cb5657a
Compare
Choose a tag to compare
Pre-release

Fixed SpliceMLContext code, add safety check around cloud service usage.