Tutorial, Parte 2: Creación de una aplicación de recuperación de conocimiento personalizada (RAG) con el SDK de Azure AI Foundry

Artículo
11/19/2024

En este tutorial, usará el SDK de Azure AI Foundry (y otras bibliotecas) para compilar, configurar, evaluar e implementar una aplicación de chat para su empresa minorista denominada Contoso Trek. Su empresa minorista se especializa en ropa y equipo de camping al aire libre. La aplicación de chat debe responder a preguntas sobre sus productos y servicios. Por ejemplo, la aplicación de chat puede responder a preguntas como "¿Qué tienda es la más impermeable?" o "¿cuál es la mejor bolsa de dormir para tiempo frío?".

En esta segunda parte se muestra cómo mejorar una aplicación de chat básica mediante la adición de la generación aumentada de recuperación (RAG) para fundamentar las respuestas en los datos personalizados. La generación aumentada de recuperación (RAG) es un patrón que usa los datos con un modelo de lenguaje grande (LLM) para generar respuestas específicas a los datos. En esta segunda parte, aprenderá a:

Creación de un índice de búsqueda de los datos de la aplicación de chat que se va a usar
Desarrollo de código de RAG personalizado

Este tutorial es la segunda parte de un tutorial de tres partes.

Requisitos previos

Complete Tutorial, Parte 1: Creación de recursos para desarrollar una aplicación de chat personalizada con el SDK de Azure AI para lo siguiente:
- Creación de un proyecto con un índice de Búsqueda de Azure AI conectado
- Instalación de la CLI de Azure, Python y los paquetes necesarios
- Configuración de las variables de entorno

Creación de datos de ejemplo para la aplicación de chat

El objetivo con esta aplicación basada en la RAG es fundamentar las respuestas del modelo en los datos personalizados. Use un índice de Búsqueda de Azure AI que almacene datos vectorizados del modelo de inserciones. El índice de búsqueda se usa para recuperar documentos relevantes en función de la pregunta del usuario.

Si ya tiene un índice de búsqueda con datos, puede ir directamente a Obtener documentos del producto. De lo contrario, puede crear un conjunto de datos de ejemplo sencillo para usarlo en la aplicación de chat.

Cree un directorio recursos y agregue estos datos de ejemplo a un archivo products.csv:

