Paso 4: Memoria y persistencia

De forma predeterminada, los agentes almacenarán el historial de chat en un InMemoryChatHistoryProvider servicio de INTELIGENCIA artificial subyacente o en función de lo que requiera el servicio subyacente.

El agente siguiente usa la finalización del chat de OpenAI, que no admite ni requiere almacenamiento del historial de chat en el servicio, por lo que, por lo tanto, crea y usa automáticamente un InMemoryChatHistoryProvider.

using System;
using Azure.AI.Projects;
using Azure.Identity;
using Microsoft.Agents.AI;

var endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
    ?? throw new InvalidOperationException("Set AZURE_OPENAI_ENDPOINT");
var deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT_NAME") ?? "gpt-4o-mini";

AIAgent agent = new AIProjectClient(new Uri(endpoint), new DefaultAzureCredential())
    .AsAIAgent(
        model: deploymentName,
        instructions: "You are a friendly assistant. Keep your answers brief.",
        name: "MemoryAgent");

Para usar un elemento personalizado ChatHistoryProvider, se puede pasar uno a las opciones del agente.

using System;
using Azure.AI.Projects;
using Azure.Identity;
using Microsoft.Agents.AI;

var endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
    ?? throw new InvalidOperationException("Set AZURE_OPENAI_ENDPOINT");
var deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT_NAME") ?? "gpt-4o-mini";

AIAgent agent = new AIProjectClient(new Uri(endpoint), new DefaultAzureCredential())
    .AsAIAgent(model: deploymentName, options: new ChatClientAgentOptions()
    {
        ChatOptions = new() { Instructions = "You are a helpful assistant." },
        ChatHistoryProvider = new CustomChatHistoryProvider()
    });

Use una sesión para compartir contexto entre ejecuciones:

AgentSession session = await agent.CreateSessionAsync();

Console.WriteLine(await agent.RunAsync("Hello! What's the square root of 9?", session));
Console.WriteLine(await agent.RunAsync("My name is Alice", session));
Console.WriteLine(await agent.RunAsync("What is my name?", session));

Defina un proveedor de contexto que almacene información de usuario en estado de sesión e inserte instrucciones de personalización:

class UserMemoryProvider(ContextProvider):
    """A context provider that remembers user info in session state."""

    DEFAULT_SOURCE_ID = "user_memory"

    def __init__(self):
        super().__init__(self.DEFAULT_SOURCE_ID)

    async def before_run(
        self,
        *,
        agent: Any,
        session: AgentSession | None,
        context: SessionContext,
        state: dict[str, Any],
    ) -> None:
        """Inject personalization instructions based on stored user info."""
        user_name = state.get("user_name")
        if user_name:
            context.extend_instructions(
                self.source_id,
                f"The user's name is {user_name}. Always address them by name.",
            )
        else:
            context.extend_instructions(
                self.source_id,
                "You don't know the user's name yet. Ask for it politely.",
            )

    async def after_run(
        self,
        *,
        agent: Any,
        session: AgentSession | None,
        context: SessionContext,
        state: dict[str, Any],
    ) -> None:
        """Extract and store user info in session state after each call."""
        for msg in context.input_messages:
            text = msg.text if hasattr(msg, "text") else ""
            if isinstance(text, str) and "my name is" in text.lower():
                state["user_name"] = text.lower().split("my name is")[-1].strip().split()[0].capitalize()

Cree un agente con el proveedor de contexto:

client = FoundryChatClient(
    project_endpoint="https://your-project.services.ai.azure.com",
    model="gpt-4o",
    credential=AzureCliCredential(),
)

agent = Agent(
    client=client,
    name="MemoryAgent",
    instructions="You are a friendly assistant.",
    context_providers=[UserMemoryProvider()],
)

Ejecútelo: el agente ahora tiene acceso al contexto:

session = agent.create_session()

# The provider doesn't know the user yet — it will ask for a name
result = await agent.run("Hello! What's the square root of 9?", session=session)
print(f"Agent: {result}\n")

# Now provide the name — the provider stores it in session state
result = await agent.run("My name is Alice", session=session)
print(f"Agent: {result}\n")

# Subsequent calls are personalized — name persists via session state
result = await agent.run("What is 2 + 2?", session=session)
print(f"Agent: {result}\n")

