Use loguru instead of printing (#57)

main.py

@@ -1,6 +1,7 @@
 from decimal import Decimal
 
 import typer
+from loguru import logger
 from prediction_market_agent_tooling.markets.agent_market import SortBy
 from prediction_market_agent_tooling.markets.markets import (
     MarketType,
@@ -43,7 +44,7 @@ def main(
         do_bet = user_input.lower().strip() == "y" if user_input else True
 
     if do_bet:
-        print(
+        logger.info(
             f"Placing bet with position {pma.utils.parse_result_to_str(result)} on market '{market.question}'"
         )
         amount = Decimal(

prediction_market_agent/agents/custom_agent.py

@@ -2,6 +2,7 @@
 from typing import Optional
 
 import requests
+from loguru import logger
 from prediction_market_agent_tooling.markets.agent_market import AgentMarket
 from prediction_market_agent_tooling.tools.utils import (
     check_not_none,
@@ -62,9 +63,9 @@ def __init__(self, model: AbstractAiChatModel, max_cycles: int, verbose: bool):
         )(web_scrape_structured_and_summarized)
         self.model = model
 
-    def verbose_print(self, message: str) -> None:
+    def verbose_log(self, message: str) -> None:
         if self.verbose:
-            print(f"{message}\n")
+            logger.info(f"{message}\n")
 
     def answer_binary_market(self, market: AgentMarket) -> bool:
         cycle_count = 0
@@ -78,7 +79,7 @@ def answer_binary_market(self, market: AgentMarket) -> bool:
             cycle_count += 1
             if cycle_count == self.max_cycles - 1:
                 # If we reached the maximum number of cycles, ask to provide an answer based on available information.
-                self.verbose_print(
+                self.verbose_log(
                     f"Agent: Reached {cycle_count} cycles, asking to provide answer now."
                 )
                 messages.append(
@@ -97,7 +98,7 @@ def answer_binary_market(self, market: AgentMarket) -> bool:
             completion_str = check_not_none(
                 self.model.complete(messages), "Couldn't complete the prompt."
             )
-            self.verbose_print(f"Completion: {completion_str}")
+            self.verbose_log(f"Completion: {completion_str}")
             try:
                 completion_dict = json.loads(completion_str)
             except json.decoder.JSONDecodeError:
@@ -108,7 +109,7 @@ def answer_binary_market(self, market: AgentMarket) -> bool:
             # - tool_name: Name of the tool to be used.
             # - answer: Final answer to the question.
             if thinking := completion_dict.get("thinking"):
-                self.verbose_print(f"Agent: Thinking about {thinking=}.")
+                self.verbose_log(f"Agent: Thinking about {thinking=}.")
                 messages.append(
                     Message(
                         role="assistant", content=json.dumps({"thinking": thinking})
@@ -117,7 +118,7 @@ def answer_binary_market(self, market: AgentMarket) -> bool:
 
             elif tool_name := completion_dict.get("tool_name"):
                 tool_params = completion_dict.get("tool_params")
-                self.verbose_print(f"Agent: Using {tool_name=} with {tool_params=}.")
+                self.verbose_log(f"Agent: Using {tool_name=} with {tool_params=}.")
                 tool_output = (  # type: ignore # Untyped for the sake of simplicity when the LLM is used.
                     self.google_search
                     if tool_name == "GoogleSearchTool"
@@ -129,7 +130,7 @@ def answer_binary_market(self, market: AgentMarket) -> bool:
                 )(
                     **tool_params
                 )
-                self.verbose_print(f"Tool: {tool_name=} returns {tool_output=}.")
+                self.verbose_log(f"Tool: {tool_name=} returns {tool_output=}.")
                 messages.append(
                     Message(
                         role="assistant",
@@ -140,7 +141,7 @@ def answer_binary_market(self, market: AgentMarket) -> bool:
                 )
 
             elif answer := completion_dict.get("answer"):
-                self.verbose_print(f"Agent: Answering {answer=}.")
+                self.verbose_log(f"Agent: Answering {answer=}.")
                 messages.append(
                     Message(role="assistant", content=json.dumps({"answer": answer}))
                 )

prediction_market_agent/agents/known_outcome_agent/benchmark.py

@@ -3,6 +3,7 @@
 from datetime import timedelta
 
 from dotenv import load_dotenv
+from loguru import logger
 from prediction_market_agent_tooling.benchmark.agents import AbstractBenchmarkedAgent
 from prediction_market_agent_tooling.benchmark.benchmark import Benchmarker
 from prediction_market_agent_tooling.benchmark.utils import (
@@ -62,7 +63,9 @@ def predict(self, market_question: str) -> Prediction:
             question=market_question,
             max_tries=self.max_tries,
         )
-        print(f"Answered {market_question=} with {answer.result=}, {answer.reasoning=}")
+        logger.info(
+            f"Answered {market_question=} with {answer.result=}, {answer.reasoning=}"
+        )
         if not answer.has_known_result():
             return Prediction(
                 is_predictable=False,
@@ -161,7 +164,7 @@ def predict(self, market_question: str) -> Prediction:
 