id,name,price,category,brand,description
1,TrailMaster X4 Tent,250.0,Tents,OutdoorLiving,"Unveiling the TrailMaster X4 Tent from OutdoorLiving, your home away from home for your next camping adventure. Crafted from durable polyester, this tent boasts a spacious interior perfect for four occupants. It ensures your dryness under drizzly skies thanks to its water-resistant construction, and the accompanying rainfly adds an extra layer of weather protection. It offers refreshing airflow and bug defence, courtesy of its mesh panels. Accessibility is not an issue with its multiple doors and interior pockets that keep small items tidy. Reflective guy lines grant better visibility at night, and the freestanding design simplifies setup and relocation. With the included carry bag, transporting this convenient abode becomes a breeze. Be it an overnight getaway or a week-long nature escapade, the TrailMaster X4 Tent provides comfort, convenience, and concord with the great outdoors. Comes with a two-year limited warranty to ensure customer satisfaction."
2,Adventurer Pro Backpack,90.0,Backpacks,HikeMate,"Venture into the wilderness with the HikeMate's Adventurer Pro Backpack! Uniquely designed with ergonomic comfort in mind, this backpack ensures a steadfast journey no matter the mileage. It boasts a generous 40L capacity wrapped up in durable nylon fabric ensuring its long-lasting performance on even the most rugged pursuits. It's meticulously fashioned with multiple compartments and pockets for organized storage, hydration system compatibility, and adjustable padded shoulder straps all in a lightweight construction. The added features of a sternum strap and hip belt enhance stability without compromising on comfort. The Adventurer Pro Backpack also prioritizes your safety with its reflective accents for when night falls. This buoyant beauty does more than carry your essentials; it carries the promise of a stress-free adventure!"
3,Summit Breeze Jacket,120.0,Hiking Clothing,MountainStyle,"Discover the joy of hiking with MountainStyle's Summit Breeze Jacket. This lightweight jacket is your perfect companion for outdoor adventures. Sporting a trail-ready, windproof design and a water-resistant fabric, it's ready to withstand any weather. The breathable polyester material and adjustable cuffs keep you comfortable, whether you're ascending a mountain or strolling through a park. And its sleek black color adds style to function. The jacket features a full-zip front closure, adjustable hood, and secure zippered pockets. Experience the comfort of its inner lining and the convenience of its packable design. Crafted for night trekkers too, the jacket has reflective accents for enhanced visibility. Rugged yet chic, the Summit Breeze Jacket is more than a hiking essential, it's the gear that inspires you to reach new heights. Choose adventure, choose the Summit Breeze Jacket."
4,TrekReady Hiking Boots,140.0,Hiking Footwear,TrekReady,"Introducing the TrekReady Hiking Boots - stepping up your hiking game, one footprint at a time! Crafted from leather, these stylistic Trailmates are made to last. TrekReady infuses durability with its reinforced stitching and toe protection, making sure your journey is never stopped short. Comfort? They have that covered too! The boots are a haven with their breathable materials, cushioned insole, with padded collar and tongue; all nestled neatly within their lightweight design. As they say, it's what's inside that counts - so inside you'll find a moisture-wicking lining that quarantines stank and keeps your feet fresh as that mountaintop breeze. Remember the fear of slippery surfaces? With these boots, you can finally tell it to 'take a hike'! Their shock-absorbing midsoles and excellent traction capabilities promise stability at your every step. Beautifully finished in a traditional lace-up system, every adventurer deserves a pair of TrekReady Hiking Boots. Hike more, worry less!"
5,BaseCamp Folding Table,60.0,Camping Tables,CampBuddy,"CampBuddy's BaseCamp Folding Table is an adventurer's best friend. Lightweight yet powerful, the table is a testament to fun-meets-function and will elevate any outing to new heights. Crafted from resilient, rust-resistant aluminum, the table boasts a generously sized 48 x 24 inches tabletop, perfect for meal times, games and more. The foldable design is a godsend for on-the-go explorers. Adjustable legs rise to the occasion to conquer uneven terrains and offer height versatility, while the built-in handle simplifies transportation. Additional features like non-slip feet, integrated cup holders and mesh pockets add a pinch of finesse. Quick to set up without the need for extra tools, this table is a silent yet indispensable sidekick during camping, picnics, and other outdoor events. Don't miss out on the opportunity to take your outdoor experiences to a new level with the BaseCamp Folding Table. Get yours today and embark on new adventures tomorrow! "
6,EcoFire Camping Stove,80.0,Camping Stoves,EcoFire,"Introducing EcoFire's Camping Stove, your ultimate companion for every outdoor adventure! This portable wonder is precision-engineered with a lightweight and compact design, perfect for capturing that spirit of wanderlust. Made from high-quality stainless steel, it promises durability and steadfast performance. This stove is not only fuel-efficient but also offers an easy, intuitive operation that ensures hassle-free cooking. Plus, it's flexible, accommodating a variety of cooking methods whether you're boiling, grilling, or simmering under the starry sky. Its stable construction, quick setup, and adjustable flame control make cooking a breeze, while safety features protect you from any potential mishaps. And did we mention it also includes an effective wind protector and a carry case for easy transportation? But that's not all! The EcoFire Camping Stove is eco-friendly, designed to minimize environmental impact. So get ready to enhance your camping experience and enjoy delicious outdoor feasts with this unique, versatile stove!"
7,CozyNights Sleeping Bag,100.0,Sleeping Bags,CozyNights,"Embrace the great outdoors in any season with the lightweight CozyNights Sleeping Bag! This durable three-season bag is superbly designed to give hikers, campers, and backpackers comfort and warmth during spring, summer, and fall. With a compact design that folds down into a convenient stuff sack, you can whisk it away on any adventure without a hitch. The sleeping bag takes comfort seriously, featuring a handy hood, ample room and padding, and a reliable temperature rating. Crafted from high-quality polyester, it ensures long-lasting use and can even be zipped together with another bag for shared comfort. Whether you're gazing at stars or catching a quick nap between trails, the CozyNights Sleeping Bag makes it a treat. Don't just sleep— dream with CozyNights."
8,Alpine Explorer Tent,350.0,Tents,AlpineGear,"Welcome to the joy of camping with the Alpine Explorer Tent! This robust, 8-person, 3-season marvel is from the responsible hands of the AlpineGear brand. Promising an enviable setup that is as straightforward as counting sheep, your camping experience is transformed into a breezy pastime. Looking for privacy? The detachable divider provides separate spaces at a moment's notice. Love a tent that breathes? The numerous mesh windows and adjustable vents fend off any condensation dragon trying to dampen your adventure fun. The waterproof assurance keeps you worry-free during unexpected rain dances. With a built-in gear loft to stash away your outdoor essentials, the Alpine Explorer Tent emerges as a smooth balance of privacy, comfort, and convenience. Simply put, this tent isn't just a shelter - it's your second home in the heart of nature! Whether you're a seasoned camper or a nature-loving novice, this tent makes exploring the outdoors a joyous journey."
9,SummitClimber Backpack,120.0,Backpacks,HikeMate,"Adventure waits for no one! Introducing the HikeMate SummitClimber Backpack, your reliable partner for every exhilarating journey. With a generous 60-liter capacity and multiple compartments and pockets, packing is a breeze. Every feature points to comfort and convenience; the ergonomic design and adjustable hip belt ensure a pleasantly personalized fit, while padded shoulder straps protect you from the burden of carrying. Venturing into wet weather? Fear not! The integrated rain cover has your back, literally. Stay hydrated thanks to the backpack's hydration system compatibility. Travelling during twilight? Reflective accents keep you visible in low-light conditions. The SummitClimber Backpack isn't merely a carrier; it's a wearable base camp constructed from ruggedly durable nylon and thoughtfully designed for the great outdoors adventurer, promising to withstand tough conditions and provide years of service. So, set off on that quest - the wild beckons! The SummitClimber Backpack - your hearty companion on every expedition!"
10,TrailBlaze Hiking Pants,75.0,Hiking Clothing,MountainStyle,"Meet the TrailBlaze Hiking Pants from MountainStyle, the stylish khaki champions of the trails. These are not just pants; they're your passport to outdoor adventure. Crafted from high-quality nylon fabric, these dapper troopers are lightweight and fast-drying, with a water-resistant armor that laughs off light rain. Their breathable design whisks away sweat while their articulated knees grant you the flexibility of a mountain goat. Zippered pockets guard your essentials, making them a hiker's best ally. Designed with durability for all your trekking trials, these pants come with a comfortable, ergonomic fit that will make you forget you're wearing them. Sneak a peek, and you are sure to be tempted by the sleek allure that is the TrailBlaze Hiking Pants. Your outdoors wardrobe wouldn't be quite complete without them."
11,TrailWalker Hiking Shoes,110.0,Hiking Footwear,TrekReady,"Meet the TrekReady TrailWalker Hiking Shoes, the ideal companion for all your outdoor adventures. Constructed with synthetic leather and breathable mesh, these shoes are tough as nails yet surprisingly airy. Their cushioned insoles offer fabulous comfort for long hikes, while the supportive midsoles and traction outsoles with multidirectional lugs ensure stability and excellent grip. A quick-lace system, padded collar and tongue, and reflective accents make these shoes a dream to wear. From combating rough terrain with the reinforced toe cap and heel, to keeping off trail debris with the protective mudguard, the TrailWalker Hiking Shoes have you covered. These waterproof warriors are made to endure all weather conditions. But they're not just about being rugged, they're light as a feather too, minimizing fatigue during epic hikes. Each pair can be customized for a perfect fit with removable insoles and availability in multiple sizes and widths. Navigate hikes comfortably and confidently with the TrailWalker Hiking Shoes. Adventure, here you come!"
12,TrekMaster Camping Chair,50.0,Camping Tables,CampBuddy,"Gravitate towards comfort with the TrekMaster Camping Chair from CampBuddy. This trusty outdoor companion boasts sturdy construction using high-quality materials that promise durability and enjoyment for seasons to come. Impeccably lightweight and portable, it's designed to be your go-to seat whether you're camping, at a picnic, cheering at a sporting event, or simply relishing in your backyard pleasures. Beyond its foldable design ensuring compact storage and easy transportation, its ergonomic magic is in the details. An adjustable recline, padded seat and backrest, integrated cup holder, and side pockets ensure the greatest outdoor comfort. Weather resistant, easy to clean, and capable of supporting diverse body types, this versatile chair also comes with a carry bag, ready for your next adventure."
13,PowerBurner Camping Stove,100.0,Camping Stoves,PowerBurner,"Unleash your inner explorer with the PowerBurner Dual Burner Camping Stove. It's designed for the adventurous heart, with sturdy construction and a high heat output that makes boiling and cooking a breeze. This stove isn't just about strength—it's got finesse too. With adjustable flame control, you can simmer, sauté, or sizzle with absolute precision. Its compact design and integrated carrying handle make transportation effortless. Moreover, it's crafted to defy the elements, boasting a wind-resistant exterior and piezo ignition system for quick, reliable starts. And when the cooking's done, its removable grates make cleanup swift and easy. Rugged, versatile and reliable, the PowerBurner marks a perfect blend of practicality and performance. So, why wait? Let's turn up the heat on your outdoor culinary adventures today."
14,MountainDream Sleeping Bag,130.0,Sleeping Bags,MountainDream,"Meet the MountainDream Sleeping Bag: your new must-have companion for every outdoor adventure. Designed to handle 3-season camping with ease, it comes equipped with a premium synthetic insulation that will keep you cozy even when temperatures fall down to 15°F! Sporting a durable water-resistant nylon shell and soft breathable polyester lining, this bag doesn't sacrifice comfort for toughness. The star of the show is the contoured mummy shape that not only provides optimal heat retention but also cuts down on the weight. A smooth, snag-free YKK zipper with a unique anti-snag design allows for hassle-free operation, while the adjustable hood and full-length zipper baffle work together to ensure you stay warm all night long. Need to bring along some essentials? Not to worry! There's an interior pocket just for that. And when it's time to pack up? Just slip it into the included compression sack for easy storage and transport. Whether you're a backpacking pro or a camping novice, the MountainDream Sleeping Bag is the perfect blend of durability, warmth, and comfort that you've been looking for."
