Commit cb377359 by 陈正乐

项目文件整理

parent b0dfbbcf
import os
from typing import Dict, Optional, List
from langchain.llms.base import BaseLLM, LLM
from langchain.callbacks.manager import CallbackManagerForLLMRun, Callbacks
import torch
from transformers import AutoTokenizer, AutoModel, AutoConfig, AutoModelForCausalLM
from transformers.generation.utils import GenerationConfig
from pydantic import root_validator
class BaichuanLLM(LLM):
model_name: str = "baichuan-inc/Baichuan-13B-Chat"
quantization_bit: Optional[int] = None
tokenizer: AutoTokenizer = None
model: AutoModel = None
def _llm_type(self) -> str:
return "chatglm_local"
@root_validator()
def validate_environment(self, values: Dict) -> Dict:
if not values["model_name"]:
raise ValueError("No model name provided.")
model_name = values["model_name"]
tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype=torch.float16,
# device_map="auto",
trust_remote_code=True
)
model.generation_config = GenerationConfig.from_pretrained(
model_name
)
if values["quantization_bit"]:
print(f"Quantized to {values['quantization_bit']} bit")
model = model.quantize(values["quantization_bit"]).cuda()
else:
model = model.half().cuda()
model = model.eval()
values["tokenizer"] = tokenizer
values["model"] = model
return values
def _call(self, prompt: str, stop: Optional[List[str]] = None,
run_manager: Optional[CallbackManagerForLLMRun] = None, **kwargs) -> str:
message = [{"role": "user", "content": prompt}]
resp = self.model.chat(self.tokenizer, message)
# print(f"prompt:{prompt}\nresponse:{resp}\n")
return resp
import logging
import os
from typing import Any, Dict, List, Mapping, Optional
from langchain.llms.base import BaseLLM, LLM
from langchain.schema import LLMResult
from langchain.utils import get_from_dict_or_env
from langchain.callbacks.manager import CallbackManagerForLLMRun, Callbacks
from enum import Enum
from pydantic import root_validator, Field
from .xinghuo import SparkApi
from .xinghuo.ws import SparkAPI
logger = logging.getLogger(__name__)
text = []
# length = 0
def getText(role, content):
jsoncon = {"role": role, "content": content}
text.append(jsoncon)
return text
def getlength(text):
length = 0
for content in text:
temp = content["content"]
leng = len(temp)
length += leng
return length
def checklen(_text):
while getlength(_text) > 8000:
del _text[0]
return _text
class SparkLLM(LLM):
"""
ErnieLLM is a LLM that uses Ernie to generate text.
"""
appid: str = Field(
None,
description="APPID",
)
api_key: str = Field(
None,
description="API_KEY",
)
api_secret: str = Field(
None,
description="API_SECRET",
)
version: str = Field(
None,
description="version",
)
api: SparkAPI = Field(
None,
description="api",
)
@root_validator()
def validate_environment(self, values: Dict) -> Dict:
"""Validate the environment."""
# print(values)
appid = get_from_dict_or_env(values, "appid", "XH_APPID", "")
api_key = get_from_dict_or_env(values, "api_key", "XH_API_KEY", "")
api_secret = get_from_dict_or_env(values, "api_secret", "XH_API_SECRET", "")
version = values.get("version", "v1")
if not appid:
raise ValueError("No appid provided.")
if not api_key:
raise ValueError("No api_key provided.")
if not api_secret:
raise ValueError("No api_secret provided.")
values["appid"] = appid
values["api_key"] = api_key
values["api_secret"] = api_secret
api = SparkAPI(appid, api_key, api_secret, version)
values["api"] = api
return values
def _call(self, prompt: str, stop: Optional[List[str]] = None,
run_manager: Optional[CallbackManagerForLLMRun] = None, **kwargs) -> str:
question = self.getText("user", prompt)
try:
# 你的代码
# SparkApi.main(self.appid,self.api_key,self.api_secret,self.Spark_url,self.domain,question)
self.api.call(question)
response = self.api.answer
return response
except Exception as e:
# 处理异常
print("exception:", e)
raise e
def getText(self, role, content):
text = []
jsoncon = {}
jsoncon["role"] = role
jsoncon["content"] = content
text.append(jsoncon)
return text
@property
def _llm_type(self) -> str:
"""Return type of llm."""
return "xinghuo"
from langchain.llms.base import LLM
from langchain.callbacks.manager import CallbackManagerForLLMRun
from pydantic import root_validator
from typing import Dict, List, Optional
from transformers import PreTrainedModel, PreTrainedTokenizer
class WrapperLLM(LLM):
tokenizer: PreTrainedTokenizer = None
model: PreTrainedModel = None
@root_validator()
def validate_environment(self, values: Dict) -> Dict:
"""Validate the environment."""
