teradatamlwidgets 20.0.0.5

Teradataml Widgets

Teradata Widgets

teradatamlwidgets makes available to Python users a user interface to a collection of analytic functions and plot functions that reside on Teradata Vantage. This package provides Data Scientists and Teradata users a simple UI experience within a Jupyter Notebook to perform analytics and visualization on Teradata Vantage with no SQL coding and limited python coding required.

For documentation and tutorial notebooks please visit Documentation.

For Teradata customer support, please visit Teradata Support.

Copyright 2024, Teradata. All Rights Reserved.

Table of Contents

• Release Notes
• Installation and Requirements
• Using the Teradata Widgets Package
• Documentation
• License

Release Notes:

teradatamlwidgets 20.0.0.5

• New Features/Functionality

  • Exploratory Data Analysis UI
  • BYOM Scoring
  • Login UI
  • AutoML
  • Script Table Operator
  • Describe
  • Persist

• New APIs:

  • teradatamlwidgets.eda.Ui()
  • teradatamlwidgets.byom_functions.Ui()
  • teradatamlwidgets.login.Ui()
  • teradatamlwidgets.auto_ml.Ui()
  • teradatamlwidgets.script.Ui()
  • teradatamlwidgets.describe.Ui()
  • teradatamlwidgets.persist.Ui()
  • Added support for notebook utilities.
    • tdnb.text() to create text widget.
    • tdnb.dropdown() to create dropdown widget.
    • tdnb.multi_select() to create multiselect widget.
    • tdnb.get() to get the value of widget.
    • tdnb.get_all() to get all the values of widgets in current session.
    • tdnb.remove() to remove the widget.
    • tdnb.remove_all() to remove all the widgets in current session.
    • tdnb.run_notebook() to run a notebook from another notebook.
    • tdnb.exit() to exit the notebook with a message.

• Bug Fixes

  • Fully qualified table names are now correctly recognized.

teradatamlwidgets 20.0.0.4

• New Features/Functionality

  • None

• New APIs:

  • None

• Bug Fixes

  • Fixed list of SQLE functions

teradatamlwidgets 20.0.0.3

• New Features/Functionality

  • None

• New APIs:

  • None

• Bug Fixes

  • Using native dialog boxes
  • Parameter name change for Plot (color).

teradatamlwidgets 20.0.0.2

• New Features/Functionality

  • Updated documentation

• New APIs:

  • None

• Bug Fixes

  • Initialized default database

teradatamlwidgets 20.0.0.1

• New Features/Functionality

  • Updated documentation

• New APIs:

  • None

• Bug Fixes

  • None

teradatamlwidgets 20.0.0.0

• New Features/Functionality

• New APIs:

  • Analytic functions
    • teradatamlwidgets.analytic_functions.Ui()
    • teradatamlwidgets.analytic_functions.get_output_dataframe()
  • Plotting
    • teradatamlwidgets.plot.ShowPlots()

• Bug Fixes

  • None

Installation and Requirements

Package Requirements:

• Python 3.5 or later

Note: 32-bit Python is not supported.

Minimum System Requirements:

• Windows 7 (64Bit) or later
• macOS 10.9 (64Bit) or later
• Red Hat 7 or later versions
• Ubuntu 16.04 or later versions
• CentOS 7 or later versions
• SLES 12 or later versions
• Teradata Vantage Advanced SQL Engine:
  • Advanced SQL Engine 16.20 Feature Update 1 or later
• For a Teradata Vantage system with the ML Engine:
  • Teradata Machine Learning Engine 08.00.03.01 or later

Installation

Use pip to install the Teradata Widgets Package for Advanced Analytics.

Platform	Command
macOS/Linux	pip install teradatamlwidgets
Windows	py -3 -m pip install teradatamlwidgets

When upgrading to a new version of the Teradata Widgets Package, you may need to use pip install's --no-cache-dir option to force the download of the new version.

Platform	Command
macOS/Linux	pip install --no-cache-dir -U teradatamlwidgets
Windows	py -3 -m pip install --no-cache-dir -U teradatamlwidgets

Using the Teradata Python Package

Your Python script must import the teradatamlwidgets package in order to use the Teradata Widgets Package:

from teradatamlwidgets import login
ui = login.Ui()

from teradataml import *
from teradatamlwidgets import analytic_functions
Load the example data.
load_example_data("movavg", ["ibm_stock"])
load_example_data("teradataml", "titanic")
inputs = ["ibm_stock"]
outputs = ["Project_OutMovingAverageTest"]
ui = analytic_functions.Ui(
		function='MovingAverage',
		outputs=outputs, 
		inputs=inputs)

from teradataml import *
from teradatamlwidgets import plot
# Load the example data.
load_example_data("movavg", "ibm_stock")
load_example_data("teradataml", "iris_input")
# Plot
plot1 = plot.Ui(
		table_name="ibm_stock", 
		current_plot="Line", 
		x='period', 
		series='stockprice', 
		style='green')
plot2 = plot.Ui(
		table_name="iris_input", 
		current_plot="Scatter", 
		x='sepal_length', 
		series='petal_length', 
		xlabel='sepal_length',
		ylabel='petal_length',
		grid_color='black',
		grid_linewidth=1, 
		grid_linestyle="-",
		style='red', 
		title='Scatter Plot of sepal_length vs petal_length',
		heading= 'Scatter Plot Example')
# Combine Plots
plot.ShowPlots([plot1, plot2], nrows=1, ncols=2) 

from teradatamlwidgets import byom_functions 
# BYOM Scoring Functions
byom = byom_functions.Ui(
		function = "DataRobotPredict", 
		byom_location = "mldb", 
		input_table="iris_test", 
		model_id="dr_iris_rf", 
		model_table="byom_models")

from teradatamlwidgets import auto_ml
ui = auto_ml.Ui(
		task="Classification", 
		training_table=iris_train, 
		testing_table=iris_test,
		predict_table='iris_test', 
		algorithms=['xgboost', 'knn'],
		verbose=0,
		max_runtime_secs=300,
		max_models=5)

from teradatasqlalchemy import (CHAR, VARCHAR, CLOB, INTEGER, FLOAT)
from teradatamlwidgets import script 
ui = script.Ui(
		script_name='ex1pSco.py',
        files_local_path='.', 
        script_command='python3  ./<db_name>/ex1pScoViaDSS.py',
        returns=OrderedDict({"Cust_ID": INTEGER(), "Prob_0": FLOAT(), "Prob_1": FLOAT(), "Actual_Value": INTEGER()}))

from teradatamlwidgets import eda
from teradataml import DataFrame
df = DataFrame("ibm_stock")
ui = eda.Ui(df = df)

from teradatamlwidgets import describe
from teradataml import DataFrame
df = DataFrame("ibm_stock")
ui = describe.Ui(df = df)

from teradatamlwidgets import persist
from teradataml import DataFrame
df = DataFrame("ibm_stock")
ui = persist.Ui(df = df)

• Details

  • This package is useful to Data Scientists and Teradata users and provides following:

    • A simple UI experience within Jupyter Notebook.

    • Access to In-DB analytics

    • Visualizations

    • Integration with teradataml

    • Enable simple and easy integration with 3rd party workbenches

  • teradatamlwidgets.login.Ui Class

    • Purpose
      • Opens the function UI dialog in the notebook for the functions.
    • Function Output
      • This function will return instance of notebook UI interface.
    • Usage Considerations
      • If you are not already logged in then this will only allow you to log out otherwise the login screen is shown.

  • teradatamlwidgets.analytic_functions.Ui Class

    • Purpose
      • Opens the UI dialog in the notebook for the Analytic Functions (subset of the Analytics Database analytic functions, Vantage Analytics Library (VAL) functions, Unbounded Array Framework (UAF) time series functions).
    • Function Output
      • This function will return instance of notebook UI interface for analytic functions.
    • Usage Considerations
      • If you are not already logged in, the first time this is called, the “Login” user interface will be displayed so the user can log into a Teradata instance which creates the internal instance.

  • teradatamlwidgets.analytic_functions.get_output_dataframe Method

    • Purpose
      • Gets the DataFrame of the executed function.
    • Function Output
      • Return Value: teradataml.DataFrame. Returns the output of the function as a teradataml DataFrame.
    • Usage Considerations
      • NA

  • teradatamlwidgets.plot.Ui Class

    • Purpose
      • Allows a user interface for plotting that allows the user to set plotting parameters and then visualize the plots. The internal implementation uses the functionality of TD_PLOT exposed in teradataml DataFrame.
    • Function Output
      • This function will return instance of notebook UI interface for TD_PLOT.
    • Usage Considerations
      • If you are not already logged in, the first time this is called, the “Login” user interface will be displayed so the user can log into a Teradata instance which creates the internal instance.