15,SkyView 2-Person Tent,200.0,Tents,OutdoorLiving,"Introducing the OutdoorLiving SkyView 2-Person Tent, a perfect companion for your camping and hiking adventures. This tent offers a spacious interior that houses two people comfortably, with room to spare. Crafted from durable waterproof materials to shield you from the elements, it is the fortress you need in the wild. Setup is a breeze thanks to its intuitive design and color-coded poles, while two large doors allow for easy access. Stay organized with interior pockets, and store additional gear in its two vestibules. The tent also features mesh panels for effective ventilation, and it comes with a rainfly for extra weather protection. Light enough for on-the-go adventurers, it packs compactly into a carrying bag for seamless transportation. Reflective guy lines ensure visibility at night for added safety, and the tent stands freely for versatile placement. Experience the reliability of double-stitched seams that guarantee increased durability, and rest easy under the stars with OutdoorLiving's SkyView 2-Person Tent. It's not just a tent; it's your home away from home."
16,TrailLite Daypack,60.0,Backpacks,HikeMate,"Step up your hiking game with HikeMate's TrailLite Daypack. Built for comfort and efficiency, this lightweight and durable backpack offers a spacious main compartment, multiple pockets, and organization-friendly features all in one sleek package. The adjustable shoulder straps and padded back panel ensure optimal comfort during those long exhilarating treks. Course through nature without worry as the daypack's water-resistant fabric protects your essentials from unexpected showers. Plus, never run dry with the integrated hydration system. And did we mention it comes in a plethora of colors and designs? So you can choose one that truly speaks to your outdoorsy soul! Keeping your visibility in mind, we've added reflective accents that light up in low-light conditions. Don't just carry a backpack, adorn a companion that takes you a step ahead in your adventures. Trust the TrailLite Daypack for a hassle-free, enjoyable hiking experience."
17,RainGuard Hiking Jacket,110.0,Hiking Clothing,MountainStyle,"Introducing the MountainStyle RainGuard Hiking Jacket - the ultimate solution for weatherproof comfort during your outdoor undertakings! Designed with waterproof, breathable fabric, this jacket promises an outdoor experience that's as dry as it is comfortable. The rugged construction assures durability, while the adjustable hood provides a customizable fit against wind and rain. Featuring multiple pockets for safe, convenient storage and adjustable cuffs and hem, you can tailor the jacket to suit your needs on-the-go. And, don't worry about overheating during intense activities - it's equipped with ventilation zippers for increased airflow. Reflective details ensure visibility even during low-light conditions, making it perfect for evening treks. With its lightweight, packable design, carrying it inside your backpack requires minimal effort. With options for men and women, the RainGuard Hiking Jacket is perfect for hiking, camping, trekking and countless other outdoor adventures. Don't let the weather stand in your way - embrace the outdoors with MountainStyle RainGuard Hiking Jacket!"
18,TrekStar Hiking Sandals,70.0,Hiking Footwear,TrekReady,"Meet the TrekStar Hiking Sandals from TrekReady - the ultimate trail companion for your feet. Designed for comfort and durability, these lightweight sandals are perfect for those who prefer to see the world from a hiking trail. They feature adjustable straps for a snug, secure fit, perfect for adapting to the contours of your feet. With a breathable design, your feet will stay cool and dry, escaping the discomfort of sweaty hiking boots on long summer treks. The deep tread rubber outsole ensures excellent traction on any terrain, while the cushioned footbed promises enhanced comfort with every step. For those wild and unpredictable trails, the added toe protection and shock-absorbing midsole protect your feet from rocky surprises. Ingeniously, the removable insole makes for easy cleaning and maintenance, extending the lifespan of your sandals. Available in various sizes and a handsome brown color, the versatile TrekStar Hiking Sandals are just as comfortable on a casual walk in the park as they are navigating rocky slopes. Explore more with TrekReady!"
19,Adventure Dining Table,90.0,Camping Tables,CampBuddy,"Discover the joy of outdoor adventures with the CampBuddy Adventure Dining Table. This feature-packed camping essential brings both comfort and convenience to your memorable trips. Made from high-quality aluminum, it promises long-lasting performance, weather resistance, and easy maintenance - all key for the great outdoors! It's light, portable, and comes with adjustable height settings to suit various seating arrangements and the spacious surface comfortably accommodates meals, drinks, and other essentials. The sturdy yet lightweight frame holds food, dishes, and utensils with ease. When it's time to pack up, it fold and stows away with no fuss, ready for the next adventure!  Perfect for camping, picnics, barbecues, and beach outings - its versatility shines as brightly as the summer sun! Durable, sturdy and a breeze to set up, the Adventure Dining Table will be a loyal companion on every trip. Embark on your next adventure and make lifetime memories with CampBuddy. As with all good experiences, it'll leave you wanting more! "
20,CompactCook Camping Stove,60.0,Camping Stoves,CompactCook,"Step into the great outdoors with the CompactCook Camping Stove, a convenient, lightweight companion perfect for all your culinary camping needs. Boasting a robust design built for harsh environments, you can whip up meals anytime, anywhere. Its wind-resistant and fuel-versatile features coupled with an efficient cooking performance, ensures you won't have to worry about the elements or helpless taste buds while on adventures. The easy ignition technology and adjustable flame control make cooking as easy as a walk in the park, while its compact, foldable design makes packing a breeze. Whether you're camping with family or hiking solo, this reliable, portable stove is an essential addition to your gear. With its sturdy construction and safety-focused design, the CompactCook Camping Stove is a step above the rest, providing durability, quality, and peace of mind. Be wild, be free, be cooked for with the CompactCook Camping Stove!"

