Amazon Textract
Amazon Textract 是一种机器学习 (ML) 服务，可自动从扫描文档中提取文本、手写和数据。
它超越了简单的光学字符识别 (OCR)，可以识别、理解和提取表单和表格中的数据。如今，许多公司手动从扫描文档（如 PDF、图像、表格和表单）中提取数据，或者通过需要手动配置的简单 OCR 软件提取数据（当表单更改时，通常必须更新配置）。为了克服这些手动且昂贵的流程，
Textract使用机器学习来读取和处理任何类型的文档，准确提取文本、手写、表格和其他数据，无需人工干预。
此示例演示了 Amazon Textract 与 LangChain 结合用作 DocumentLoader 的用法。
Textract 支持 PDF、TIFF、PNG 和 JPEG 格式。
Textract 支持这些 文档大小、语言和字符。
%pip install --upgrade --quiet boto3 langchain-openai tiktoken python-dotenv
%pip install --upgrade --quiet "amazon-textract-caller>=0.2.0"
示例 1
第一个示例使用本地文件，该文件将在内部发送到 Amazon Textract 同步 API DetectDocumentText。
对于 Textract，本地文件或像 HTTP:// 这样的 URL 端点仅限于单页文档。多页文档必须驻留在 S3 上。此示例文件是 jpeg。
from langchain_community.document_loaders import AmazonTextractPDFLoader
loader = AmazonTextractPDFLoader("example_data/alejandro_rosalez_sample-small.jpeg")
documents = loader.load()
API 参考：AmazonTextractPDFLoader
来自文件的输出
documents
[Document(page_content='Patient Information First Name: ALEJANDRO Last Name: ROSALEZ Date of Birth: 10/10/1982 Sex: M Marital Status: MARRIED Email Address: Address: 123 ANY STREET City: ANYTOWN State: CA Zip Code: 12345 Phone: 646-555-0111 Emergency Contact 1: First Name: CARLOS Last Name: SALAZAR Phone: 212-555-0150 Relationship to Patient: BROTHER Emergency Contact 2: First Name: JANE Last Name: DOE Phone: 650-555-0123 Relationship FRIEND to Patient: Did you feel fever or feverish lately? Yes No Are you having shortness of breath? Yes No Do you have a cough? Yes No Did you experience loss of taste or smell? Yes No Where you in contact with any confirmed COVID-19 positive patients? Yes No Did you travel in the past 14 days to any regions affected by COVID-19? Yes No Patient Information First Name: ALEJANDRO Last Name: ROSALEZ Date of Birth: 10/10/1982 Sex: M Marital Status: MARRIED Email Address: Address: 123 ANY STREET City: ANYTOWN State: CA Zip Code: 12345 Phone: 646-555-0111 Emergency Contact 1: First Name: CARLOS Last Name: SALAZAR Phone: 212-555-0150 Relationship to Patient: BROTHER Emergency Contact 2: First Name: JANE Last Name: DOE Phone: 650-555-0123 Relationship FRIEND to Patient: Did you feel fever or feverish lately? Yes No Are you having shortness of breath? Yes No Do you have a cough? Yes No Did you experience loss of taste or smell? Yes No Where you in contact with any confirmed COVID-19 positive patients? Yes No Did you travel in the past 14 days to any regions affected by COVID-19? Yes No ', metadata={'source': 'example_data/alejandro_rosalez_sample-small.jpeg', 'page': 1})]
示例 2
下一个示例从 HTTPS 端点加载文件。它必须是单页的，因为 Amazon Textract 要求所有多页文档都存储在 S3 上。
from langchain_community.document_loaders import AmazonTextractPDFLoader
loader = AmazonTextractPDFLoader(
"https://amazon-textract-public-content.s3.us-east-2.amazonaws.com/langchain/alejandro_rosalez_sample_1.jpg"
)
documents = loader.load()
API 参考：AmazonTextractPDFLoader
documents
[Document(page_content='Patient Information First Name: ALEJANDRO Last Name: ROSALEZ Date of Birth: 10/10/1982 Sex: M Marital Status: MARRIED Email Address: Address: 123 ANY STREET City: ANYTOWN State: CA Zip Code: 12345 Phone: 646-555-0111 Emergency Contact 1: First Name: CARLOS Last Name: SALAZAR Phone: 212-555-0150 Relationship to Patient: BROTHER Emergency Contact 2: First Name: JANE Last Name: DOE Phone: 650-555-0123 Relationship FRIEND to Patient: Did you feel fever or feverish lately? Yes No Are you having shortness of breath? Yes No Do you have a cough? Yes No Did you experience loss of taste or smell? Yes No Where you in contact with any confirmed COVID-19 positive patients? Yes No Did you travel in the past 14 days to any regions affected by COVID-19? Yes No Patient Information First Name: ALEJANDRO Last Name: ROSALEZ Date of Birth: 10/10/1982 Sex: M Marital Status: MARRIED Email Address: Address: 123 ANY STREET City: ANYTOWN State: CA Zip Code: 12345 Phone: 646-555-0111 Emergency Contact 1: First Name: CARLOS Last Name: SALAZAR Phone: 212-555-0150 Relationship to Patient: BROTHER Emergency Contact 2: First Name: JANE Last Name: DOE Phone: 650-555-0123 Relationship FRIEND to Patient: Did you feel fever or feverish lately? Yes No Are you having shortness of breath? Yes No Do you have a cough? Yes No Did you experience loss of taste or smell? Yes No Where you in contact with any confirmed COVID-19 positive patients? Yes No Did you travel in the past 14 days to any regions affected by COVID-19? Yes No ', metadata={'source': 'example_data/alejandro_rosalez_sample-small.jpeg', 'page': 1})]
示例 3
处理多页文档需要该文档位于 S3 上。示例文档位于 us-east-2 的存储桶中，并且需要在同一区域调用 Textract 才能成功，因此我们在客户端上设置 region_name 并将其传递给加载器，以确保从 us-east-2 调用 Textract。您也可以在 us-east-2 中运行您的笔记本，将 AWS_DEFAULT_REGION 设置为 us-east-2，或者在不同的环境中运行时，传递一个带有该区域名称的 boto3 Textract 客户端，如下面的单元格所示。
import boto3
textract_client = boto3.client("textract", region_name="us-east-2")
file_path = "s3://amazon-textract-public-content/langchain/layout-parser-paper.pdf"
loader = AmazonTextractPDFLoader(file_path, client=textract_client)
documents = loader.load()
现在获取页数以验证响应（打印完整响应会很长……）。我们预期有 16 页。
len(documents)
16
示例 4
您可以选择将一个名为 linearization_config 的附加参数传递给 AmazonTextractPDFLoader，这将决定 Textract 运行后解析器如何线性化文本输出。
from langchain_community.document_loaders import AmazonTextractPDFLoader
from textractor.data.text_linearization_config import TextLinearizationConfig
loader = AmazonTextractPDFLoader(
"s3://amazon-textract-public-content/langchain/layout-parser-paper.pdf",
linearization_config=TextLinearizationConfig(
hide_header_layout=True,
hide_footer_layout=True,
hide_figure_layout=True,
),
)
documents = loader.load()
API 参考：AmazonTextractPDFLoader
在 LangChain 链中使用 AmazonTextractPDFLoader（例如 OpenAI）
AmazonTextractPDFLoader 可以像使用其他加载器一样在链中使用。Textract 本身确实具有 查询功能，该功能与此示例中的 QA 链提供类似的功能，也值得检查。
# You can store your OPENAI_API_KEY in a .env file as well
# import os
# from dotenv import load_dotenv
# load_dotenv()
# Or set the OpenAI key in the environment directly
import os
os.environ["OPENAI_API_KEY"] = "your-OpenAI-API-key"
from langchain.chains.question_answering import load_qa_chain
from langchain_openai import OpenAI
chain = load_qa_chain(llm=OpenAI(), chain_type="map_reduce")
query = ["Who are the autors?"]
chain.run(input_documents=documents, question=query)
API 参考：load_qa_chain | OpenAI
' The authors are Zejiang Shen, Ruochen Zhang, Melissa Dell, Benjamin Charles Germain Lee, Jacob Carlson, Weining Li, Gardner, M., Grus, J., Neumann, M., Tafjord, O., Dasigi, P., Liu, N., Peters, M., Schmitz, M., Zettlemoyer, L., Lukasz Garncarek, Powalski, R., Stanislawek, T., Topolski, B., Halama, P., Gralinski, F., Graves, A., Fernández, S., Gomez, F., Schmidhuber, J., Harley, A.W., Ufkes, A., Derpanis, K.G., He, K., Gkioxari, G., Dollár, P., Girshick, R., He, K., Zhang, X., Ren, S., Sun, J., Kay, A., Lamiroy, B., Lopresti, D., Mears, J., Jakeway, E., Ferriter, M., Adams, C., Yarasavage, N., Thomas, D., Zwaard, K., Li, M., Cui, L., Huang,'