  • teradatamlwidgets.plot.ShowPlots Method

    • Purpose
      • ShowPlots combines multiple plots together into one figure.

  • teradatamlwidgets.eda.Ui Class

    • Purpose
      • The Exploratory Data Analysis UI allows the user to take a deeper look into their dataset. The tabs include Data, Analyze, Visualize, Describe, and Persist. This provides visual components for scaled, in-Database Analytics with data that you keep in the Teradata Vantage Analytics Database within a notebook.
    • Function Output
      • This function will return instance of notebook EDA UI interface.
    • Usage Considerations
      • You must login before either using Login UI Class, or using teradataml create_context().

  • teradatamlwidgets.byom_functions.Ui Class

    • Purpose
      • Opens the UI dialog in the notebook for the BYOM functions.
    • Function Output
      • This function will return instance of notebook BYOM functions UI interface.
    • Usage Considerations
      • If you are not already logged in, the first time this is called, the “Login” user interface will be displayed so the user can log into a Teradata instance which creates the internal instance.

  • teradatamlwidgets.byom_functions.get_output_dataframe Method

    • Purpose
      • Gets the DataFrame of the executed function.
    • Function Output
      • Return Value: teradataml.DataFrame. Returns the output of the function as a teradataml DataFrame.
    • Usage Considerations
      • NA

  • teradatamlwidgets.auto_ml.Ui Class

    • Purpose
      • Opens the UI dialog in the notebook for the AutoML functions.
    • Function Output
      • This function will return instance of notebook AutoML UI interface.
    • Usage Considerations
      • If you are not already logged in, the first time this is called, the “Login” user interface will be displayed so the user can log into a Teradata instance which creates the internal instance.

  • teradatamlwidgets.auto_ml.get_prediction_dataframe Method

    • Purpose
      • To access the predicted output table.
    • Function Output
      • Return Value: teradataml.DataFrame.
    • Usage Considerations
      • NA

  • teradatamlwidgets.auto_ml.get_auto_ml Method

    • Purpose
      • To access the AutoML instance.
    • Function Output
      • Return Value: Pandas DataFrame.
    • Usage Considerations
      • NA

  • teradatamlwidgets.auto_ml.get_leaderboard Method

    • Purpose
      • To access the leaderboard.
    • Function Output
      • Return Value: teradataml.automl.AutoClassifier or teradataml.automl.AutoRegressor.
    • Usage Considerations
      • NA

  • teradatamlwidgets.script.Ui Class

    • Purpose
      • Opens the UI dialog in the notebook for the Script function.
    • Function Output
      • This function will return instance of notebook Script UI interface.
    • Usage Considerations
      • If you are not already logged in, the first time this is called, the “Login” user interface will be displayed so the user can log into a Teradata instance which creates the internal instance.

  • teradatamlwidgets.script.get_output_dataframe Method

    • Purpose
      • Gets the DataFrame of the script result.
    • Function Output
      • Return Value: teradataml.DataFrame. Returns the output of the function as a teradataml DataFrame.
    • Usage Considerations
      • NA

  • teradatamlwidgets.describe.Ui Class

    • Purpose
      • Allows user to see the dataFrame description and information, including Shape and Size, Column Statistics, Column Types, Column Summary, Categorical Summary, Futile Columns and Source Query.
    • Function Output
      • This function will return instance of notebook Describe UI interface.
    • Usage Considerations
      • You must login before either using Login UI Class, or using teradataml create_context().

  • teradatamlwidgets.persist.Ui Class

    • Purpose
      • Allows user to write records stored in a teradataml DataFrame to Teradata Vantage.
    • Function Output
      • This function will return instance of notebook Persist UI interface.
    • Usage Considerations
      • You must login before either using Login UI Class, or using teradataml create_context().

Documentation

General product information, including installation instructions, is available in the Teradata Documentation website

License

Use of the Teradata Widgets Package is governed by the License Agreement for the Teradata Widgets Package for Advanced Analytics. After installation, the LICENSE and LICENSE-3RD-PARTY files are located in the teradata_widget directory of the Python installation directory.