Deploy and Serve a Feature List for Model Predictions¶

Once you have a feature list ready, it's essential to serve these features effectively to make real-time or batch predictions using your machine learning model.

In this section, we'll take the feature list we previously crafted and explore two primary ways to serve its values:

REST API: Ideal for real-time predictions where you need instantaneous results. For instance, in applications where user interactions require immediate feedback based on model predictions.

Batch Processing: Best suited for scenarios where you have a bulk of data and don't require instant results.

In [1]:

import featurebyte as fb
from datetime import datetime

# Set your profile to the tutorial environment
fb.use_profile("tutorial")

catalog_name = "Grocery Dataset Tutorial"
catalog = fb.Catalog.activate(catalog_name)

import featurebyte as fb from datetime import datetime # Set your profile to the tutorial environment fb.use_profile("tutorial") catalog_name = "Grocery Dataset Tutorial" catalog = fb.Catalog.activate(catalog_name)

22:04:02 | INFO     | Using configuration file at: /Users/gxav/.featurebyte/config.yaml
22:04:02 | INFO     | Active profile: tutorial (https://tutorials.featurebyte.com/api/v1)
22:04:02 | WARNING  | Remote SDK version (0.5.0.dev6) is different from local (0.5.0.dev1). Update local SDK to avoid unexpected behavior.
22:04:02 | INFO     | No catalog activated.
22:04:03 | INFO     | 6 feature lists, 31 features deployed
22:04:03 | INFO     | Using profile: tutorial
22:04:03 | INFO     | Using configuration file at: /Users/gxav/.featurebyte/config.yaml
22:04:03 | INFO     | Active profile: tutorial (https://tutorials.featurebyte.com/api/v1)
22:04:03 | WARNING  | Remote SDK version (0.5.0.dev6) is different from local (0.5.0.dev1). Update local SDK to avoid unexpected behavior.
22:04:03 | INFO     | No catalog activated.
22:04:04 | INFO     | 6 feature lists, 31 features deployed
22:04:05 | INFO     | Catalog activated: Grocery Dataset Tutorial

List feature lists in Catalog¶

In [2]:

catalog.list_feature_lists()

catalog.list_feature_lists()

Out[2]:

	id	name	num_feature	status	deployed	readiness_frac	online_frac	tables	entities	primary_entities	created_at
0	64ff1dec72f1e0466e55f6a3	Customer Simple FeatureList	7	DRAFT	False	0.0	0.0	[GROCERYCUSTOMER, GROCERYINVOICE, INVOICEITEMS...	[customer]	[customer]	2023-09-11T14:02:35.008000

Get a feature list from Catalog¶

In [3]:

simple_feature_list = catalog.get_feature_list("Customer Simple FeatureList")

simple_feature_list = catalog.get_feature_list("Customer Simple FeatureList")

Loading Feature(s) |████████████████████████████████████████| 7/7 [100%] in 1.3s

Deploy feature list¶

In [4]:

# Create a deployment
deployment = simple_feature_list.deploy(
  deployment_name="Customer Spending forecast",
  make_production_ready=True,
)

# Create a deployment deployment = simple_feature_list.deploy( deployment_name="Customer Spending forecast", make_production_ready=True, )

Loading Feature(s) |████████████████████████████████████████| 7/7 [100%] in 1.0s
Done! |████████████████████████████████████████| 100% in 3.4s (0.30%/s)         

In [5]:

# Enable deployment
deployment.enable()

# Enable deployment deployment.enable()

Done! |████████████████████████████████████████| 100% in 40.6s (0.02%/s)        

In [6]:

# Check that the deployment is enabled
catalog.list_deployments()

# Check that the deployment is enabled catalog.list_deployments()

Out[6]:

	id	name	feature_list_name	feature_list_version	num_feature	enabled
0	64ff1e5c1b8535f0fa03767c	Customer Spending forecast	Customer Simple FeatureList	V230911	7	True

In [7]:

# Check status of the feature list
print("simple_feature_list.status:", simple_feature_list.status)

# Check status of the feature list print("simple_feature_list.status:", simple_feature_list.status)

simple_feature_list.status: DEPLOYED

Get template for online serving¶

In [8]:

# Get a python template for consuming the feature serving API
deployment.get_online_serving_code(language="python")

# Get a python template for consuming the feature serving API deployment.get_online_serving_code(language="python")

Loading Feature(s) |████████████████████████████████████████| 7/7 [100%] in 0.9s

Out[8]:

from typing import Any, Dict

import pandas as pd
import requests


def request_features(entity_serving_names: Dict[str, Any]) -> pd.DataFrame:
    """
    Send POST request to online serving endpoint

    Parameters
    ----------
    entity_serving_names: Dict[str, Any]
        Entity serving name values to used for serving request

    Returns
    -------
    pd.DataFrame
    """
    response = requests.post(
        url="https://tutorials.featurebyte.com/api/v1/deployment/64ff1e5c1b8535f0fa03767c/online_features",
        headers={"Content-Type": "application/json", "active-catalog-id": "64ff1c7d9a683e4e22365111", "Authorization": "Bearer gUaCOUiCSYpHazO10CQ4qXSccT0j6U2vsyFYxWhc5yA"},
        json={"entity_serving_names": entity_serving_names},
    )
    assert response.status_code == 200, response.json()
    return pd.DataFrame.from_dict(response.json()["features"])


request_features([{"GROCERYCUSTOMERGUID": "9bf67c68-0d34-48e2-9b0a-b5c3ee495b1c"}])

In [9]:

# Get shell script
deployment.get_online_serving_code(language="sh")

# Get shell script deployment.get_online_serving_code(language="sh")

Loading Feature(s) |████████████████████████████████████████| 7/7 [100%] in 0.9s

Out[9]:

#!/bin/sh

curl -X POST \
    -H 'Content-Type: application/json' \
    -H 'active-catalog-id: 64ff1c7d9a683e4e22365111' \
    -H 'Authorization: Bearer gUaCOUiCSYpHazO10CQ4qXSccT0j6U2vsyFYxWhc5yA' \
    -d '{"entity_serving_names": [{"GROCERYCUSTOMERGUID": "9bf67c68-0d34-48e2-9b0a-b5c3ee495b1c"}]}' \
    https://tutorials.featurebyte.com/api/v1/deployment/64ff1e5c1b8535f0fa03767c/online_features

Create batch request table¶

In [10]:

# Get view of current customer
customer_table = catalog.get_table("GROCERYCUSTOMER")
customer_view = customer_table.get_view(view_mode="manual")
cond = customer_view.CurrentRecord == True
current_customer_view = customer_view[cond]

# Get view of current customer customer_table = catalog.get_table("GROCERYCUSTOMER") customer_view = customer_table.get_view(view_mode="manual") cond = customer_view.CurrentRecord == True current_customer_view = customer_view[cond]

In [11]:

# Check primary entity of the deployed feature list to obtain its serving name
simple_feature_list.primary_entity

# Check primary entity of the deployed feature list to obtain its serving name simple_feature_list.primary_entity

Out[11]:

[<featurebyte.api.entity.Entity at 0x7f79d8061780>
 {
   'name': 'customer',
   'created_at': '2023-09-11T13:56:58.863000',
   'updated_at': '2023-09-11T13:57:16.943000',
   'description': None,
   'serving_names': [
     'GROCERYCUSTOMERGUID'
   ],
   'catalog_name': 'Grocery Dataset Tutorial'
 }]

In [12]:

# Create batch request table from the view
batch_request_table = current_customer_view.create_batch_request_table(
    name="Current Customers at " + datetime.now().strftime("%Y%m%d:%H%M"),
    columns=["GroceryCustomerGuid"],
    columns_rename_mapping={
        "GroceryCustomerGuid": "GROCERYCUSTOMERGUID",
    }
)

# Create batch request table from the view batch_request_table = current_customer_view.create_batch_request_table( name="Current Customers at " + datetime.now().strftime("%Y%m%d:%H%M"), columns=["GroceryCustomerGuid"], columns_rename_mapping={ "GroceryCustomerGuid": "GROCERYCUSTOMERGUID", } )

