Merge pull request #10 from Azure/samuel100-patch1

general tidy of docs

Author: samuel100

provider/README.md

@@ -14,63 +14,42 @@ The _interoperable_ design of feast means that many Azure services can be used t
 
 ## 🐱‍👤 Getting Started
 
-### 1. Install Feast Azure Provider
+### Pre-requisites
+
+You will need to have:
 
-Git clone this repo and then:
+- An Azure account with an active subscription. 
+    - Don't have an Account? [You can create an Azure account with $200 free credit](https://azure.microsoft.com/free/).
+- Either an Azure SQL DB or Synapse SQL
+- A provisioned Azure Cache for Redis
+- An Azure Storage Account
+
+### 1. Install Feast Azure Provider
+Install the provider using `pip` (we recommend using either conda or virtualenv for environment isolation):
 
 ```bash
-cd provider/sdk
-pip install code/.
+pip install feast-azure-provider
 ```
 
-**Note: Best practice is to use environment isolation (for example: virtualenv or conda) when installing**
+> **Note**
+There is a dependency on the `unixodbc` operating system package. You can install this on Debian/Ubuntu using `sudo apt-get install unixodbc-dev`
 
-### 2. Connect to your feature store
+### 2. Create a feature repository
 
-To connect to your feature store, you can use either:
-
-- Feast Python SDK, or
-- Feast CLI
-
-#### a.) Using the Feast Python SDK
-
-```python
-import os
-from feast import FeatureStore, RepoConfig
-from feast.registry import RegistryConfig
-from feast_azure_provider.mssqlserver import MsSqlServerOfflineStoreConfig, MsSqlServerSource
-
-from feast import Entity, Feature, FeatureView, ValueType
-from datetime import timedelta
-
-# update this to your location
-registry_blob_url = "https://<ACCOUNT_NAME>.blob.core.windows.net/<CONTAINER>/<PATH>/registry.db"
-
-# update with your connection strings
-sql_conn_string = "mssql+pyodbc://<USER_NAME>:<PASSWORD>@<SERVER_NAME>.database.windows.net:1433/<DB_NAME>?driver=ODBC+Driver+17+for+SQL+Server&autocommit=True"
-
-redis_conn_string = "<CACHE_NAME>.redis.cache.windows.net:6380,password=<PASSWORD>,ssl=True"
+```bash
+feast init -m my_feature_repo
+cd my_feature_repo
+```
 
-reg_config = RegistryConfig(
-    registry_store_type="feast_azure_provider.registry_store.AzBlobRegistryStore",
-    path=registry_blob_url,
-)
+Rather than store credentials in your `feature_store.yaml` file, we recommend using environment variables:
 
-repo_cfg = RepoConfig(
-    registry=reg_config,
-    project="production",
-    provider="feast_azure_provider.azure_provider.AzureProvider",
-    offline_store=MsSqlServerOfflineStoreConfig(
-        connection_string=sql_conn_string
-        ),
-)
+```bash
+export SQL_CONN='mssql+pyodbc://<USER_NAME>:<PASSWORD>@<SERVER_NAME>.database.windows.net:1433/<DB_NAME>?driver=ODBC+Driver+17+for+SQL+Server&autocommit=True'
 
-store = FeatureStore(config=repo_cfg)
+export REDIS_CONN='<CACHE_NAME>.redis.cache.windows.net:6380,password=<PASSWORD>,ssl=True'
 ```
 
-#### b.) Using the Feast CLI
-
-Create your feature_store.yaml, which should contain the following:
+Update the `feature_store.yaml`:
 
 ```yaml
 project: production
@@ -80,28 +59,85 @@ registry:
 provider: feast_azure_provider.azure_provider.AzureProvider
 offline_store:
     type: feast_azure_provider.mssqlserver.MsSqlServerOfflineStore
-    connection_string: mssql+pyodbc://<USER_NAME>:<PASSWORD>@<SERVER_NAME>.database.windows.net:1433/<DB_NAME>?driver=ODBC+Driver+17+for+SQL+Server&autocommit=True
+    connection_string: ${SQL_CONN}
 online_store:
     type: redis
-    connection_string: <CACHE_NAME>.redis.cache.windows.net:6380,password=<PASSWORD>,ssl=True
+    connection_string: ${REDIS_CONN}
 ```
 
-To apply features:
+### 3. Register your feature definitions and set up your feature store
 
 ```bash
 feast apply
 ```
 
-To materialize features (adjust start and end times accordingly):
+### 4. Build a training dataset
+
+```python
+from feast import FeatureStore
+import pandas as pd
+from datetime import datetime
+
+entity_df = pd.DataFrame.from_dict({
+    "entity_id": [], # list of entities e.g. customer_id
+    "event_timestamp": [
+        datetime(2021, 4, 12, 10, 59, 42),
+        # list of datetimes
+    ]
+})
+
+store = FeatureStore(repo_path=".")
+
+training_df = store.get_historical_features(
+    entity_df=entity_df,
+    features = [
+        'featureview1:feature1',
+        'featureview1:feature2',
+        'featureview2:feature1'
+    ],
+).to_df()
+
+print(training_df.head())
+
+# Train model
+model = ml.fit(training_df)
+```
+
+### 5. Load feature values into your online store
 
 ```bash
-feast materialize '2021-08-01T19:20:01' '2021-08-03T19:20:01'
+CURRENT_TIME=$(date -u +"%Y-%m-%dT%H:%M:%S")
+feast materialize-incremental $CURRENT_TIME
+```
+
+### 6. Read online features at low latency
+
+```python
+from pprint import pprint
+from feast import FeatureStore
+
+store = FeatureStore(repo_path=".")
+
+feature_vector = store.get_online_features(
+    features=[
+        'featureview1:feature1',
+        'featureview1:feature2',
+        'featureview2:feature1'
+    ],
+    entity_rows=[{"entity_id": 1001}] # for example: driver_id, customer_id
+).to_dict()
+
+pprint(feature_vector)
+
+# Make prediction
+# model.predict(feature_vector)
 ```
 
+
 ## 🎓 Learn more
 
 - [Feast website](http://feast.dev)
-- [Feast on Azure tutorial](./docs/tutorial/README.md)
+- [Feast on Azure tutorial](./tutorial/README.md)
 
 ## Contributing