Ragas

Iniciação rápida

Conceito

概念	含义
Question
Contexts	Retrieved contexts：实际找到的Context
Answer	最终生成的答案
Ground truths	参考答案

Datasets

from datasets import Dataset 

data_samples = {
    'question': ['When was the first super bowl?', 'Who won the most super bowls?'],
    'answer': ['The first superbowl was held on January 15, 1967', 'The most super bowls have been won by The New England Patriots'],
    'contexts' : [['The Super Bowl....season since 1966,','replacing the NFL...in February.'], 
    ['The Green Bay Packers...Green Bay, Wisconsin.','The Packers compete...Football Conference']],
    'ground_truth': ['The first superbowl was held on January 15, 1967', 'The New England Patriots have won the Super Bowl a record six times']
}

dataset = Dataset.from_dict(data_samples)

Métrica

官方文档

Metric
Context Precision	Retrieval	Question	是否跑题：检索结果 与 Quesion 是否相关
Answer Relevance	Anwser	Question	是否跑题：生成的答案 是否与 Question 相关
Faithfulness	Anwser	Retrieval	是否参考引用：生成的答案 是否忠诚于 检索结果
Context Recall	Retrieval	参考答案 Ground Truth	检索的准确性： 检索结果 与 参考答案 是否相关

Pronto

Contexto de precisão

Given question, answer and context verify if the context was useful in arriving at the given answer. Give verdict as "1" if useful and "0" if not with json output.

The output should be a well-formatted JSON instance that conforms to the JSON schema below.

……

Your actual task:

question: 法国的首都是什么？
context: 巴黎是法国的首都。
answer: 巴黎
verification: 

Resposta Relevância

Não entendo o prompt.

Generate a question for the given answer and Identify if answer is noncommittal. Give noncommittal as 1 if the answer is noncommittal and 0 if the answer is committal. A noncommittal answer is one that is evasive, vague, or ambiguous. For example, "I don't know" or "I'm not sure" are noncommittal answers

……

Your actual task:

answer: 巴黎
context: 巴黎是法国的首都。
output: 

fidelidade

Create one or more statements from each sentence in the given answer.

……

Your actual task:

question: 法国的首都是什么？
answer: 巴黎
statements: 

Your task is to judge the faithfulness of a series of statements based on a given context. For each statement you must return verdict as 1 if the statement can be verified based on the context or 0 if the statement can not be verified based on the context.

……

Your actual task:

context: 巴黎是法国的首都。
statements: ["\u6cd5\u56fd\u7684\u9996\u90fd\u662f\u5df4\u9ece\u3002"]
answer: 

Contexto Recall

Given a context, and an answer, analyze each sentence in the answer and classify if the sentence can be attributed to the given context or not. Use only "Yes" (1) or "No" (0) as a binary classification. Output json with reason.

……

Your actual task:

question: 法国的首都是什么？
context: 巴黎是法国的首都。
answer: 巴黎
classification: 

Geração de dados sintéticos

Motivação: Criação manual de centenas de amostras de QA (pergunta-contexto - resposta) a partir de um documento pode ser demorada e laboriosa. Geração automática usando o LLM.

方案：Evol-Instruct

Dividido em várias categorias: simples, * * raciocínio, condicionamento, multi-contexto**

As categorias são chamadas de *

Especifique as proporções para essas três categorias quando você gera.

from ragas.testset.generator import TestsetGenerator
from ragas.testset.evolutions import simple, reasoning, multi_context
from langchain_openai import ChatOpenAI, OpenAIEmbeddings

# documents = load your documents

# generator with openai models
generator_llm = ChatOpenAI(model="gpt-3.5-turbo-16k")
critic_llm = ChatOpenAI(model="gpt-4")
embeddings = OpenAIEmbeddings()

generator = TestsetGenerator.from_langchain(
    generator_llm,
    critic_llm,
    embeddings
)

# Change resulting question type distribution
distributions = {
    simple: 0.5,
    multi_context: 0.4,
    reasoning: 0.1
}

# use generator.generate_with_llamaindex_docs if you use llama-index as document loader
testset = generator.generate_with_langchain_docs(documents, 10, distributions) 
testset.to_pandas()

Dados de leitura

Não é fácil ter problemas com o uso oficial da LangChain e continuar usando a LangChain.

Adaptação automática de linguagem

官网

Avaliações

O Prompt usado durante a avaliação foi traduzido para o chinês usando o modelo gpt - 4 - turbo-preview; armazenado em cache localmente.

As solicitações pertencentes às respectivas métricas serão agora adaptadas automaticamente ao idioma de destino.

A etapa de salva salva o arquivo para . cacha / ragas por padrão para reutilizar mais tarde.

# 将Metric中的Prompt翻译成中文
from datasets import Dataset
# from langchain.chat_models import ChatOpenAI
from langchain_openai import ChatOpenAI, OpenAI

from ragas.metrics import (
    answer_relevancy,
    faithfulness,
    context_recall,
    context_precision,
    answer_correctness,
    answer_similarity,
)
from ragas import evaluate
from ragas import adapt

eval_model = ChatOpenAI(model="gpt-3.5-turbo", temperature=0)

# llm used for adaptation
openai_model = ChatOpenAI(model_name="gpt-4-turbo-preview")
# openai_model = OpenAI(model_name="gpt-4-0125-preview", temperature=0)

adapt(
    metrics=[
        answer_relevancy,
        # faithfulness,
        context_recall,
        context_precision,
        
        answer_correctness,
        # answer_similarity,
    ],
    language="Chinese",
    llm=openai_model,
)

# Eval
dataset = Dataset.from_dict(
    {
        "question": ["法国的首都是什么？"],
        "contexts": [["巴黎是法国的首都。"]],
        "answer": ["巴黎"],
        "ground_truths": [["巴黎"]],
    }
)
print(dataset)

results = evaluate(dataset, llm=eval_model)
print(results)

Dados cachés gerados

Geração de dados sintéticos

from ragas.testset.generator import TestsetGenerator
from ragas.testset.evolutions import simple, reasoning, multi_context,conditional
from langchain_openai import ChatOpenAI, OpenAIEmbeddings

# generator with openai models
generator_llm = ChatOpenAI(model="gpt-3.5-turbo-16k")
critic_llm = ChatOpenAI(model="gpt-4")
embeddings = OpenAIEmbeddings()

generator = TestsetGenerator.from_langchain(
    generator_llm,
    critic_llm,
    embeddings
)

# adapt to language
language = "Chinese"

generator.adapt(language, evolutions=[simple, reasoning,conditional,multi_context])
generator.save(evolutions=[simple, reasoning, multi_context,conditional])

Dados cachés gerados

Campos em um dataset

• Pergunta
• contexts: contexto recuperado
• ground _ truth: resposta de referência
• Anwser: Resposta Gerada

1. Question: A set of questions.
2. Contexts: Retrieved contexts corresponding to each question. This is a list[list] since each question can retrieve multiple text chunks.
3. Answer: Generated answer corresponding to each question.
4. Verdades fundamentais: Verdades fundamentais correspondentes a cada questão. Esta é uma string que corresponde à resposta esperada para cada pergunta.

Contexto de referência não usado temporariamente

        metrics=[
            # 这几个都不需要原始的context
            context_precision,
            answer_relevancy,
            faithfulness,
            context_recall,
        ],

Usando o BGE

from ragas.llama_index import evaluate

flag_model = HuggingFaceEmbeddings(model_name="BAAI/bge-small-en-v1.5")
query_engine2 = build_query_engine(flag_model)
result = evaluate(query_engine2, metrics, test_questions, test_answers)

Integração LangSmith

Definir as variáveis de ambiente

export LANGCHAIN_TRACING_V2=true
export LANGCHAIN_ENDPOINT=https://api.smith.langchain.com
export LANGCHAIN_API_KEY=<your-api-key>
export LANGCHAIN_PROJECT=<your-project>  # if not specified, defaults to "default"

Criar um rastreador

# langsmith
from langchain.callbacks.tracers import LangChainTracer

tracer = LangChainTracer(project_name="callback-experiments")

Usado durante a avaliação

from datasets import load_dataset
from ragas.metrics import context_precision
from ragas import evaluate

dataset = load_dataset("explodinggradients/amnesty_qa","english")
evaluate(dataset["train"],metrics=[context_precision],callbacks=[tracer])

Integração do LlamaIndex

官方文档有问题。