Creación de un índice de búsqueda

El índice de búsqueda se usa para almacenar datos vectorizados del modelo de incrustaciones. El índice de búsqueda se usa para recuperar documentos relevantes en función de la pregunta del usuario.

Cree el archivo create_search_index.py en la carpeta principal (es decir, el mismo directorio donde colocó la carpeta recursos, no dentro de la carpeta recursos).
Copie y pegue el siguiente código en el archivo create_search_index.py.

Agregue el código para importar las bibliotecas necesarias, cree un cliente de proyecto y configure algunas opciones:

import os
from azure.ai.projects import AIProjectClient
from azure.ai.projects.models import ConnectionType
from azure.identity import DefaultAzureCredential
from azure.core.credentials import AzureKeyCredential
from azure.search.documents import SearchClient
from azure.search.documents.indexes import SearchIndexClient
from config import get_logger

# initialize logging object
logger = get_logger(__name__)

# create a project client using environment variables loaded from the .env file
project = AIProjectClient.from_connection_string(
    conn_str=os.environ["AIPROJECT_CONNECTION_STRING"], credential=DefaultAzureCredential()
)

# create a vector embeddings client that will be used to generate vector embeddings
embeddings = project.inference.get_embeddings_client()

# use the project client to get the default search connection
search_connection = project.connections.get_default(
    connection_type=ConnectionType.AZURE_AI_SEARCH, include_credentials=True
)

# Create a search index client using the search connection
# This client will be used to create and delete search indexes
index_client = SearchIndexClient(
    endpoint=search_connection.endpoint_url, credential=AzureKeyCredential(key=search_connection.key)
)

Ahora agregue la función para definir un índice de búsqueda:

import pandas as pd
from azure.search.documents.indexes.models import (
    SemanticSearch,
    SearchField,
    SimpleField,
    SearchableField,
    SearchFieldDataType,
    SemanticConfiguration,
    SemanticPrioritizedFields,
    SemanticField,
    VectorSearch,
    HnswAlgorithmConfiguration,
    VectorSearchAlgorithmKind,
    HnswParameters,
    VectorSearchAlgorithmMetric,
    ExhaustiveKnnAlgorithmConfiguration,
    ExhaustiveKnnParameters,
    VectorSearchProfile,
    SearchIndex,
)


def create_index_definition(index_name: str, model: str) -> SearchIndex:
    dimensions = 1536  # text-embedding-ada-002
    if model == "text-embedding-3-large":
        dimensions = 3072

    # The fields we want to index. The "embedding" field is a vector field that will
    # be used for vector search.
    fields = [
        SimpleField(name="id", type=SearchFieldDataType.String, key=True),
        SearchableField(name="content", type=SearchFieldDataType.String),
        SimpleField(name="filepath", type=SearchFieldDataType.String),
        SearchableField(name="title", type=SearchFieldDataType.String),
        SimpleField(name="url", type=SearchFieldDataType.String),
        SearchField(
            name="contentVector",
            type=SearchFieldDataType.Collection(SearchFieldDataType.Single),
            searchable=True,
            # Size of the vector created by the text-embedding-ada-002 model.
            vector_search_dimensions=dimensions,
            vector_search_profile_name="myHnswProfile",
        ),
    ]

    # The "content" field should be prioritized for semantic ranking.
    semantic_config = SemanticConfiguration(
        name="default",
        prioritized_fields=SemanticPrioritizedFields(
            title_field=SemanticField(field_name="title"),
            keywords_fields=[],
            content_fields=[SemanticField(field_name="content")],
        ),
    )

    # For vector search, we want to use the HNSW (Hierarchical Navigable Small World)
    # algorithm (a type of approximate nearest neighbor search algorithm) with cosine
    # distance.
    vector_search = VectorSearch(
        algorithms=[
            HnswAlgorithmConfiguration(
                name="myHnsw",
                kind=VectorSearchAlgorithmKind.HNSW,
                parameters=HnswParameters(
                    m=4,
                    ef_construction=1000,
                    ef_search=1000,
                    metric=VectorSearchAlgorithmMetric.COSINE,
                ),
            ),
            ExhaustiveKnnAlgorithmConfiguration(
                name="myExhaustiveKnn",
                kind=VectorSearchAlgorithmKind.EXHAUSTIVE_KNN,
                parameters=ExhaustiveKnnParameters(metric=VectorSearchAlgorithmMetric.COSINE),
            ),
        ],
        profiles=[
            VectorSearchProfile(
                name="myHnswProfile",
                algorithm_configuration_name="myHnsw",
            ),
            VectorSearchProfile(
                name="myExhaustiveKnnProfile",
                algorithm_configuration_name="myExhaustiveKnn",
            ),
        ],
    )

    # Create the semantic settings with the configuration
    semantic_search = SemanticSearch(configurations=[semantic_config])

    # Create the search index definition
    return SearchIndex(
        name=index_name,
        fields=fields,
        semantic_search=semantic_search,
        vector_search=vector_search,
    )

