新增 千帆大模型的接入；对llm相关源码加载和逻辑处理做了大优化，可能有bug，待测试

requirements.txt

@@ -61,4 +61,5 @@ pure-protobuf==3.0.0a5
 pyaudio
 flask
 flask_cors
-xingchen
+xingchen
+git+https://github.com/Ikaros-521/WenxinWorkshop-Python-SDK

requirements_common.txt

@@ -144,4 +144,5 @@ flask==2.3.3
 flask_cors==4.0.0
 xingchen==1.0.7
 qianfan==0.2.2
-socketio==0.2.1
+socketio==0.2.1
+git+https://github.com/Ikaros-521/WenxinWorkshop-Python-SDK

src/Function/index.js

@@ -0,0 +1,6 @@
+exports.handler = async event => {
+  // Log the event argument for debugging and for use in local development.
+  console.log(JSON.stringify(event, undefined, 2));
+
+  return {};
+};

src/Function/package.json

@@ -0,0 +1,4 @@
+{
+  "name": "function",
+  "version": "1.0.0"
+}

tests/test_wenxinworkshop/demo.py

@@ -0,0 +1,131 @@
+import json, logging, traceback
+from wenxinworkshop import LLMAPI, EmbeddingAPI, PromptTemplateAPI
+from wenxinworkshop import Message, Messages, Texts
+
+# 前往官网：https://cloud.baidu.com/product/wenxinworkshop 申请服务获取
+
+class My_WenXinWorkShop:
+    def __init__(self, data):
+        # self.common = Common()
+        # # 日志文件路径
+        # file_path = "./log/log-" + self.common.get_bj_time(1) + ".txt"
+        # Configure_logger(file_path)
+
+        self.config_data = data
+        self.history = []
+
+        try:
+            model_url_map = {
+                "ERNIEBot": LLMAPI.ERNIEBot,
+                "ERNIEBot_turbo": LLMAPI.ERNIEBot_turbo,
+                "ERNIEBot_4_0": LLMAPI.ERNIEBot_4_0,
+                "BLOOMZ_7B": LLMAPI.BLOOMZ_7B,
+                "LLAMA_2_7B": LLMAPI.LLAMA_2_7B,
+                "LLAMA_2_13B": LLMAPI.LLAMA_2_13B,
+                "LLAMA_2_70B": LLMAPI.LLAMA_2_70B,
+                "ERNIEBot_4_0": LLMAPI.ERNIEBot_4_0,
+                "QIANFAN_BLOOMZ_7B_COMPRESSED": LLMAPI.QIANFAN_BLOOMZ_7B_COMPRESSED,
+                "QIANFAN_CHINESE_LLAMA_2_7B": LLMAPI.QIANFAN_CHINESE_LLAMA_2_7B,
+                "CHATGLM2_6B_32K": LLMAPI.CHATGLM2_6B_32K,
+                "AQUILACHAT_7B": LLMAPI.AQUILACHAT_7B,
+                "ERNIE_BOT_8K": LLMAPI.ERNIE_BOT_8K,
+                "CODELLAMA_7B_INSTRUCT": LLMAPI.CODELLAMA_7B_INSTRUCT,
+                "XUANYUAN_70B_CHAT": LLMAPI.XUANYUAN_70B_CHAT,
+                "CHATLAW": LLMAPI.QIANFAN_BLOOMZ_7B_COMPRESSED,
+                "QIANFAN_BLOOMZ_7B_COMPRESSED": LLMAPI.CHATLAW,
+            }
+
+            selected_model = self.config_data["model"]
+            if selected_model in model_url_map:
+                self.my_bot = LLMAPI(
+                    api_key=self.config_data["api_key"],
+                    secret_key=self.config_data["secret_key"],
+                    url=model_url_map[selected_model]
+                )
+        except Exception as e:
+            logging.error(traceback.format_exc())
+
+
+
+    def get_resp(self, prompt):
+        """请求对应接口，获取返回值
+
+        Args:
+            prompt (str): 你的提问
+
+        Returns:
+            str: 返回的文本回答
+        """
+        try:
+            # create messages
+            messages: Messages = []
+
+            for history in self.history:
+                messages.append(Message(
+                    role=history["role"],
+                    content=history["content"]
+                ))
+
+            messages.append(Message(
+                role='user',
+                content=prompt
+            ))
+
+            logging.info(f"self.history={self.history}")
+
+            # get response from LLM API
+            resp_content = self.my_bot(
+                messages=messages,
+                temperature=self.config_data["temperature"],
+                top_p=self.config_data["top_p"],
+                penalty_score=self.config_data["penalty_score"],
+                stream=None,
+                user_id=None,
+                chunk_size=512
+            )
+
+            # 启用历史就给我记住！
+            if self.config_data["history_enable"]:
+                while True:
+                    # 获取嵌套列表中所有字符串的字符数
+                    total_chars = sum(len(item['content']) for item in self.history if 'content' in item)
+                    # 如果大于限定最大历史数，就剔除第一个元素
+                    if total_chars > self.config_data["history_max_len"]:
+                        self.history.pop(0)
+                        self.history.pop(0)
+                    else:
+                        # self.history.pop()
+                        self.history.append({"role": "user", "content": prompt})
+                        self.history.append({"role": "assistant", "content": resp_content})
+                        break
+
+            return resp_content
+
+        except Exception as e:
+            logging.error(e)
+
+        return None
+
+if __name__ == '__main__':
+    # 配置日志输出格式
+    logging.basicConfig(
+        level=logging.DEBUG,  # 设置日志级别，可以根据需求调整
+        format="%(asctime)s [%(levelname)s] %(message)s",
+        datefmt="%Y-%m-%d %H:%M:%S",
+    )
+
+    data = {
+        "model": "ERNIEBot",
+        "api_key": "",
+        "secret_key": "",
+        "top_p": 0.8,
+        "temperature": 0.9,
+        "penalty_score": 1.0,
+        "history_enable": True,
+        "history_max_len": 300
+    }
+
+    # 实例化并调用
+    my_wenxinworkshop = My_WenXinWorkShop(data)
+    logging.info(my_wenxinworkshop.get_resp("你可以扮演猫娘吗，每句话后面加个喵"))
+    logging.info(my_wenxinworkshop.get_resp("早上好"))

utils/gpt_model/gpt.py

@@ -23,7 +23,7 @@
 from utils.gpt_model.tongyi import TongYi
 from utils.gpt_model.tongyixingchen import TongYiXingChen
 from utils.gpt_model.my_qianfan import My_QianFan
-
+from utils.gpt_model.my_wenxinworkshop import My_WenXinWorkShop
 
 class GPT_Model:
     openai = None
@@ -42,6 +42,7 @@ def set_model_config(self, model_name, config):
             "yiyan": Yiyan,
             "tongyi": TongYi,
             "tongyixingchen": TongYiXingChen,
+            "my_wenxinworkshop": My_WenXinWorkShop,
             "my_qianfan": My_QianFan
         }
 

utils/gpt_model/my_wenxinworkshop.py

@@ -0,0 +1,110 @@
+import json, logging, traceback
+from wenxinworkshop import LLMAPI, EmbeddingAPI, PromptTemplateAPI
+from wenxinworkshop import Message, Messages, Texts
+
+# 前往官网：https://cloud.baidu.com/product/wenxinworkshop 申请服务获取
+
+class My_WenXinWorkShop:
+    def __init__(self, data):
+        # self.common = Common()
+        # # 日志文件路径
+        # file_path = "./log/log-" + self.common.get_bj_time(1) + ".txt"
+        # Configure_logger(file_path)
+
+        self.config_data = data
+        self.history = []
+
+        try:
+            # create a LLM API
+            self.my_bot = LLMAPI(
+                api_key=self.config_data["api_key"],
+                secret_key=self.config_data["secret_key"],
+                url=LLMAPI.ERNIEBot
+            )
+        except Exception as e:
+            logging.error(traceback.format_exc())
+
+
+
+    def get_resp(self, prompt):
+        """请求对应接口，获取返回值
+
+        Args:
+            prompt (str): 你的提问
+
+        Returns:
+            str: 返回的文本回答
+        """
+        try:
+            # create messages
+            messages: Messages = []
+
+            for history in self.history:
+                messages.append(Message(
+                    role=history["role"],
+                    content=history["content"]
+                ))
+
+            messages.append(Message(
+                role='user',
+                content=prompt
+            ))
+
+            logging.info(f"self.history={self.history}")
+
+            # get response from LLM API
+            resp_content = self.my_bot(
+                messages=messages,
+                temperature=self.config_data["temperature"],
+                top_p=self.config_data["top_p"],
+                penalty_score=self.config_data["penalty_score"],
+                stream=None,
+                user_id=None,
+                chunk_size=512
+            )
+
+            # 启用历史就给我记住！
+            if self.config_data["history_enable"]:
+                while True:
+                    # 获取嵌套列表中所有字符串的字符数
+                    total_chars = sum(len(item['content']) for item in self.history if 'content' in item)
+                    # 如果大于限定最大历史数，就剔除第一个元素
+                    if total_chars > self.config_data["history_max_len"]:
+                        self.history.pop(0)
+                        self.history.pop(0)
+                    else:
+                        # self.history.pop()
+                        self.history.append({"role": "user", "content": prompt})
+                        self.history.append({"role": "assistant", "content": resp_content})
+                        break
+
+            return resp_content
+
+        except Exception as e:
+            logging.error(e)
+
+        return None
+
+if __name__ == '__main__':
+    # 配置日志输出格式
+    logging.basicConfig(
+        level=logging.DEBUG,  # 设置日志级别，可以根据需求调整
+        format="%(asctime)s [%(levelname)s] %(message)s",
+        datefmt="%Y-%m-%d %H:%M:%S",
+    )
+
+    data = {
+        "model": "ERNIEBot",
+        "api_key": "",
+        "secret_key": "",
+        "top_p": 0.8,
+        "temperature": 0.9,
+        "penalty_score": 1.0,
+        "history_enable": True,
+        "history_max_len": 300
+    }
+
+    # 实例化并调用
+    my_wenxinworkshop = My_WenXinWorkShop(data)
+    logging.info(my_wenxinworkshop.get_resp("你可以扮演猫娘吗，每句话后面加个喵"))
+    logging.info(my_wenxinworkshop.get_resp("早上好"))