Ragas

Быстрый старт

Концепции

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

Набор данных

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)

Metric

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

По запросу

Контекстная точность

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: 

Ответы на релевантность

Промп не понимает.

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: 

Верность

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: 

Контекстный вызов

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: 

Создание синтезированных данных

Мотивация: создание сотен образцов QA (вопрос-контекст-ответ) из документа вручную может быть трудоемким и трудоемким.Используйте LLM для автоматического создания.

方案：Evol-Instruct

Классификация: Simple, * * reasoning, conditioning, multi-context * *

Категория называется: * * эволюции * *

При формировании укажите пропорции для этих трех категорий.

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()

Читать данные

Официальное использование LangChain, продолжение использования LangChain, нелегко возникнуть проблемы.

Автоматическая языковая адаптация

官网

Оценки

Переведите запрос, использованный в процессе оценки, на китайский язык с использованием модели gpt-4-turbo-preview; кэшируйте локально.

Подсказки, относящиеся к соответствующим метрикам, теперь автоматически адаптируются к целевому языку.

Сохранение сохраняет его в файле «. cacha / ragas » по умолчанию для повторного использования позже.

# 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)

Созданные данные в кэши

Создание синтезированных данных

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])

Созданные данные в кэши

поля в наборе данных

• вопрос
• contexts: извлеченный контекст
• ground_truth: справочные ответы
• anwser: сгенерированные ответы

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. Основные истины: Основные истины, соответствующие каждому вопросу.Это строка, которая соответствует ожидаемому ответу на каждый вопрос.

Ссылка на контекст не используется

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

Использование 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)

Интеграция LangSmith

Настройка переменной среды

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"

Создание трекеров

# langsmith
from langchain.callbacks.tracers import LangChainTracer

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

Использование при оценке

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])

Интеграция LlamaIndex

官方文档有问题。