WizataDSAPIClient

client wrapper to cloud data science API

accessible preferably through wizata_dsapi.api() using os variables, you can also create an instance using init()
please contact your administrator to receive configuration and connection details

Attributes

Name	Type	Description
domain	str	URL of data science API
client_id	str	azure authentication - client id
client_secret	str	azure authentication - your client secret API key
scope	str	azure authentication - scope
tenant_id	str	azure authentication - tenant
protocol	str	default - https
username	str	azure authentication - deprecated - username
password	str	azure authentication - deprecated - password

Methods

abort()

send an abort request for executions and return a result message

Name	Type	Default	Description
executions	list		must be a list containing uuid or Execution.

access_check()

call the who am I request, to read proper headers and get specific information.

return: roles, twins

add_template_property()

Name	Type	Default
template
property_name	str
property_type	str	datapoint
is_required	bool	True

authenticate()

perform authentication on client side

await_deployment()

poll a deployment until it stops moving.

Name	Type	Default	Description
deployment			Deployment or its id.
timeout	int	900	seconds before giving up waiting.
interval	int	5	seconds between polls.
return: the Deployment in its final state - check `status`, which is 'deployed' on success and 'failed' otherwise.

await_execution()

wait for an execution to reach a terminal status (completed, failed or aborted).

Name	Type	Default	Description
execution	ExecutionLog		ExecutionLog returned by experiment() or run().
timeout	int	180	maximum seconds to wait before raising TimeoutError (default 180).
poll_interval	int	5	seconds between status checks (default 5).
return: ExecutionLog with terminal status.

batch_query()

Name	Type	Default
step	timedelta
datapoints	list	None
start	datetime	None
end	datetime	None
interval	int	None
agg_method	str	mean
template	str	None
twin	str	None
null	str	None
filters	dict	None
options	dict	None
group	dict	None
field		None
bucket	str	None
tags	dict	None
verbose	bool	True

branches()

list the branches relevant to a pipeline.

Name	Type	Default	Description
pipeline			pipeline key, UUID or Pipeline.
return: dict with 'default', 'feature' and 'release' - feature branches are this pipeline's drafts, release branches are the frozen refs deployments run from.

build_image()

build an image of a pipeline, store it on the image repository and return its pipeline image id.

Name	Type	Default	Description
key	str		pipeline key.
branch	str	None	release branch (`release/<name>`) to freeze the image from. REQUIRED when the tenant has git enabled — the pipeline definition and the raw script sources are both read at that commit, and the resolved sha is recorded in the image (DEV-5965). Ignored when git is disabled.
return: image id.

create()

create and save an object on the server

Name	Type	Default	Description
obj			object from any supported entity (see info()) or python callable function (Script)
return: id of created object

create_branch()

create a branch for a pipeline and return the pipeline as it stands on it.

Name	Type	Default	Description
pipeline			pipeline key, UUID or Pipeline.
name	str	None	branch name. Leave None for a draft, which is named for you. A release is named `release/<name>`.
source	str	None	ref to branch from; the default branch when None.
return: the Pipeline read at the new branch, with 'branch' set.

create_component()

create a component based on its ID.

Name	Type	Default	Description
component	SolutionComponent

create_mobile_tile()

Create a mobile asset tile.

Name	Type	Default	Description
tile	MobileAssetTile		MobileAssetTile to create.
return: created MobileAssetTile with server-generated id.

delete()

delete an object on the server

Name	Type	Default	Description
obj			object to delete including all content

delete_branch()

discard a branch.

A draft goes on demand. A release is refused with 409 while a deployment or a

trigger still runs from it - pause or redeploy those first, rather than

leaving a schedule pointing at a ref that no longer exists.

Name	Type	Default	Description
pipeline			pipeline key, UUID or Pipeline.
branch	str		branch ref.

delete_component()

delete component

Name	Type	Default	Description
component_id	UUID

delete_mobile_tile()

Delete a mobile asset tile.

Name	Type	Default	Description
tile_id	str		tile UUID.

delete_model()

delete a model version from the repository.

Name	Type	Default	Description
identifier	str		full identifier including alias (e.g. 'my_model@alias' or 'my_model.twin@alias').

deploy()

deploy a pipeline: merge the draft, cut the release, build the image and create

the schedule - the whole sequence, as one call.

Returns immediately with a Deployment to poll; an image build outlasts the

gateway timeout, so the record carries the progress instead of the request.

Poll it with :meth:get on the returned id, or :meth:await_deployment.