# print(values)
if values.get("model") is None:
raise ValueError("No model provided.")
if values.get("tokenizer") is None:
raise ValueError("No tokenizer provided.")
return values
def _call(self, prompt: str, stop: Optional[List[str]] = None,
run_manager: Optional[CallbackManagerForLLMRun] = None, **kwargs) -> str:
resp, his = self.model.chat(self.tokenizer, prompt)
return resp
@property
def _llm_type(self) -> str:
"""Return type of llm."""
return "wrapper"
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
answer = ""
class Ws_Param(object):
# 初始化
def __init__(self, APPID, APIKey, APISecret, Spark_url):
self.APPID = APPID
self.APIKey = APIKey
self.APISecret = APISecret
self.host = urlparse(Spark_url).netloc
self.path = urlparse(Spark_url).path
self.Spark_url = Spark_url
# 生成url
def 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url
url = self.Spark_url + '?' + urlencode(v)
# 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, domain= ws.domain,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:
choices = data["payload"]["choices"]
status = choices["status"]
content = choices["text"][0]["content"]
print(content,end ="")
global answer
answer += content
# print(1)
if status == 2:
ws.close()
def gen_params(appid, domain,question):
"""
通过appid和用户的提问来生成请参数
"""
data = {
"header": {
"app_id": appid,
"uid": "1234"
},
"parameter": {
"chat": {
"domain": domain,
"random_threshold": 0.5,
"max_tokens": 2048,
"auditing": "default"
}
},
"payload": {
"message": {
"text": question
}
}
}
return data
def main(appid, api_key, api_secret, Spark_url,domain, question):
# print("星火:")
global answer
answer = ""
wsParam = Ws_Param(appid, api_key, api_secret, Spark_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 = appid
ws.question = question
ws.domain = domain
ws.run_forever(sslopt={"cert_reqs": ssl.CERT_NONE})
import SparkApi
#以下密钥信息从控制台获取
appid = "XXXXXXXX" #填写控制台中获取的 APPID 信息
api_secret = "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX" #填写控制台中获取的 APISecret 信息
api_key ="XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX" #填写控制台中获取的 APIKey 信息
#用于配置大模型版本,默认“general/generalv2”
domain = "general" # v1.5版本
# domain = "generalv2" # v2.0版本
#云端环境的服务地址
Spark_url = "ws://spark-api.xf-yun.com/v1.1/chat" # v1.5环境的地址
# Spark_url = "ws://spark-api.xf-yun.com/v2.1/chat" # v2.0环境的地址
text =[]
# length = 0
def getText(role,content):
jsoncon = {}
jsoncon["role"] = role
jsoncon["content"] = content
text.append(jsoncon)
return text
def getlength(text):
length = 0
for content in text:
temp = content["content"]
leng = len(temp)
length += leng
return length
def checklen(text):
while (getlength(text) > 8000):
del text[0]
return text
if __name__ == '__main__':
text.clear
while(1):
Input = input("\n" +"我:")
question = checklen(getText("user",Input))
SparkApi.answer =""
print("星火:",end = "")
SparkApi.main(appid,api_key,api_secret,Spark_url,domain,question)
getText("assistant",SparkApi.answer)
# print(str(text))
import _thread as thread
import base64
import datetime
import hashlib
import hmac
import json
import ssl
from datetime import datetime
from time import mktime
from urllib.parse import urlparse, urlencode
from wsgiref.handlers import format_date_time
import websocket # 使用websocket_client
URL_V1_5="ws://spark-api.xf-yun.com/v1.1/chat"
URL_V2="ws://spark-api.xf-yun.com/v2.1/chat"
Domain_V1_5="general"
Domain_V2="generalv2"
class SparkAPI:
def __init__(self, APPID, APIKey, APISecret, Version="v1"):
self.APPID = APPID
self.APIKey = APIKey
self.APISecret = APISecret
if Version == "v1":
self.Spark_url = URL_V1_5
self.domain = Domain_V1_5
elif Version == "v2":
self.Spark_url = URL_V2
self.domain = Domain_V2
self.host = urlparse(self.Spark_url).netloc
self.path = urlparse(self.Spark_url).path
self.answer = ""
def create_url(self):
now = datetime.now()
date = format_date_time(mktime(now.timetuple()))
signature_origin = "host: " + self.host + "\n"
signature_origin += "date: " + date + "\n"
signature_origin += "GET " + self.path + " HTTP/1.1"
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')
v = {
"authorization": authorization,
"date": date,
"host": self.host
}
url = self.Spark_url + '?' + urlencode(v)
return url
def on_error(self, ws, error):
print("### error:", error)
def on_close(self, ws, one, two):
print(" ")
def on_open(self, ws):
thread.start_new_thread(self.run, (ws,))
def run(self, ws, *args):
data = json.dumps(self.gen_params(appid=self.APPID, domain=self.domain, question=ws.question))
ws.send(data)
def on_message(self, ws, message):
data = json.loads(message)
code = data['header']['code']
if code != 0:
print(f'请求错误: {code}, {data}')
ws.close()
else:
choices = data["payload"]["choices"]
status = choices["status"]
content = choices["text"][0]["content"]
# print(content, end="")
self.answer += content
if status == 2:
ws.close()
def gen_params(self, appid, domain, question):
data = {
"header": {
"app_id": appid,
"uid": "1234"
},
"parameter": {
"chat": {
"domain": domain,
"random_threshold": 0.5,
"max_tokens": 2048,
"auditing": "default"
}
},
"payload": {
"message": {
"text": question
}
}
}
return data
def call(self, question):
self.answer = ""
wsUrl = self.create_url()
websocket.enableTrace(False)
ws = websocket.WebSocketApp(wsUrl, on_message=self.on_message, on_error=self.on_error, on_close=self.on_close, on_open=self.on_open)
ws.question = question
ws.run_forever(sslopt={"cert_reqs": ssl.CERT_NONE})
\ No newline at end of file
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment