星火大模型——多模态API调用（图像+文本）

本文主要是介绍星火大模型——多模态API调用（图像+文本），希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

参考文档

图片理解 API 文档

环境配置

pip3 install websocket-client

核心代码

import _thread as thread
import base64
import datetime
import hashlib
import hmac
import json
from urllib.parse import urlparse
import ssl
from datetime import datetime
from time import mktime
from urllib.parse import urlencode
from wsgiref.handlers import format_date_time
import websocket  # 使用websocket_client
from PIL import Image
import ioappid = "xxx"    #填写控制台中获取的 APPID 信息
api_secret = "xxxx"   #填写控制台中获取的 APISecret 信息
api_key ="xxxx"    #填写控制台中获取的 APIKey 信息
#  = open("images\PointAClouth.png",'rb').read()# Function to resize the image
def resize_image(image_path, size):image = Image.open(image_path)resized_image = image.resize(size)print(resized_image.size)# import pdb; pdb.set_trace()return resized_image# Function to encode the image
def encode_image(image):image_byte_array = io.BytesIO()image.save(image_byte_array, format='PNG')image_byte_array = image_byte_array.getvalue()return image_byte_array, base64.b64encode(image_byte_array).decode('utf-8')# Path to your image
image_path = "images/PointAClouth.png"# Getting the base64 string of the resized image
new_size = (1000, 1000)
resized_image = resize_image(image_path, new_size)
imagedata, base64_image = encode_image(resized_image)imageunderstanding_url = "wss://spark-api.cn-huabei-1.xf-yun.com/v2.1/image"#云端环境的服务地址
text =[{"role": "user", "content": base64_image, "content_type":"image"}]class Ws_Param(object):# 初始化def __init__(self, APPID, APIKey, APISecret, imageunderstanding_url):self.APPID = APPIDself.APIKey = APIKeyself.APISecret = APISecretself.host = urlparse(imageunderstanding_url).netlocself.path = urlparse(imageunderstanding_url).pathself.ImageUnderstanding_url = imageunderstanding_url# 生成urldef create_url(self):# 生成RFC1123格式的时间戳now = datetime.now()date = format_date_time(mktime(now.timetuple()))# 拼接字符串signature_origin = "host: " + self.host + "\n"signature_origin += "date: " + date + "\n"signature_origin += "GET " + self.path + " HTTP/1.1"# 进行hmac-sha256进行加密signature_sha = hmac.new(self.APISecret.encode('utf-8'), signature_origin.encode('utf-8'),digestmod=hashlib.sha256).digest()signature_sha_base64 = base64.b64encode(signature_sha).decode(encoding='utf-8')authorization_origin = f'api_key="{self.APIKey}", algorithm="hmac-sha256", headers="host date request-line", signature="{signature_sha_base64}"'authorization = base64.b64encode(authorization_origin.encode('utf-8')).decode(encoding='utf-8')# 将请求的鉴权参数组合为字典v = {"authorization": authorization,"date": date,"host": self.host}# 拼接鉴权参数，生成urlurl = self.ImageUnderstanding_url + '?' + urlencode(v)#print(url)# 此处打印出建立连接时候的url,参考本demo的时候可取消上方打印的注释，比对相同参数时生成的url与自己代码生成的url是否一致return url# 收到websocket错误的处理
def on_error(ws, error):print("### error:", error)# 收到websocket关闭的处理
def on_close(ws,one,two):print(" ")# 收到websocket连接建立的处理
def on_open(ws):thread.start_new_thread(run, (ws,))def run(ws, *args):data = json.dumps(gen_params(appid=ws.appid, question= ws.question ))ws.send(data)# 收到websocket消息的处理
def on_message(ws, message):#print(message)data = json.loads(message)code = data['header']['code']if code != 0:print(f'请求错误: {code}, {data}')ws.close()else:# print("text_token:", payload.usage.text.prompt_tokens)choices = data["payload"]["choices"]status = choices["status"]content = choices["text"][0]["content"]print(content,end ="")# print(usage)global answeranswer += content# print(1)if status == 2:# import pdb; pdb.set_trace()print("usage:", data["payload"]['usage'])ws.close()def gen_params(appid, question):"""通过appid和用户的提问来生成请参数"""data = {"header": {"app_id": appid},"parameter": {"chat": {"domain": "image","temperature": 0.5,"top_k": 4,"max_tokens": 2028,"auditing": "default"}},"payload": {"message": {"text": question}}
}return datadef main(appid, api_key, api_secret, imageunderstanding_url,imagedata,question):wsParam = Ws_Param(appid, api_key, api_secret, imageunderstanding_url)websocket.enableTrace(False)wsUrl = wsParam.create_url()ws = websocket.WebSocketApp(wsUrl, on_message=on_message, on_error=on_error, on_close=on_close, on_open=on_open)ws.appid = appidws.imagedata = imagedataws.question = questionws.run_forever(sslopt={"cert_reqs": ssl.CERT_NONE})def getText(role, content):jsoncon = {}jsoncon["role"] = rolejsoncon["content"] = contenttext.append(jsoncon)return textdef getlength(text):length = 0for content in text:temp = content["content"]leng = len(temp)length += lengreturn lengthdef checklen(text):#print("text-content-tokens:", getlength(text[1:]))while (getlength(text[1:])> 8000):del text[1]return textif __name__ == '__main__':text.clear# while(1):# Input = input("\n" +"问:")getText("system","xxxx")Input = "xxxxx"question = checklen(getText("user",Input))answer = ""print("答:",end = "")main(appid, api_key, api_secret, imageunderstanding_url, imagedata,question)getText("assistant", answer)print(text[1:])# print(str(text))