Done! |████████████████████████████████████████| 100% in 6.7s (0.15%/s)         

In [13]:

# Get name of the batch request table
batch_request_table.name

# Get name of the batch request table batch_request_table.name

Out[13]:

'Current Customers at 20230911:2205'

In [14]:

# List batch request tables in catalog
catalog.list_batch_request_tables()

# List batch request tables in catalog catalog.list_batch_request_tables()

Out[14]:

	id	name	type	shape	feature_store_name	created_at
0	64ff1ea91b8535f0fa03767d	Current Customers at 20230911:2205	view	[500, 1]	playground	2023-09-11T14:05:31.608000

Compute batch feature table¶

In [15]:

# Get deployment and batch request table
deployment = catalog.get_deployment("Customer Spending forecast")
batch_request_table = catalog.get_batch_request_table(batch_request_table.name)

# Get deployment and batch request table deployment = catalog.get_deployment("Customer Spending forecast") batch_request_table = catalog.get_batch_request_table(batch_request_table.name)

In [16]:

# Compute batch features
batch_features = deployment.compute_batch_feature_table(
    batch_request_table=batch_request_table,
    batch_feature_table_name = 
    f"Customer Simple FeatureList for Spending forecast with {batch_request_table.name}"
)

# Compute batch features batch_features = deployment.compute_batch_feature_table( batch_request_table=batch_request_table, batch_feature_table_name = f"Customer Simple FeatureList for Spending forecast with {batch_request_table.name}" )

Done! |████████████████████████████████████████| 100% in 6.7s (0.15%/s)         

In [17]:

# List observation tables
catalog.list_batch_feature_tables()

# List observation tables catalog.list_batch_feature_tables()

Out[17]:

	id	name	feature_store_name	batch_request_table_name	shape	created_at
0	64ff1eb41b8535f0fa03767e	Customer Simple FeatureList for Spending forec...	playground	Current Customers at 20230911:2205	[500, 8]	2023-09-11T14:05:44.586000

In [18]:

# Convert to pandas 
batch_features.to_pandas()

# Convert to pandas batch_features.to_pandas()

Downloading table |████████████████████████████████████████| 500/500 [100%] in 0

Out[18]:

	GROCERYCUSTOMERGUID	CUSTOMER_Age_band	CUSTOMER_Latest_invoice_Amount	CUSTOMER_Count_of_invoice_14d	CUSTOMER_Avg_of_invoice_Amount_14d	CUSTOMER_Std_of_invoice_Amount_14d	CUSTOMER_Latest_invoice_Amount_Z_Score_to_invoice_Amount_28d	CUSTOMER_vs_OVERALL_item_TotalCost_across_product_ProductGroups_26w
0	217d44e9-d85b-4dba-9847-2c4c071f7303	40-44	11.86	4.0	9.2025	2.485482	0.218373	0.458607
1	59232f53-fd8c-4ea6-abe1-91c7da41e793	75-79	12.16	4.0	18.1350	5.300818	-0.187217	0.432918
2	f1ee8601-2f7e-4aa3-9756-19ca5c8f3bcc	70-74	57.16	1.0	57.1600	0.000000	NaN	0.328127
3	aa5a6b30-7acd-4d96-a3b3-f09f75345ca9	80-84	0.99	1.0	0.9900	0.000000	NaN	0.601628
4	682fb64a-5ca6-48f6-9de1-6d9d6321db39	65-69	12.75	1.0	12.7500	0.000000	0.886151	0.662089
...	...	...	...	...	...	...	...	...
495	da9493e7-46b5-4e97-9f4c-497048901a0d	25-29	6.29	1.0	6.2900	0.000000	1.000000	0.467478
496	71b4cc67-21e7-4a73-819c-adc6da260127	40-44	29.75	0.0	NaN	NaN	NaN	0.655595
497	29dc945d-1c9e-46a7-a5d7-ba6d24182195	65-69	7.51	1.0	7.5100	0.000000	NaN	0.540942
498	dbb91a85-9e60-4f3a-90bc-f156d6763249	50-54	11.95	0.0	NaN	NaN	NaN	0.668405
499	77fac0cb-6a8d-4f59-9a1a-c324ab36e885	65-69	34.45	0.0	NaN	NaN	NaN	0.247293