Name	Type	Default	Description
pipeline			pipeline key, UUID or Pipeline.
twins	list	None	hardware ids or twin UUIDs this schedule covers - both forms are accepted - or `{twin, target, deviceId}` entries where target is "cloud" (default) or "edge".
interval	int	None	milliseconds between runs.
delay	int	None	milliseconds to wait after each interval boundary.
packaging	str	None	"image" (default) or "release".
source	str	None	branch or ref to deploy from.
properties	dict	None	pipeline properties for every run.
plot	bool	None	produce plots on each run. Refused on edge.
train	bool	None	train models on each run. Refused on edge, and on an image-packaged deployment - the model would be trained and silently discarded, because the image's frozen model is what predict uses.
write	bool	None	write results back. Defaults on.
return: Deployment to poll.

deploy_preview()

plan a deploy without performing it: what would be merged, cut, built and

scheduled, and every reason it would be refused.

Writes nothing, and always answers 200 - read ok and problems rather

than catching an exception. Takes the same arguments as :meth:deploy.

Name	Type	Default	Description
pipeline			pipeline key, UUID or Pipeline.
kwargs
return: the plan, including 'ok' and 'problems'.

deployed_plot()

fetch one of a deployed pipeline's LIVE plots.

The plot step is identified by BOTH its script and its input dataframe - that

pair is the plot's identity, as :meth:deployed_plots lists it.

Returns the envelope, not the bare figure, because the provenance is the point:

a live plot is overwritten on every run and 'generatedAt' / 'executionId' say

which one wrote it. Render it with client.plot(figure=envelope["figure"]).

Name	Type	Default	Description
pipeline			pipeline key, UUID or Pipeline - both references are accepted.
script	str		the plot step's script.
dataframe	str		the plot step's input dataframe.
twin	str	None	hardware id - required on a templated pipeline.
shape	str	figure	'figure' (default) for the stored Plotly figure, or 'data' for its traces projected to 'series'.
return: dict with {pipelineKey, twin, script, dataframe, name, executionId, release, sha, imageId, generatedAt} plus 'figure', or 'series' and 'notProjected' when shape is 'data'.

deployed_plots()

list a deployed pipeline's LIVE plots - what its deployments are currently

drawing, one per (twin, plot step).

A live plot is overwritten on every triggered run, so there is exactly one

current figure per plot and no history to page through. Fetch one with

per execution, see :meth:plots.

Name	Type	Default	Description
pipeline			pipeline key, UUID or Pipeline - both references are accepted.
twin	str	None	hardware id, to list only that twin's plots.
return: list of dict {pipelineKey, twin, script, dataframe, key, deploymentId, produced, lastModified, size}. 'produced' is False when the definition has this plot step but no run has written it yet.

download_file()

download a model extra file from the repository.

Name	Type	Default	Description
model	ModelInfo	None	model (ModelInfo), alternatively can use identifier
file	ModelFile	None	file (ModelFile), alternatively can use path
identifier	str	None	identifier including alias of the model.
path	str	None	path name of the file including the extension.
return: bytes content.

download_image()

download a pipeline image from the repository

Name	Type	Default	Description
pipeline_image_id	str		id of image
return: packaged (unzipped) as a PipelineImage

download_model()

download a model image directly from the repository, auto-select alias if necessary.

Name	Type	Default	Description
identifier	str		exact identifier including alias of the model or no alias to auto-select active version.
return: ModelInfo with trained_model loaded, extra file must be downloaded separately.

download_script()

Download a script with its artifacts (function/source) as a dill binary.

Name	Type	Default	Description
name	str		script name.
script_type	str	api	one of 'api', 'raw', 'module' (default: 'api').
return: Script with function loaded.

experiment()

experiment and train models with a pipeline.

existing experiment is required (use create or upsert_experiment(key, name)).
if your pipeline is templated please provide a twin.
please provide all variables and parameters required through properties.
return an execution
check status with "wizata_dsapi.api().get(execution).status"
see plots with "wizata_dsapi.api().plots(execution)"

Name	Type	Default	Description
pipeline			pipeline identified by key (str) .
experiment		None	existing experiment identified by key (str).
twin		None	twin identified by hardware ID (str)(optional).
image	str	None	pipeline image id to use.
properties	dict	None	dictionary containing override for variables or additional parameters for your script.
train	bool	True	train machine learning model on model steps.
plot	bool	True	if False plot steps are ignored.
write	bool	False	if False write steps are ignored.
version	str	None	version to use, by default use your python environment version.
branch	str	None	git branch to run against (git mode); default branch when omitted.

get()

get record content from DS API.