Cree la función para agregar un archivo CSV al índice:

# define a function for indexing a csv file, that adds each row as a document
# and generates vector embeddings for the specified content_column
def create_docs_from_csv(path: str, content_column: str, model: str) -> list[dict[str, any]]:
    products = pd.read_csv(path)
    items = []
    for product in products.to_dict("records"):
        content = product[content_column]
        id = str(product["id"])
        title = product["name"]
        url = f"/products/{title.lower().replace(' ', '-')}"
        emb = embeddings.embed(input=content, model=model)
        rec = {
            "id": id,
            "content": content,
            "filepath": f"{title.lower().replace(' ', '-')}",
            "title": title,
            "url": url,
            "contentVector": emb.data[0].embedding,
        }
        items.append(rec)

    return items


def create_index_from_csv(index_name, csv_file):
    # If a search index already exists, delete it:
    try:
        index_definition = index_client.get_index(index_name)
        index_client.delete_index(index_name)
        logger.info(f"🗑️  Found existing index named '{index_name}', and deleted it")
    except Exception:
        pass

    # create an empty search index
    index_definition = create_index_definition(index_name, model=os.environ["EMBEDDINGS_MODEL"])
    index_client.create_index(index_definition)

    # create documents from the products.csv file, generating vector embeddings for the "description" column
    docs = create_docs_from_csv(path=csv_file, content_column="description", model=os.environ["EMBEDDINGS_MODEL"])

    # Add the documents to the index using the Azure AI Search client
    search_client = SearchClient(
        endpoint=search_connection.endpoint_url,
        index_name=index_name,
        credential=AzureKeyCredential(key=search_connection.key),
    )

    search_client.upload_documents(docs)
    logger.info(f"➕ Uploaded {len(docs)} documents to '{index_name}' index")

Por último, ejecute las funciones para compilar el índice y registrarlo en el proyecto en la nube:

if __name__ == "__main__":
    import argparse

    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--index-name",
        type=str,
        help="index name to use when creating the AI Search index",
        default=os.environ["AISEARCH_INDEX_NAME"],
    )
    parser.add_argument(
        "--csv-file", type=str, help="path to data for creating search index", default="assets/products.csv"
    )
    args = parser.parse_args()
    index_name = args.index_name
    csv_file = args.csv_file

    create_index_from_csv(index_name, csv_file)

Desde la consola, inicie sesión en su cuenta de Azure y siga las instrucciones para autenticar su cuenta:
```
az login
```
Ejecute el código para compilar el índice localmente y registrarlo en el proyecto en la nube:
```
python create_search_index.py
```
Una vez ejecutado el script, puede ver el índice recién creado en la página Datos e índices del proyecto de Estudio de IA de Azure. Para obtener más información, consulte Cómo compilar y consumir índices vectoriales en Inteligencia artificial de Azure Studio.
Si vuelve a ejecutar el script con el mismo nombre de índice, crea una nueva versión del mismo índice.

Obtener documentos del producto

A continuación, creará un script para obtener los documentos del producto a partir del índice de búsqueda. El script consulta el índice de búsqueda de documentos que coinciden con la pregunta de un usuario.

Creación de un script para obtener los documentos del producto

Cuando el chat recibe una solicitud, busca en los datos para encontrar la información pertinente. Este script usa el SDK de Azure AI para consultar el índice de búsqueda de documentos que coinciden con la pregunta de un usuario. A continuación, devuelve los documentos a la aplicación de chat.

Cree el archivo get_product_documents.py en el directorio principal. Copie y pegue el código siguiente en el archivo .

Comience con el código para importar las bibliotecas necesarias, cree un cliente de proyecto y configure las opciones:

import os
from pathlib import Path
from opentelemetry import trace
from azure.ai.projects import AIProjectClient
from azure.ai.projects.models import ConnectionType
from azure.identity import DefaultAzureCredential
from azure.core.credentials import AzureKeyCredential
from azure.search.documents import SearchClient
from config import ASSET_PATH, get_logger

# initialize logging and tracing objects
logger = get_logger(__name__)
tracer = trace.get_tracer(__name__)

# create a project client using environment variables loaded from the .env file
project = AIProjectClient.from_connection_string(
    conn_str=os.environ["AIPROJECT_CONNECTION_STRING"], credential=DefaultAzureCredential()
)

# create a vector embeddings client that will be used to generate vector embeddings
chat = project.inference.get_chat_completions_client()
embeddings = project.inference.get_embeddings_client()

# use the project client to get the default search connection
search_connection = project.connections.get_default(
    connection_type=ConnectionType.AZURE_AI_SEARCH, include_credentials=True
)

# Create a search index client using the search connection
# This client will be used to create and delete search indexes
search_client = SearchClient(
    index_name=os.environ["AISEARCH_INDEX_NAME"],
    endpoint=search_connection.endpoint_url,
    credential=AzureKeyCredential(key=search_connection.key),
)

Agregue la función para obtener los documentos del producto:

from azure.ai.inference.prompts import PromptTemplate
from azure.search.documents.models import VectorizedQuery