500 rows × 8 columns

In [19]:

# download parquet file
batch_features.download()

# download parquet file batch_features.download()

Downloading table |████████████████████████████████████████| 500/500 [100%] in 0

Out[19]:

PosixPath('BATCH_FEATURE_TABLE_64ff1eb5ceb61c9075514aaf.parquet')

In [20]:

# delete if not needed any more
batch_features.delete()
batch_request_table.delete()

# delete if not needed any more batch_features.delete() batch_request_table.delete()

Done! |████████████████████████████████████████| 100% in 6.8s (0.15%/s)         
Done! |████████████████████████████████████████| 100% in 6.7s (0.15%/s)         

Manage Deployment¶

In [21]:

deployment = catalog.get_deployment("Customer Spending forecast")

deployment = catalog.get_deployment("Customer Spending forecast")

In [22]:

# get feature jobs status (this will produce meaningful results once multiple jobs have been run)
deployment.get_feature_jobs_status()

# get feature jobs status (this will produce meaningful results once multiple jobs have been run) deployment.get_feature_jobs_status()

Loading Feature(s) |████████████████████████████████████████| 7/7 [100%] in 1.0s

Out[22]:

Job statistics (last 1 hours)

	request_date	job_history_window	job_duration_tolerance
0	2023-09-11T14:06:13.209534	1	60

	feature_name	aggregation_hash
0	CUSTOMER_Avg_of_invoice_Amount_14d	b67eafff
1	CUSTOMER_Count_of_invoice_14d	081a79b3
2	CUSTOMER_Latest_invoice_Amount	3e664139
3	CUSTOMER_Latest_invoice_Amount_Z_Score_to_invoice_Amount_28d	3e664139
4	CUSTOMER_Latest_invoice_Amount_Z_Score_to_invoice_Amount_28d	b67eafff
5	CUSTOMER_Latest_invoice_Amount_Z_Score_to_invoice_Amount_28d	cf255092
6	CUSTOMER_Std_of_invoice_Amount_14d	cf255092
7	CUSTOMER_vs_OVERALL_item_TotalCost_across_product_ProductGroups_26w	c8d4aa8c
8	CUSTOMER_vs_OVERALL_item_TotalCost_across_product_ProductGroups_26w	6d77adb1

	aggregation_hash	frequency(min)	completed_jobs	max_duration(s)	95 percentile	frac_late	exceed_period	failed_jobs	incomplete_jobs	time_since_last
0	3e664139	60	0	NaN	NaN	NaN	0	0	1	NaT
1	081a79b3	60	0	NaN	NaN	NaN	0	0	1	NaT
2	b67eafff	60	0	NaN	NaN	NaN	0	0	1	NaT
3	cf255092	60	0	NaN	NaN	NaN	0	0	1	NaT
4	c8d4aa8c	60	0	NaN	NaN	NaN	0	0	1	NaT
5	6d77adb1	60	0	NaN	NaN	NaN	0	0	1	NaT

In [23]:

# Disable deployment
deployment.disable()

# Disable deployment deployment.disable()

Done! |████████████████████████████████████████| 100% in 10.2s (0.10%/s)        

In [24]:

# The deployment is still part of the catalog but disabled
catalog.list_deployments()

# The deployment is still part of the catalog but disabled catalog.list_deployments()

Out[24]:

	id	name	feature_list_name	feature_list_version	num_feature	enabled
0	64ff1e5c1b8535f0fa03767c	Customer Spending forecast	Customer Simple FeatureList	V230911	7	False

Concepts in this tutorial¶

• Feature list deployment
• REST API and batch serving
• Feature jobs

SDK reference for¶

• Deployment
• Batch Request Table
• Batch Feature Table
• FeatureList.deploy()
• Deployment.enable()
• Deployment.get_online_serving_code()
• Deployment.compute batch feature_table()
• Table.get_view() in a manual mode
• View.create_batch_request_table()
• SourceTable.create_batch_request_table()

In [ ]: