使用 Amazon Bedrock 中的 Cohere Embed 4 多模态嵌入模型驱动企业搜索

import boto3
import json # 初始化 Bedrock Runtime 客户端
bedrock_runtime = boto3.client('bedrock-runtime', region_name='us-east-1') # 请求体
body = json.dumps({
"texts": [
text1, text2], "input_type":"search_document", "embedding_types": ["float"]
})
# 调用模型
model_id = 'cohere.embed-v4:0'
response = bedrock_runtime.invoke_model(
    modelId=model_id,
    body=json.dumps(body),
    accept= '*/*',
    contentType='application/json'
)
# 解析响应
result = json.loads(response['body'].read())

import base64 # 初始化 Bedrock Runtime 客户端
bedrock_runtime = boto3.client('bedrock-runtime', region_name='us-east-1') # 请求体
body = json.dumps({
"inputs": [
{
"content": [
{ "type": "text", "text": text },
{ "type": "image_url", {"image_url":image_base64_uri}}
]
}
],
"input_type":"search_document",
"embedding_types": ["int8","float"]
})
# 调用模型
model_id = 'cohere.embed-v4:0'
response = bedrock_runtime.invoke_model(
    modelId=model_id,
    body=json.dumps(body),
    accept= '*/*',
    contentType='application/json'
)
# 解析响应
result = json.loads(response['body'].read())

# S3 Vector 搜索金融文档的函数
@tool
def search(query_text: str, bucket_name: str = "my-s3-vector-bucket", index_name: str = "my-s3-vector-index-1536", top_k: int = 3, category_filter: str = None) -> str: """使用语义向量搜索金融文档""" bedrock = boto3.client("bedrock-runtime", region_name="us-east-1") s3vectors = boto3.client("s3vectors", region_name="us-east-1") # 使用 Cohere Embed v4 生成嵌入
response = bedrock.invoke_model(
    modelId="cohere.embed-v4:0",
    body=json.dumps({
        "texts": [query_text],
        "input_type": "search_query",
        "embedding_types": ["float"]
    }),
    accept='*/*',
    contentType='application/json'
)
response_body = json.loads(response["body"].read())
embedding = response_body["embeddings"]["float"][0]
# 查询向量
query_params = {
    "vectorBucketName": bucket_name,
    "indexName": index_name,
    "queryVector": {"float32": embedding},
    "topK": top_k,
    "returnDistance": True,
    "returnMetadata": True
}
if category_filter:
    query_params["filter"] = {"category": category_filter}
response = s3vectors.query_vectors(**query_params)
return json.dumps(response["vectors"], indent=2)

# 使用 Strands 创建金融研究代理
agent = Agent(
    name="FinancialResearchAgent",
    system_prompt="You are a financial research assistant that can search through financial documents, earnings reports, regulatory filings, and market analysis. Use the search tool to find relevant financial information and provide helpful analysis.",
    tools=[search]
)

result = search("Compare earnings growth rates mentioned in the documents")
print(result)
{
    "key": "doc_0_en",
    "metadata": {
        "language": "en",
        "source_text": "Q3 2024 earnings report shows revenue growth of 15% year-over-year driven by strong performance in cloud services and AI products",
        "doc_id": 0
    },
    "distance": 0.7292724251747131
},
{
    "key": "doc_18_zh",
    "metadata": {
        "source_text": "2024年上半年财务报告显示净利润增长20%，主要受益于金融科技和数字支付业务的快速发展",
        "language": "zh",
        "doc_id": 18
    },
    "distance": 0.7704182863235474
},
{
    "key": "doc_12_de",
    "metadata": {
        "language": "de",
        "doc_id": 12,
        "source_text": "Halbjahresbericht 2024 verzeichnet 12% Gewinnsteigerung durch erfolgreiche Digitalisierung und Automatisierungslösungen"
    },
    "distance": 0.7738275527954102
}
]

response = agent("Compare earnings growth rates mentioned in the documents")

Agent Query:
I'll search for information about earnings growth rates in the financial documents to help you compare them.
Tool #1: search
Let me search for more specific earnings growth information and financial performance metrics.
Tool #2: search
Let me search for more comprehensive financial performance data and specific earnings metrics.
Tool #3: search
Based on the financial documents I've searched through, here's a comparison of the earnings growth rates mentioned:
...
## Key Insights: 1. **Growth Range:** The earnings growth rates span from 12% to 20%, indicating generally healthy performance across different markets and sectors. 2. **Technology Focus:** All companies showing strong growth are heavily invested in technology sectors (fintech, AI, cloud services, cybersecurity, automation). 3. **Geographic Diversity:** The strong performers represent different regions (Asia, Europe, North America), suggesting broad-based growth in tech-enabled services. 4. **Growth Sustainability:** The Chinese fintech company leads with 20% net profit growth, while the others show strong revenue growth in the 12-18% range. The data suggests that companies with strong technology components, particularly in emerging areas like AI, fintech, and cybersecurity, are experiencing the most robust earnings growth rates in 2024.Based on the financial documents I've searched through, here's a comparison of the earnings growth rates mentioned:
## Earnings Growth Rate Comparison The data suggests that companies with strong technology components, particularly in emerging areas like AI, fintech, and cybersecurity, are experiencing the most robust earnings growth rates in 2024.

# 使用 bedrock-agentcore<=0.1.5 和 bedrock-agentcore-starter-toolkit==0.1.14
from bedrock_agentcore_starter_toolkit import Runtime
from boto3.session import Session
boto_session = Session()
region = boto_session.region_name
agentcore_runtime = Runtime()
agent_name = "search_agent"
response = agentcore_runtime.configure(
    entrypoint="example.py", # 替换为您自定义的代理和工具
    auto_create_execution_role=True,
    auto_create_ecr=True,
    requirements_file="requirements.txt",
    region=region,
    agent_name=agent_name
)
response
launch_result = agentcore_runtime.launch()
invoke_response = agentcore_runtime.invoke({"prompt": "Compare earnings growth rates mentioned in the documents"})