@tracer.start_as_current_span(name="get_product_documents")
def get_product_documents(messages: list, context: dict = None) -> dict:
    if context is None:
        context = {}

    overrides = context.get("overrides", {})
    top = overrides.get("top", 5)

    # generate a search query from the chat messages
    intent_prompty = PromptTemplate.from_prompty(Path(ASSET_PATH) / "intent_mapping.prompty")

    intent_mapping_response = chat.complete(
        model=os.environ["INTENT_MAPPING_MODEL"],
        messages=intent_prompty.create_messages(conversation=messages),
        **intent_prompty.parameters,
    )

    search_query = intent_mapping_response.choices[0].message.content
    logger.debug(f"🧠 Intent mapping: {search_query}")

    # generate a vector representation of the search query
    embedding = embeddings.embed(model=os.environ["EMBEDDINGS_MODEL"], input=search_query)
    search_vector = embedding.data[0].embedding

    # search the index for products matching the search query
    vector_query = VectorizedQuery(vector=search_vector, k_nearest_neighbors=top, fields="contentVector")

    search_results = search_client.search(
        search_text=search_query, vector_queries=[vector_query], select=["id", "content", "filepath", "title", "url"]
    )

    documents = [
        {
            "id": result["id"],
            "content": result["content"],
            "filepath": result["filepath"],
            "title": result["title"],
            "url": result["url"],
        }
        for result in search_results
    ]

    # add results to the provided context
    if "thoughts" not in context:
        context["thoughts"] = []

    # add thoughts and documents to the context object so it can be returned to the caller
    context["thoughts"].append(
        {
            "title": "Generated search query",
            "description": search_query,
        }
    )

    if "grounding_data" not in context:
        context["grounding_data"] = []
    context["grounding_data"].append(documents)

    logger.debug(f"📄 {len(documents)} documents retrieved: {documents}")
    return documents

Por último, agregue el código para probar la función al ejecutar el script directamente:

if __name__ == "__main__":
    import logging
    import argparse

    # set logging level to debug when running this module directly
    logger.setLevel(logging.DEBUG)

    # load command line arguments
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--query",
        type=str,
        help="Query to use to search product",
        default="I need a new tent for 4 people, what would you recommend?",
    )

    args = parser.parse_args()
    query = args.query

    result = get_product_documents(messages=[{"role": "user", "content": query}])

Creación de una plantilla de indicación para la asignación de intenciones

El script get_product_documents.py usa una plantilla de indicación para convertir la conversación en una consulta de búsqueda. En la plantilla se indica cómo extraer la intención del usuario de la conversación.

Antes de ejecutar el script, cree la plantilla de indicación. Agregue el archivo intent_mapping.prompty a la carpeta recursos:

---
name: Chat Prompt
description: A prompty that extract users query intent based on the current_query and chat_history of the conversation
model:
    api: chat
    configuration:
        azure_deployment: gpt-4o
inputs:
    conversation:
        type: array
---
system:
# Instructions
- You are an AI assistant reading a current user query and chat_history.
- Given the chat_history, and current user's query, infer the user's intent expressed in the current user query.
- Once you infer the intent, respond with a search query that can be used to retrieve relevant documents for the current user's query based on the intent
- Be specific in what the user is asking about, but disregard parts of the chat history that are not relevant to the user's intent.
- Provide responses in json format

# Examples
Example 1:
With a conversation like below:
```
 - user: are the trailwalker shoes waterproof?
 - assistant: Yes, the TrailWalker Hiking Shoes are waterproof. They are designed with a durable and waterproof construction to withstand various terrains and weather conditions.
 - user: how much do they cost?
```
Respond with:
{
    "intent": "The user wants to know how much the Trailwalker Hiking Shoes cost.",
    "search_query": "price of Trailwalker Hiking Shoes"
}

Example 2:
With a conversation like below:
```
 - user: are the trailwalker shoes waterproof?
 - assistant: Yes, the TrailWalker Hiking Shoes are waterproof. They are designed with a durable and waterproof construction to withstand various terrains and weather conditions.
 - user: how much do they cost?
 - assistant: The TrailWalker Hiking Shoes are priced at $110.
 - user: do you have waterproof tents?
 - assistant: Yes, we have waterproof tents available. Can you please provide more information about the type or size of tent you are looking for?
 - user: which is your most waterproof tent?
 - assistant: Our most waterproof tent is the Alpine Explorer Tent. It is designed with a waterproof material and has a rainfly with a waterproof rating of 3000mm. This tent provides reliable protection against rain and moisture.
 - user: how much does it cost?
```
Respond with:
{
    "intent": "The user would like to know how much the Alpine Explorer Tent costs.",
    "search_query": "price of Alpine Explorer Tent"
}

user:
Return the search query for the messages in the following conversation:
{{#conversation}}
 - {{role}}: {{content}}
{{/conversation}}

Prueba del script de recuperación de documentos del producto

Ahora que tiene el script y la plantilla, ejecute el script para probar qué documentos devuelve el índice de búsqueda a partir de una consulta. En una ventana de terminal, ejecute lo siguiente:

python get_product_documents.py --query "I need a new tent for 4 people, what would you recommend?"

Desarrollo del código de recuperación de conocimiento personalizada (RAG)

A continuación, cree código personalizado para agregar funcionalidades de generación aumentada de recuperación (RAG) a una aplicación de chat básica.

Creación de un script de chat con funcionalidades de RAG

En la carpeta principal, cree un archivo denominado chat_with_products.py. Este script recupera los documentos del producto y genera una respuesta a la pregunta de un usuario.

Agregue el código para importar las bibliotecas necesarias, cree un cliente de proyecto y configure las opciones:

import os
from pathlib import Path
from opentelemetry import trace
from azure.ai.projects import AIProjectClient
from azure.identity import DefaultAzureCredential
from config import ASSET_PATH, get_logger, enable_telemetry
from get_product_documents import get_product_documents


# initialize logging and tracing objects
logger = get_logger(__name__)
tracer = trace.get_tracer(__name__)

# create a project client using environment variables loaded from the .env file
project = AIProjectClient.from_connection_string(
    conn_str=os.environ["AIPROJECT_CONNECTION_STRING"], credential=DefaultAzureCredential()
)

# create a chat client we can use for testing
chat = project.inference.get_chat_completions_client()

Cree la función de chat que usa las funcionalidades de RAG:

from azure.ai.inference.prompts import PromptTemplate


@tracer.start_as_current_span(name="chat_with_products")
def chat_with_products(messages: list, context: dict = None) -> dict:
    if context is None:
        context = {}

    documents = get_product_documents(messages, context)

    # do a grounded chat call using the search results
    grounded_chat_prompt = PromptTemplate.from_prompty(Path(ASSET_PATH) / "grounded_chat.prompty")

    system_message = grounded_chat_prompt.create_messages(documents=documents, context=context)
    response = chat.complete(
        model=os.environ["CHAT_MODEL"],
        messages=system_message + messages,
        **grounded_chat_prompt.parameters,
    )
    logger.info(f"💬 Response: {response.choices[0].message}")

    # Return a chat protocol compliant response
    return {"message": response.choices[0].message, "context": context}

Por último, agregue el código para ejecutar la función de chat:

if __name__ == "__main__":
    import argparse

    # load command line arguments
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--query",
        type=str,
        help="Query to use to search product",
        default="I need a new tent for 4 people, what would you recommend?",
    )
    parser.add_argument(
        "--enable-telemetry",
        action="store_true",
        help="Enable sending telemetry back to the project",
    )
    args = parser.parse_args()
    if args.enable_telemetry:
        enable_telemetry(True)

    # run chat with products
    response = chat_with_products(messages=[{"role": "user", "content": args.query}])

Creación de una plantilla de indicación de chat exhaustiva

El script chat_with_products.py llama a una plantilla de indicación para generar una respuesta a la pregunta del usuario. En la plantilla se indica cómo generar una respuesta basada en la pregunta del usuario y los documentos recuperados. Cree esta plantilla ahora.

En la carpeta recursos, agregue el archivo grounded_chat.prompty:

---
name: Chat with documents
description: Uses a chat completions model to respond to queries grounded in relevant documents
model:
    api: chat
    configuration:
        azure_deployment: gpt-4o
inputs:
    conversation:
        type: array
---
system:
You are an AI assistant helping users with queries related to outdoor outdooor/camping gear and clothing.
If the question is not related to outdoor/camping gear and clothing, just say 'Sorry, I only can answer queries related to outdoor/camping gear and clothing. So, how can I help?'
Don't try to make up any answers.
If the question is related to outdoor/camping gear and clothing but vague, ask for clarifying questions instead of referencing documents. If the question is general, for example it uses "it" or "they", ask the user to specify what product they are asking about.
Use the following pieces of context to answer the questions about outdoor/camping gear and clothing as completely, correctly, and concisely as possible.
Do not add documentation reference in the response.

# Documents

{{#documents}}

## Document {{id}}: {{title}}
{{content}}
{{/documents}}

Ejecución del script de chat con funcionalidades de RAG

Ahora que tiene el script y la plantilla, ejecute el script para probar la aplicación de chat con las funcionalidades de RAG:

python chat_with_products.py --query "I need a new tent for 4 people, what would you recommend?"

Para habilitar el registro de telemetría en el proyecto:

Instale azure-monitor-opentelemetry:

pip install azure-monitor-opentelemetry

Agregue la marca --enable-telemetry cuando use el script chat_with_products.py:

python chat_with_products.py --query "I need a new tent for 4 people, what would you recommend?" --enable-telemetry

Limpieza de recursos

Para evitar incurrir en costos innecesarios de Azure, debe eliminar los recursos que creó en este tutorial si ya no son necesarios. Para administrar recursos, puede usar Azure Portal.

Pero aún no los elimine, si quiere implementar su aplicación de chat en Azure en la siguiente parte de esta serie de tutoriales.

Paso siguiente

Parte 3: Evaluación e implementación de la aplicación de chat en Azure

Compartir a través de