Python Client History
3.x Client Version
3.4
3.4.0
Removed
The
get_fairness
method is removed. Please use charts and custom metrics to track / compute fairness metrics on your model.
3.3
3.3.2
Modifications
Fixed the error while setting notification config for alert rule.
3.3.1
Modifications
Added validation while adding notifications to alert rules.
Upgraded dependencies to resolve known vulnerabilities - deepdiff, mypy, pytest, pytest-mock, python-decouple, types-requests and types-simplejson.
3.3.0
New Features
Introduced
upload_feature_impact()
method to upload or update feature impact manually.
3.2
3.2.0
New Features
Introduced evaluation delay in Alerts Rule.
Optional
evaluation_delay
parameter added toAlertRule.__init__
method.It is used to introduce a delay in the evaluation of the alert.
Modifications
Fix windows file permission error bug with publish method.
3.1
3.1.2
Modifications
Adds support to get schema of Column object by
fdl.Column
3.1.1
Modifications
Updated
pydantic
andtyping-extensions
dependencies to support Python 3.12.
3.1.0
New Features
Introduced the native support for model versions.
Optional
version
parameter added toModel
,Model.from_data
,Model.from_name
methods.New
duplicate()
method to seamlessly create new version from existing model.Optional
name
parameter added toModel.list
to offer the ability to list all the versions of a model.
3.0
3.0.5
New Features
Allowed usage of
group_by()
to form the grouped data for ranking models.
3.0.4
Modifications
Return Job in ModelDeployment update.
3.0.3
New Features
Added
Webhook.from_name()
Modifications
Import path fix for packtools.
3.0.2
Modifications
Fix pydantic issue with typing-extensions versions > 4.5.0
3.0.1
New Features
General
Moving all functions of client to an Object oriented approach
Methods return resource object or a deserialised object wherever possible.
Support to search model, project, dataset, baselines by their names using
from_name()
method.List methods will return iterator which handles pagination internally.
Data
Concept of environments was introduced.
Ability to download slice data to a parquet file.
Publish dataframe as stream instead of batch.
New methods for baselines.
Multiple datasets can be added to a single model. Ability to choose which dataset to use for computing feature impact / importance, surrogate generation etc.
Model can be added without dataset.
Ability to generate schema for a model.
Model delete is async and returns job details.
Added cached properties for
model
:datasets
,model_deployment
.
Alerts
New methods for alerts:
enable_notification
,disable_notification
,set_notification_config
andget_notification_config
.
Explainability
New methods in explainability:
precompute_feature_impact
,precompute_feature_importance
,get_precomputed_feature_importance
,get_precomputed_feature_impact
,precompute_predictions
.Decoupled model artifact / surrogate upload and feature impact / importance pre-computation.
Modifications
All IDs will be UUIDs instead of strings
Dataset delete is not allowed anymore
\
2.x Client Version
2.5
2.5.3
Modifications
Import path fix for packtools.
2.5.2
Modifications
Fix pydantic issue with typing-extensions versions > 4.5.0
2.5.1
New Features
Support tf_idf vectors in
TextEmbedding
type column definition
2.5.0
New Features
Add support for enrichments
Allow pausing multiple alerts at once
Add percentage violation metrics for alert
Support for alert revisions
2.4
2.4.1
Modifications
Fix slice query with
vector
type columns
2.4.0
New Features
Add support for segments
Modifications
Ensure
TextEmbedding
andImageEmbedding
is supported for vectors in metadata
2.3
2.3.0
New Features
Added support for creating alerts on the
Frequency
metric.
Modifications
Relax pydantic version to allow any version between 1.9 and 2
2.2
2.2.1
Modifications
Relax pydantic version to allow any version between 1.9 and 2
2.2.0
Modifications
Method
add_custom_metric
supports an optionaldescription
parameterparameter
fql
is renamed todefinition
inadd_custom_metric
methodA new
get_custom_metric
function to get details about a single custom metric.
