Bump PMAT & Pydantic AI

prediction_prophet/app.py

@@ -7,7 +7,7 @@
 from pydantic_ai.models import KnownModelName
 from pydantic_ai.settings import ModelSettings
 from prediction_market_agent_tooling.tools.utils import secret_str_from_env
-from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain_text_splitters import RecursiveCharacterTextSplitter
 from prediction_prophet.functions.create_embeddings_from_results import create_embeddings_from_results
 from prediction_prophet.functions.generate_subqueries import generate_subqueries
 from prediction_prophet.functions.prepare_report import prepare_report
@@ -18,8 +18,8 @@
 def research(
     goal: str,
     tavily_api_key: SecretStr,
-    model: KnownModelName = "gpt-4o",
-    temperature: float = 0.7, 
+    model: KnownModelName = "openai:gpt-4o",
+    temperature: float = 0.7,
     initial_subqueries_limit: int = 20,
     subqueries_limit: int = 4,
     scrape_content_split_chunk_size: int = 800,
@@ -30,19 +30,19 @@ def research(
 
     with st.status("Generating subqueries"):
         queries = generate_subqueries(query=goal, limit=initial_subqueries_limit, agent=agent)
-    
+
         stringified_queries = '\n- ' + '\n- '.join(queries)
         st.write(f"Generated subqueries: {stringified_queries}")
-        
+
     with st.status("Reranking subqueries"):
         queries = rerank_subqueries(queries=queries, goal=goal, agent=agent)[:subqueries_limit] if initial_subqueries_limit > subqueries_limit else queries
 
         stringified_queries = '\n- ' + '\n- '.join(queries)
         st.write(f"Reranked subqueries. Will use top {subqueries_limit}: {stringified_queries}")
-    
+
     with st.status("Searching the web"):
         search_results_with_queries = search(
-            queries, 
+            queries,
             lambda result: not result.url.startswith("https://www.youtube"),
             tavily_api_key=tavily_api_key
         )
@@ -52,36 +52,36 @@ def research(
 
         scrape_args = [result for (_, result) in search_results_with_queries]
         websites_to_scrape = set([result.url for result in scrape_args])
-        
+
         stringified_websites = '\n- ' + '\n- '.join(websites_to_scrape)
         st.write(f"Found the following relevant results: {stringified_websites}")
-    
+
     with st.status(f"Scraping web results"):
         scraped = scrape_results(scrape_args)
         scraped = [result for result in scraped if result.content != ""]
-        
+
         st.write(f"Scraped content from {len(scraped)} websites")
 
     text_splitter = RecursiveCharacterTextSplitter(
         separators=["\n\n", "\n", ". ", "  "],
         chunk_size=scrape_content_split_chunk_size,
         chunk_overlap=scrape_content_split_chunk_overlap
     )
-    
+
     with st.status(f"Performing similarity searches"):
         collection = create_embeddings_from_results(scraped, text_splitter)
         st.write("Created embeddings")
 
         vector_result_texts: list[str] = []
         url_to_content_deemed_most_useful: dict[str, str] = {}
-        
+
         for query in queries:
             top_k_per_query_results = collection.similarity_search(query, k=top_k_per_query)
             vector_result_texts += [result.page_content for result in top_k_per_query_results if result.page_content not in vector_result_texts]
 
             for x in top_k_per_query_results:
                 url_to_content_deemed_most_useful[x.metadata["url"]] = x.metadata["content"]
-        
+
             st.write(f"Similarity searched for: {query}")
 
         st.write(f"Found {len(vector_result_texts)} relevant information chunks")
@@ -106,7 +106,7 @@ def research(
     st.title('Prediction Prophet')
     st.markdown("A web3 agent by [Agentcoin](https://www.agentcoin.tv/)")
     st.image('https://raw.githubusercontent.com/agentcoinorg/predictionprophet/main/docs/imgs/banner.png')
-    
+
     st.markdown('#')
     st.markdown('#')
     st.markdown('#')
@@ -116,35 +116,35 @@ def research(
     st.caption('View the source code on our [github](https://github.com/agentcoinorg/predictionprophet)')
     st.caption('Learn more on our [substack](https://www.agentcoin.tv/blog/prediction-prophet)')
     st.caption('Join our [discord](https://discord.com/invite/6gk85fetcT)')
-    
+
 
 # TODO: find a better way to clear the history
 progress_placeholder = st.empty()
 
 if question := st.chat_input(placeholder='Will Twitter implement a new misinformation policy before the end of 2024?'):
     progress_placeholder.empty()
     time.sleep(0.1) # https://github.com/streamlit/streamlit/issues/5044
-    
+
     with progress_placeholder.container():
         st.chat_message("user").write(question)
-        
+
         with st.chat_message("assistant"):
             st.write(f"I will evaluate the probability of '{question}' occurring")
-            
+
             with st.status("Evaluating question") as status:
-                is_predictable = is_predictable_binary(question=question) 
+                is_predictable = is_predictable_binary(question=question)
                 if not is_predictable:
                     st.container().error(f"The agent thinks this question is not predictable.")
                     status.update(label="Error evaluating question", state="error", expanded=True)
                     st.stop()
-            
+
             report = research(
                 goal=question,
                 subqueries_limit=6,
                 top_k_per_query=15,
                 tavily_api_key=tavily_api_key,
             )
-                    
+
             with st.status("Making prediction"):
                 prediction = _make_prediction(market_question=question, additional_information=report, agent=Agent("gpt-4o", model_settings=ModelSettings(temperature=0.0)))
                 st.write(f"Probability: {prediction.p_yes * 100}%. Confidence: {prediction.confidence * 100}%")
@@ -153,6 +153,5 @@ def research(
                 if not prediction:
                     st.container().error("No prediction was generated.")
                     st.stop()
-
-            st.markdown(f"With **{prediction.confidence * 100}% confidence**, I'd say this outcome has a **{prediction.p_yes * 100}% probability** of happening")
 
+            st.markdown(f"With **{prediction.confidence * 100}% confidence**, I'd say this outcome has a **{prediction.p_yes * 100}% probability** of happening")