# Inspect session state to see what the provider stored
provider_state = session.state.get("user_memory", {})
print(f"[Session State] Stored user name: {provider_state.get('user_name')}")

from agent_framework import InMemoryHistoryProvider
from agent_framework.mem0 import Mem0ContextProvider

memory_store = InMemoryHistoryProvider(load_messages=True) # add local history for a reused or serialized session
agent_memory = Mem0ContextProvider("user-memory", api_key=..., agent_id="my-agent")  # add Mem0 provider for agent memory
audit_store = InMemoryHistoryProvider(
    "audit",
    load_messages=False,
    store_context_messages=True,  # include context added by other providers
)

agent = client.as_agent(
    name="MemoryAgent",
    instructions="You are a friendly assistant.",
    context_providers=[memory_store, agent_memory, audit_store],  # audit store last
)

De forma predeterminada, los agentes usan el historial en memoria local o el historial administrado por el servicio en función del proveedor y la sesión.

El siguiente agente de Foundry usa una implementación de modelo respaldada por proyectos. Agregue un proveedor de contexto cuando desee un estado de personalización o memoria específico de la aplicación más allá del historial de conversaciones.

a := foundryprovider.NewAgent(
    endpoint,
    token,
    foundryprovider.ModelDeployment(model),
    foundryprovider.AgentConfig{
        Instructions: "You are a friendly assistant. Keep your answers brief.",
        Config: agent.Config{
            Name: "MemoryAgent",
        },
    },
)

Defina un proveedor de contexto que almacene información de usuario en estado de sesión e inserte instrucciones de personalización:

import (
    "context"
    "fmt"
    "strings"

    "github.com/microsoft/agent-framework-go/agent"
    "github.com/microsoft/agent-framework-go/message"
)

const userMemorySourceID = "user_memory"

type providerState struct {
    UserName string `json:"user_name,omitempty"`
}

func newUserMemoryProvider() agent.ContextProvider {
    return agent.NewContextProvider(agent.ContextProviderConfig{
        SourceID: userMemorySourceID,
        Provide:  provideUserMemory,
        Store:    storeUserMemory,
    })
}

func provideUserMemory(ctx context.Context, invoking agent.InvokingContext) ([]*message.Message, []agent.Option, error) {
    session, _ := agent.GetOption(invoking.Options, agent.WithSession)
    var state providerState
    _, _ = session.Get(userMemorySourceID, &state)

    instructions := "You don't know the user's name yet. Ask for it politely."
    if state.UserName != "" {
        instructions = fmt.Sprintf("The user's name is %s. Always address them by name.", state.UserName)
    }
    return nil, []agent.Option{agent.WithInstructions(instructions)}, nil
}

func storeUserMemory(ctx context.Context, invoked agent.InvokedContext) error {
    session, _ := agent.GetOption(invoked.Options, agent.WithSession)
    var state providerState
    _, _ = session.Get(userMemorySourceID, &state)
    for _, msg := range invoked.RequestMessages {
        text := strings.TrimSpace(msg.Contents.Text())
        lower := strings.ToLower(text)
        if idx := strings.Index(lower, "my name is"); idx >= 0 {
            parts := strings.Fields(text[idx+len("my name is"):])
            if len(parts) == 0 {
                continue
            }
            state.UserName = strings.Trim(parts[0], ".,!?")
            session.Set(userMemorySourceID, state)
            break
        }
    }
    return nil
}

Cree un agente con el proveedor de contexto:

a := foundryprovider.NewAgent(
    endpoint,
    token,
    foundryprovider.ModelDeployment(model),
    foundryprovider.AgentConfig{
        Instructions: "You are a friendly assistant.",
        Config: agent.Config{
            Name:             "MemoryAgent",
            ContextProviders: []agent.ContextProvider{newUserMemoryProvider()},
        },
    },
)

Ejecútelo: el agente ahora tiene acceso al contexto:

ctx := context.Background()
session, err := a.CreateSession(ctx)
if err != nil {
    panic(err)
}

// The provider doesn't know the user yet.
resp, err := a.RunText(ctx, "Hello, what is the square root of 9?", agent.WithSession(session)).Collect()
fmt.Println(resp, err)

// Teach the provider the user's name.
resp, err = a.RunText(ctx, "My name is Alice", agent.WithSession(session)).Collect()
fmt.Println(resp, err)

// Subsequent calls are personalized using session state.
resp, err = a.RunText(ctx, "What is 2 + 2?", agent.WithSession(session)).Collect()
fmt.Println(resp, err)