2.1
2.1.2
Modifications
Relax pydantic version to allow any version between 1.9 and 2
2.1.1
Modifications
Update
pyarrow
requirement to7.0.0
.
2.1.0
New Features
Introduce Model Tasks
NOT_SET
andLLM
Relax Target / Output specification for model tasks
NOT_SET
andLLM
Support custom metrics
Modifications
DatasetDataSource
andEventIdDataSource
will takedataset_id
instead ofdataset_name
Method
list_baselines
to return baseline names instead of baseline objects
2.0
2.0.8
Modifications
Relax pydantic version to allow any version between 1.9 and 2
2.0.7
Modifications
Support string metric types for alert creation, for server 23.7.
2.0.6
Modifications
Update
pyarrow
requirement to7.0.0
.
2.0.5
Modifications
Fix for a minor bug in
fdl.DatasetInfo.from_dataframe
2.0.4
Modifications
Update
pyarrow
requirement to13.0.0
.
2.0.3
Modifications
Relax pandas version for 2.0.
In
get_slice
method, parameterquery
is reverted back tosql_query
2.0.2
Modifications
Fix Parquet conversion issue in
upload_dataset
andpublish_event_batch
2.0.1
Removed
Following methods are removed
register_model
upload_model_package
update_model
trigger_pre_computation
_trigger_model_predictions
generate_sample_events
list_teams
list_project_roles
list_org_roles
unshare_project
share_project
process_avro
process_csv
New Features
Add
monitor_components
as an attribute forCustomFeature
of typeFROM_COLUMNS
. Default asFalse
Adds new statistic type
SUM
to supported alert metricsSupport
CustomFeature
of typeFROM_VECTOR
FROM_TEXT_EMBEDDING
andFROM_IMAGE_EMBEDDING
Modificatiosn
Remove
column
as a parameter inadd_alert_rule
andget_alert_rules
functionsDefault
FileType
Parquet inupload_dataset
andpublish_event_batch
In
get_slice
method,slice_query
parameter is changed toquery
Versions prior to 2.0 are deprecated
1.x Client Version
1.8
1.8.6
Modifications
Relax pandas version for 2.0.
1.8.4
New Features
New DeploymentType enum for
MANUAL
deployment
1.8.3
Modifications
New
columns
parameter in add_alert_rule and get_alert_rules to support multiple columns to be used for server version >= 23.3.0get_triggered_alerts supports
alert_value
as a float as well as a dict
1.8.2
Modifications
Fixed a bug where
min
andmax
for columns of typefloat
indataset_info
are cast intoint
after uploaded
1.8.1
Modifications
Fixed a bug wherein null string was going in request body if body wasn't specified.
Fix
categorical_target_class_details
when passed as an arrayFix a bug where
fdl.ModelInputType.TEXT
were not being accepted properlyFix
categorical_target_class_details
when passed as an empty list
1.8.0
Modifications
Add new alert type -
statistic
for setting alertsAdd
target_class_order
as a required field ofModelInfo
object whenmodel_task
isMULTICLASS_CLASSIFICATION
,RANKING
orBINARY_CLASSIFICATION
. Only applies forBINARY_CLASSIFICATION
when target column is of typeCATEGORY
Add
columns
as a parameter inadd_alert_rule
andget_alert_rules
functionsAdd deprecation warning for
column
as a parameter inadd_alert_rule
andget_alert_rules
functions
1.7
1.7.4
Modification
Do not typecast column with strings in
get_slice
1.7.3
Modification
Send row and column count information to dataset upload api
1.7.2
Modification
Bring back
WeightingParams
object
1.7.1
Modification
Relaxed boto3 version constraint
1.7.0
Removed
Remove support for initializing fiddler client with version=1
Following methods are removed
get_segment_info
delete_segment
deactivate_segment
activate_segment
list_segments
upload_segment
add_monitoring_config
publish_parquet_s3
publish_events_log
Last updated