     output = f"./known_outcome_agent_benchmark_report.{int(time.time())}.md"
     with open(output, "w") as f:
-        print(f"Writing benchmark report to: {output}")
+        logger.info(f"Writing benchmark report to: {output}")
         f.write(md)
 
     # Check all predictions are correct, i.e. mean-squared-error == 0

prediction_market_agent/agents/known_outcome_agent/deploy.py

@@ -2,6 +2,7 @@
 import random
 from decimal import Decimal
 
+from loguru import logger
 from prediction_market_agent_tooling.config import APIKeys
 from prediction_market_agent_tooling.deploy.agent import DeployableAgent
 from prediction_market_agent_tooling.deploy.constants import OWNER_KEY
@@ -40,17 +41,17 @@ def pick_markets(self, markets: list[AgentMarket]) -> list[AgentMarket]:
                     "This agent only supports predictions on Omen markets"
                 )
 
-            print(f"Looking at market {market.id=} {market.question=}")
+            logger.info(f"Looking at market {market.id=} {market.question=}")
 
             # Assume very high probability markets are already known, and have
             # been correctly bet on, and therefore the value of betting on them
             # is low.
             if market_is_saturated(market=market):
-                print(
+                logger.info(
                     f"Skipping market {market.id=} {market.question=}, because it is already saturated."
                 )
             elif market.get_liquidity_in_xdai() < 5:
-                print(
+                logger.info(
                     f"Skipping market {market.id=} {market.question=}, because it has insufficient liquidity."
                 )
             else:
@@ -77,12 +78,12 @@ def answer_binary_market(self, market: AgentMarket) -> bool | None:
                 max_tries=3,
             )
         except Exception as e:
-            print(
-                f"Error: Failed to predict market {market.id=} {market.question=}: {e}"
+            logger.error(
+                f"Failed to predict market {market.id=} {market.question=}: {e}"
             )
             answer = None
         if answer and answer.has_known_result():
-            print(
+            logger.info(
                 f"Picking market {market.id=} {market.question=} with answer {answer.result=}"
             )
             return answer.result.to_boolean()

prediction_market_agent/agents/known_outcome_agent/known_outcome_agent.py

@@ -5,6 +5,7 @@
 
 from langchain.prompts import ChatPromptTemplate
 from langchain_openai import ChatOpenAI
+from loguru import logger
 from prediction_market_agent_tooling.tools.utils import utcnow
 from pydantic import BaseModel
 
@@ -195,7 +196,9 @@ def has_question_event_happened_in_the_past(model: str, question: str) -> bool:
         if parsed_answer == 1:
             return True
     except Exception as e:
-        print("Exception occured, cannot assert if title happened in the past. ", e)
+        logger.error(
+            "Exception occured, cannot assert if title happened in the past. ", e
+        )
 
     return False
 
@@ -214,9 +217,9 @@ def get_known_outcome(model: str, question: str, max_tries: int) -> Answer:
         search_prompt = ChatPromptTemplate.from_template(
             template=GENERATE_SEARCH_QUERY_PROMPT
         ).format_messages(date_str=date_str, question=question)
-        print(f"Invoking LLM for {search_prompt=}")
+        logger.debug(f"Invoking LLM for {search_prompt=}")
         search_query = str(llm.invoke(search_prompt).content).strip('"')
-        print(f"Searching for {search_query=}")
+        logger.debug(f"Searching for {search_query=}")
         search_results = web_search(query=search_query, max_results=5)
         if not search_results:
             raise ValueError("No search results found.")

prediction_market_agent/tools/web_scrape/markdown.py

@@ -1,6 +1,7 @@
 import requests
 import tenacity
 from bs4 import BeautifulSoup
+from loguru import logger
 from markdownify import markdownify
 from prediction_market_agent_tooling.tools.cache import persistent_inmemory_cache
 from requests import Response
@@ -45,9 +46,9 @@ def web_scrape(url: str, timeout: int = 10) -> str:
 
             return text
         else:
-            print("Non-HTML content received")
+            logger.warning("Non-HTML content received")
             return ""
 
     except requests.RequestException as e:
-        print(f"HTTP request failed: {e}")
+        logger.error(f"HTTP request failed: {e}")
         return ""