botrun-flow-lang 5.9.301__py3-none-any.whl → 5.10.82__py3-none-any.whl

botrun_flow_lang/langgraph_agents/agents/util/gemini_grounding.py

@@ -1,66 +1,66 @@
-from typing import AsyncGenerator
-from google.generativeai import GenerativeModel
-from google.generativeai.protos import Candidate
-
-from pydantic import BaseModel
-
-
-class GeminiGroundingEvent(BaseModel):
-    chunk: str
-
-
-async def respond_with_gemini_grounding(
-    input_content, messages_for_llm
-) -> AsyncGenerator[GeminiGroundingEvent, None]:
-    model = GenerativeModel("models/gemini-1.5-pro-002")
-
-    # 將 litellm 格式轉換為 Gemini 格式的對話歷史
-    chat_history = []
-    for message in messages_for_llm:
-        role = "user" if message["role"] == "user" else "model"
-        chat_history.append({"role": role, "parts": [message["content"]]})
-
-    # 建立聊天實例
-    chat = model.start_chat(history=chat_history)
-
-    # 使用 chat.send() 替代 generate_content
-    responses = chat.send_message(
-        input_content,
-        tools={"google_search_retrieval": {"dynamic_retrieval_config": {}}},
-        stream=True,
-    )
-
-    # answer_message = await cl.Message(content="").send()
-    full_response = ""
-    for response in responses:
-        if response.candidates[0].finish_reason != Candidate.FinishReason.STOP:
-            # await answer_message.stream_token(response.text)
-            yield GeminiGroundingEvent(chunk=response.text)
-            full_response += response.text
-            if response.candidates[0].grounding_metadata:
-                if len(response.candidates[0].grounding_metadata.grounding_chunks) > 0:
-                    references = f"\n\n{tr('Sources:')}\n"
-                    for grounding_chunk in response.candidates[
-                        0
-                    ].grounding_metadata.grounding_chunks:
-                        references += f"- [{grounding_chunk.web.title}]({grounding_chunk.web.uri})\n"
-                    # await answer_message.stream_token(references)
-                    yield GeminiGroundingEvent(chunk=references)
-        else:
-            if response.candidates[0].grounding_metadata:
-                if len(response.candidates[0].grounding_metadata.grounding_chunks) > 0:
-                    references = f"\n\n參考來源：\n"
-                    for grounding_chunk in response.candidates[
-                        0
-                    ].grounding_metadata.grounding_chunks:
-                        references += f"- [{grounding_chunk.web.title}]({grounding_chunk.web.uri})\n"
-                    # await answer_message.stream_token(references)
-                    yield GeminiGroundingEvent(chunk=references)
-
-    # await answer_message.update()
-
-    # chat_history = cl.user_session.get("chat_history", [])
-    # chat_history.append({"role": "user", "content": input_content})
-    # chat_history.append({"role": "assistant", "content": full_response})
-    # cl.user_session.set("chat_history", chat_history)
-    # return
+from typing import AsyncGenerator
+from google.generativeai import GenerativeModel
+from google.generativeai.protos import Candidate
+
+from pydantic import BaseModel
+
+
+class GeminiGroundingEvent(BaseModel):
+    chunk: str
+
+
+async def respond_with_gemini_grounding(
+    input_content, messages_for_llm
+) -> AsyncGenerator[GeminiGroundingEvent, None]:
+    model = GenerativeModel("models/gemini-1.5-pro-002")
+
+    # 將 litellm 格式轉換為 Gemini 格式的對話歷史
+    chat_history = []
+    for message in messages_for_llm:
+        role = "user" if message["role"] == "user" else "model"
+        chat_history.append({"role": role, "parts": [message["content"]]})
+
+    # 建立聊天實例
+    chat = model.start_chat(history=chat_history)
+
+    # 使用 chat.send() 替代 generate_content
+    responses = chat.send_message(
+        input_content,
+        tools={"google_search_retrieval": {"dynamic_retrieval_config": {}}},
+        stream=True,
+    )
+
+    # answer_message = await cl.Message(content="").send()
+    full_response = ""
+    for response in responses:
+        if response.candidates[0].finish_reason != Candidate.FinishReason.STOP:
+            # await answer_message.stream_token(response.text)
+            yield GeminiGroundingEvent(chunk=response.text)
+            full_response += response.text
+            if response.candidates[0].grounding_metadata:
+                if len(response.candidates[0].grounding_metadata.grounding_chunks) > 0:
+                    references = f"\n\n{tr('Sources:')}\n"
+                    for grounding_chunk in response.candidates[
+                        0
+                    ].grounding_metadata.grounding_chunks:
+                        references += f"- [{grounding_chunk.web.title}]({grounding_chunk.web.uri})\n"
+                    # await answer_message.stream_token(references)
+                    yield GeminiGroundingEvent(chunk=references)
+        else:
+            if response.candidates[0].grounding_metadata:
+                if len(response.candidates[0].grounding_metadata.grounding_chunks) > 0:
+                    references = f"\n\n參考來源：\n"
+                    for grounding_chunk in response.candidates[
+                        0
+                    ].grounding_metadata.grounding_chunks:
+                        references += f"- [{grounding_chunk.web.title}]({grounding_chunk.web.uri})\n"
+                    # await answer_message.stream_token(references)
+                    yield GeminiGroundingEvent(chunk=references)
+
+    # await answer_message.update()
+
+    # chat_history = cl.user_session.get("chat_history", [])
+    # chat_history.append({"role": "user", "content": input_content})
+    # chat_history.append({"role": "assistant", "content": full_response})
+    # cl.user_session.set("chat_history", chat_history)
+    # return