get will look first for obj then for id then for key (e.g. if id and key specified, key is ignored )

Name	Type	Default	Description
obj		None	a wizata_dsapi.
id	UUID	None	a UUID of a specific object (use in combination of entity)
key	str	None	a logical key of a specific object (e.g. hardware id for DataPoint and Twin (use in combination of entity)
entity		None	plural name of the entity or class (use in combination of entity)(e.g. scripts, plots, mlmodels, dataframes...)
script_name		None
experiment_key		None
pipeline_key		None
model_key		None
template_key		None
twin_hardware_id		None
datapoint_hardware_id		None
return: object with all properties or None if not found.

get_business_labels()

get a name / uuid dictionary with all business labels in platform.

get_categories()

get a name / uuid dictionary with all categories in platform.

get_components()

get components

Name	Type	Default	Description
label_id	UUID	None	filter on a specific label
twin_id	UUID	None	filter on a specific twin
template_id	UUID	None	filter on a specific template
owner_id	UUID	None	filter on a specific owner_id
organization_only	bool	False	work only with organization components (by default - False)
name	str	None	filter on a specific name (contains)

get_datapoint_mappings()

get datapoint mapping from a registration.

Name	Type	Default	Description
registration

get_mobile_config()

Get resolved config for an asset (twin overrides template).

Name	Type	Default	Description
template_id	str		template UUID.
twin_id	str	None	optional twin UUID.
return: MobileAssetConfig or None.

get_mobile_tiles()

Get merged tiles for an asset (template + twin overrides).

Name	Type	Default	Description
template_id	str		template UUID.
twin_id	str	None	optional twin UUID for twin-scoped overrides.
return: list of MobileAssetTile.

get_pipeline_at()

read a pipeline as it stands on a branch, rather than on the default one.

Name	Type	Description
pipeline		pipeline key, UUID or Pipeline.
branch	str	branch ref.
return: Pipeline.

get_registrations()

retrieve all registrations for

Name	Type	Default	Description
template			template object, UUID or str key.
return: list of twin registration.

get_script_at()

read a RAW script's source as it stands on a branch.

A branch run executes the branch's script, so this is what actually runs

there - the default-branch copy may differ.

Name	Type	Description
name	str	script name.
branch	str	branch ref.
return: Script.

get_template_properties()

Retrieve all properties for a given template.

Name	Type	Default	Description
template			template object, UUID or str key.
return: list of TemplateProperty.

get_ts_query()

Get a Query string to Timeseries Database.

Name	Type	Default	Description
datapoints	list	None	list of datapoints to fetch.
start	datetime	None	start datetime of range to fetch
end	datetime	None	end datetime of range to fetch
interval	int	None	interval in milliseconds.
agg_method	str	mean
template	str	None	template to fetch.
twin	str	None	hardware ID of twin to fetch based on template.
null	str	None	By default at 'drop' and dropping NaN values. If not intended behavior please set it to 'ignore' or 'all'.
filters	dict	None	dict of filters.
options	dict	None	dict of options.
return: dataframe

get_units()

get a name / uuid dictionary with all units in platform.

info()

print insights regarding version and supported operations

lists()

lists all elements of a specific entity.

Name	Type	Default	Description
entity			plural name of the entity or class (e.g. scripts, plots, mlmodels, dataframes...)
return: list of all elements with at least the id property.

load_model()

load a model pickle or torch from the repository ready to be used.

Name	Type	Default	Description
model			ModelInfo to load
return: ModelInfo with the trained model.

merge_branch()

merge a draft into the default branch and delete it.

Name	Type	Description
pipeline		pipeline key, UUID or Pipeline.
branch	str	the feature branch to publish.
return: dict with 'merged_sha', 'deleted_branch' and 'files_changed'.

pause_deployment()

pause a deployment: delete its trigger, keep the record.

Synchronous. The manifest is kept, so :meth:resume_deployment rebuilds the

same schedule rather than asking for it again.

Name	Type	Default	Description
deployment			Deployment or its id.
return: the updated Deployment.

plot()

Fetch and show plot.

Name	Type	Default	Description
plot_id	str	None	Plot Id
plot	Plot	None	Wizata Plot Object
figure		None	JSON Figure
return: plotly figure

plots()

get all plot for an execution.

Name	Type	Default	Description
execution	ExecutionLog		id, Execution or ExecutionLog.
return: list of plots (plotly figures).

query()

Query a dataframe from API.

