updated

Dockerfile

@@ -7,12 +7,15 @@ COPY ./requirements.txt /app/requirements.txt
 RUN pip3 install --no-cache-dir --upgrade pip
 RUN pip3 install --no-cache-dir wheel setuptools build
 RUN pip3 install --no-cache-dir --use-pep517 -r /app/requirements.txt
-
+ 
 # User
 RUN useradd -m -u 1000 user
 USER user
 ENV HOME /home/user
 ENV PATH $HOME/.local/bin:$PATH
+ENV TIKTOKEN_CACHE_DIR $HOME/.cache/tiktoken
+
+RUN mkdir -p $HOME/.cache/tiktoken
 
 WORKDIR $HOME
 RUN mkdir app

agent.py

@@ -5,14 +5,15 @@ import requests
 from typing import Tuple
 from bs4 import BeautifulSoup
 
-from dotenv import load_dotenv
-load_dotenv(override=True)
-
-from pydantic import Field, BaseModel
+from pydantic import Field
 from vectara_agentic.agent import Agent
 from vectara_agentic.agent_config import AgentConfig
 from vectara_agentic.tools import ToolsFactory, VectaraToolFactory
 from vectara_agentic.tools_catalog import ToolsCatalog
+from vectara_agentic.types import ModelProvider, AgentType
+
+from dotenv import load_dotenv
+load_dotenv(override=True)
 
 initial_prompt = "How can I help you today?"
 
@@ -120,24 +121,24 @@ class AgentTools:
 
 
     def get_tools(self):
-        class QueryHackerNews(BaseModel):
-            query: str = Field(..., description="The user query.")
-
-        vec_factory = VectaraToolFactory(vectara_api_key=self.cfg.api_key, 
-                                        vectara_corpus_key=self.cfg.corpus_key)
-        summarizer = 'vectara-summary-ext-24-05-med-omni'
+        vec_factory = VectaraToolFactory(
+            vectara_api_key=self.cfg.api_key, 
+            vectara_corpus_key=self.cfg.corpus_key
+        )
+        summarizer = 'vectara-summary-table-md-query-ext-jan-2025-gpt-4o'
         ask_hackernews = vec_factory.create_rag_tool(
             tool_name = "ask_hackernews",
             tool_description = """
             Provides information on any topic or query, based on relevant hacker news stories.
             """,
-            tool_args_schema = QueryHackerNews,
-            reranker = "multilingual_reranker_v1", rerank_k = 100, 
+            reranker = "multilingual_reranker_v1", rerank_k = 100, rerank_cutoff = 0.3,
+            max_tokens = 4096, max_response_chars = 8192,
             n_sentences_before = 2, n_sentences_after = 2, lambda_val = 0.0,
-            summary_num_results = 10,
+            summary_num_results = 15,
             vectara_summarizer = summarizer,
             include_citations = True,
-            verbose=True
+            verbose = False,
+            save_history = True,
         )
         
         tools_factory = ToolsFactory()
@@ -171,13 +172,30 @@ def initialize_agent(_cfg, agent_progress_callback = None):
     - You can use the tavily_search tool to gain additional information if needed for follow up questions about Hacker News topic.
     - When including information or links provided by the tavily_search tool, make sure to notify the user in your response that this is not based on Hacker News stories.
     """
-    agent_config = AgentConfig()
+    agent_config = AgentConfig(
+        agent_type = os.getenv("VECTARA_AGENTIC_AGENT_TYPE", AgentType.OPENAI.value),
+        main_llm_provider = os.getenv("VECTARA_AGENTIC_MAIN_LLM_PROVIDER", ModelProvider.OPENAI.value),
+        main_llm_model_name = os.getenv("VECTARA_AGENTIC_MAIN_MODEL_NAME", ""),
+        tool_llm_provider = os.getenv("VECTARA_AGENTIC_TOOL_LLM_PROVIDER", ModelProvider.OPENAI.value),
+        tool_llm_model_name = os.getenv("VECTARA_AGENTIC_TOOL_MODEL_NAME", ""),
+        observer = os.getenv("VECTARA_AGENTIC_OBSERVER_TYPE", "NO_OBSERVER")
+    )
+    fallback_agent_config = AgentConfig(
+        agent_type = os.getenv("VECTARA_AGENTIC_FALLBACK_AGENT_TYPE", AgentType.OPENAI.value),
+        main_llm_provider = os.getenv("VECTARA_AGENTIC_FALLBACK_MAIN_LLM_PROVIDER", ModelProvider.OPENAI.value),
+        main_llm_model_name = os.getenv("VECTARA_AGENTIC_FALLBACK_MAIN_MODEL_NAME", ""),
+        tool_llm_provider = os.getenv("VECTARA_AGENTIC_FALLBACK_TOOL_LLM_PROVIDER", ModelProvider.OPENAI.value),
+        tool_llm_model_name = os.getenv("VECTARA_AGENTIC_FALLBACK_TOOL_MODEL_NAME", ""),
+        observer = os.getenv("VECTARA_AGENTIC_OBSERVER_TYPE", "NO_OBSERVER")
+    )
     agent = Agent(
         tools=AgentTools(_cfg, agent_config).get_tools(),
         topic="hacker news",
         custom_instructions=bot_instructions,
         agent_progress_callback=agent_progress_callback,
-        verbose=True
+        verbose=True,
+        agent_config=agent_config,
+        fallback_agent_config=fallback_agent_config,
     )
     agent.report()
     return agent

requirements.txt

@@ -1,9 +1,9 @@
 omegaconf==2.3.0
 python-dotenv==1.0.1
-streamlit==1.43.2
+streamlit==1.45.0
 streamlit_feedback==0.1.3
 uuid==1.30
 langdetect==1.0.9
 langcodes==3.4.0
-vectara-agentic==0.2.9
+vectara-agentic==0.2.15
 torch==2.6.0

st_app.py

@@ -131,7 +131,7 @@ async def launch_bot():
     if st.session_state.prompt:
         with st.chat_message("assistant", avatar='🤖'):
             st.session_state.status = st.status('Processing...', expanded=False)
-            response = st.session_state.agent.chat(st.session_state.prompt)
+            response = await st.session_state.agent.achat(st.session_state.prompt)
             res = escape_dollars_outside_latex(response.response)
             message = {"role": "assistant", "content": res, "avatar": '🤖'}
             st.session_state.messages.append(message)