Name	Type	Default	Description
datapoints	list	None	list of datapoints to fetch.
start	datetime	None	start datetime of range to fetch
end	datetime	None	end datetime of range to fetch
interval	int	None	interval in milliseconds.
agg_method	str	mean
template	str	None	template to fetch.
twin	str	None	hardware ID of twin to fetch based on template.
null	str	None	By default at 'drop' and dropping NaN values. If not intended behavior please set it to 'ignore' or 'all'.
filters	dict	None	dict of filters.
options	dict	None	dict of options.
group		None	can be use to set group system and event retrieving instructions (dict or list[dict]).
field		None	by default 'value' if none, accept str or list (e.g. value, reliability, eventId, ...)
bucket	str	None	specify on which bucket the query applies.
tags	dict	None
scopes	list	None	list of scope filters for SCOPED datapoints (e.g. ["normal", {"vibration": {"gte": 0.6}}]).
return: dataframe

query_json()

query time-series as JSON, described.

Takes the same arguments as :meth:query. Where query hands back a

DataFrame, this returns the rows as JSON — and by default wraps them in a

description of their own shape, so a caller that is not pandas can render

them without guessing:

{"columns": [...], "pivot": ..., "data": [...]}

That is what the front-end's query form is built on: an aggregated query

comes back pivoted with one column per datapoint, a raw one comes back long

with a 'value' column, and only the envelope says which you got (DEV-6031).

A separate method rather than a flag on :meth:query, deliberately — a

parameter that changes the return type between a DataFrame and a dict makes

every caller branch on what it asked for.

Name	Type	Default	Description
include_format	bool	True	False to get the bare rows without the envelope.
kwargs
return: dict with columns/pivot/data, or the rows alone when `include_format` is False.

register_twin()

Name	Type	Default	Description
template			template object, UUID or str key.
twin			twin object, UUID or str key.
properties	dict		dict where key = template property and value = datapoint name or const value (str, int, float, relative or epoch datetime).
override		True	by default at True - allow overriding any existing subscription

remove_template_property()

Name	Type	Default	Description
template
property_name	str

resume_deployment()

resume a paused deployment, re-creating its trigger from the stored manifest.

Name	Type	Default	Description
deployment			Deployment or its id.
return: the updated Deployment.

retry_deployment()

continue a failed or interrupted deployment from where it stopped.

Safe to call repeatedly: a merge or a release cut cannot be undone, so

completed steps are skipped rather than replayed.

Name	Type	Default	Description
deployment			Deployment or its id.
return: the updated Deployment.

run()

run a pipeline.

existing models are used for simulation and prediction.
caution this might affect data inside platform or trigger automation.
if your pipeline is templated please provide a twin hardware ID.
please provide all variables and parameters required through properties.
return an execution
check status with "wizata_dsapi.api().get(execution).status"
check results in platform (dashboard/explorer) or perform queries.

Name	Type	Default	Description
pipeline			pipeline identified by key (str) .
twin		None	twin identified by hardware ID (str)(optional).
properties	dict	None	dictionary containing override for variables or additional parameters for your script.
image	str	None	pipeline image id to use.
train	bool	False	train machine learning model on model steps.
plot	bool	False	if False plot steps are ignored.
write	bool	True	if False write steps are ignored.
version	str	None	version of python environment to use if different than local environment.
branch	str	None	git branch to run against (git mode); default branch when omitted.

save_script_at()

save a RAW script on a branch, leaving the default branch untouched until the

branch is merged.

Name	Type	Default	Description
script	Script		Script with its source set.
branch	str		branch ref.

search_datapoints()

get datapoints with a paged query.

Name	Type	Default	Description
page	int	1	numero of the page - default 1.
size	int	20	quantity per page - default 20 max 100.
sort	str	id	column to sort results - default id.
direction	str	asc	sorting direction by default asc, accept also desc.
hardware_id	str	None	filter on a specific hardware ID name or partial name.
categories	list	None	list of UUID or Category.
business_types	list	None	list of BusinessType or str.
twin		None	uuid or Twin element to search datapoints.
recursive	bool	False	set to True in combination of a twin to look inside all sub-twins recursively.
return: PagedQueryResults, check total for number of potential results and results for the list of entity.

search_execution()

search executions using filters, sorting and pagination via POST.

Name	Type	Default	Description
search_query		None	dict or SearchQuery with filters, sort, page, size, groups, groupSummaries, serviceName.
return: PagedQueryResult with ExecutionLog results and optional group statistics.

search_executions()

get executions with a paged query.

Name	Type	Default	Description
page	int	1	numero of the page - default 1.
size	int	20	quantity per page - default 20 max 100.
sort	str	id	column to sort results - default id.
direction	str	asc	sorting direction by default asc, accept also desc.
pipeline_id	UUID	None	filter on a specific pipeline.
twin_id	UUID	None	filter on a specific twin.
template_id	UUID	None	filter on a specific template.
status	ExecutionStatus	None	filter on a specific status.
return: PagedQueryResults, check total for number of potential results and results for the list of entity.

search_insights()

get insights with a paged query.

Name	Type	Default	Description
page	int	1	numero of the page - default 1.
size	int	20	quantity per page - default 20 max 100.
sort	str	id	column to sort results - default id.
direction	str	asc	sorting direction by default asc, accept also desc.
datapoint_id	str	None	id of the datapoint on which filtering the insights.
twin_id	str	None	id of the twin on which filtering the insights.
component_id	list	None	id of the component on which filtering the insights.
return: PagedQueryResults, check total for number of potential results and results for the list of entity.

search_models()

get all information related to models stored on Wizata.

return: ModelList structure model list

search_scripts()

Search scripts across all types using a paged query.

Name	Type	Default	Description
search_query		None	SearchQuery instance or dict with filters, sort, page, size.
return: PagedQueryResult with Script results.

search_twins()

Search twins using filters, sorting and pagination via POST.

Supported filter fields:

hardwareId, name, description, parentId, twinTypeId, type, isAsset, _search.

Name	Type	Default	Description
search_query		None	SearchQuery instance or dict with filters, sort, page, size.
return: PagedQueryResult with Twin results.

send_alerts()

send alerts through backend

Name	Type	Default
message	str
recipients	list	None
alert_type	AlertType	AlertType.SMS
subject	str
cc	list	None

unregister_twin()

un-register a twin from a specific template.

Name	Type	Default	Description
template			template object, UUID or str key.
twin			twin object, UUID or str key.

update()

update and save an object on DS API

Name	Type	Default	Description
obj			object to update on DS API (see info()) or python callable function (Script)

update_component()

update a component based on its ID.

Name	Type	Default	Description
component	SolutionComponent

update_mobile_tile()

Update an existing mobile asset tile.

Name	Type	Default	Description
tile	MobileAssetTile		MobileAssetTile with id set.
return: updated MobileAssetTile.

upload_file()

upload a model extra file to the repository.

please use upload_model for a trained_model

Name	Type	Default	Description
content			the bytes content of the file.
model	ModelInfo	None	model (ModelInfo), alternatively can use identifier
file	ModelFile	None	file (ModelFile), alternatively can use path
identifier	str	None	identifier including alias of the model.
path	str	None	path name of the file including the extension.
return: bytes content.

upload_model()

upload a model within the model repository.

by default use model_info.trained_model and convert it to a pickle
for already torch or pickle please pass the bytes_content
model_info.file_format must be set properly to 'pkl' or 'pt'

Name	Type	Default	Description
model_info	ModelInfo		model info, with at least key (+twin, +property, +alias) and trained_model.
bytes_content		None	bytes[] of your torch or pickle model.

upsert()

upsert an object on the server using its logical unique key (e.g. hardware_id, key, name).

if the object already exists (matched by key), it is updated. otherwise, it is created.

Name	Type	Default	Description
obj			object to upsert on the server (must implement api_key()) or python callable function (Script)
return: ID of the object created or updated

upsert_datapoint()

Upsert a datapoint (ignore ID, use the hardware_id).

Name	Type	Default	Description
datapoint	DataPoint
return: upserted datapoint.

upsert_experiment()

Upsert an experiment.

Name	Type	Default	Description
key	str		unique key identifying the experiment.
name	str		display name of the experiment.
pipeline		None	pipeline to set at creation only - cannot be updated.
twin		None	twin uuid, hardware id or twin.
return: upserted experiment.

upsert_mobile_config()

Upsert mobile asset config by (templateId, twinId).

Name	Type	Default	Description
config	MobileAssetConfig		MobileAssetConfig to upsert.
return: upserted MobileAssetConfig.

upsert_pipeline()

Upsert a pipeline (ignore ID, use the key).

Name	Type	Default	Description
pipeline	Pipeline
return: upserted pipeline.

upsert_template()

Upsert a template.

Name	Type	Description
key	str	unique key identifying the template.
name	str	display name of the template
return: upserted template.

upsert_twin()

Upsert a twin (ignore ID, use the hardware_id).

Name	Type	Default	Description
twin	Twin
return: upserted twin.

who_am_i()

decrypt current JWT token used to get information about current user.

return: