Squash 4 commits into one

.github/workflows/python_ci.yaml

@@ -8,7 +8,6 @@ on:
 
 env:
   MANIFOLD_API_KEY: ${{ secrets.MANIFOLD_API_KEY }}
-  SERP_API_KEY: ${{ secrets.SERP_API_KEY }}
   OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
   BET_FROM_PRIVATE_KEY: ${{ secrets.BET_FROM_PRIVATE_KEY }}
   GRAPH_API_KEY: ${{ secrets.GRAPH_API_KEY }}
@@ -38,7 +37,6 @@ jobs:
       - name: Extract environment variables for Docker container
         run: |
           echo "MANIFOLD_API_KEY=${{ secrets.MANIFOLD_API_KEY }}" >> .env
-          echo "SERP_API_KEY=${{ secrets.SERP_API_KEY }}" >> .env
           echo "OPENAI_API_KEY=${{ secrets.OPENAI_API_KEY }}" >> .env
           echo "BET_FROM_PRIVATE_KEY=${{ secrets.BET_FROM_PRIVATE_KEY }}" >> .env
           echo "GRAPH_API_KEY=${{ secrets.GRAPH_API_KEY }}" >> .env
@@ -58,8 +56,6 @@ jobs:
         env:
           RUN_PAID_TESTS: "1"
     if: github.event_name == 'workflow_dispatch'
-    # TODO: Remove once the environment variables are fixed in the new repository.
-    continue-on-error: true
 
   test-build-image:
     runs-on: ubuntu-latest

prediction_market_agent/agents/invalid_agent/deploy.py

@@ -23,7 +23,7 @@ class InvalidAgent(DeployableTraderAgent):
 
     def verify_market(self, market_type: MarketType, market: AgentMarket) -> bool:
         if self.have_bet_on_market_since(
-            market, since=self.same_market_trade_interval.get(market=market)
+            market, since=self.same_market_trade_interval.get(market)
         ):
             return False
 

prediction_market_agent/agents/known_outcome_agent/known_outcome_agent.py

@@ -9,12 +9,12 @@
     get_langfuse_langchain_config,
     observe,
 )
+from prediction_market_agent_tooling.tools.tavily.tavily_search import tavily_search
 from prediction_market_agent_tooling.tools.utils import utcnow
 from pydantic import BaseModel
 
 from prediction_market_agent.tools.web_scrape.basic_summary import _summary
 from prediction_market_agent.tools.web_scrape.markdown import web_scrape
-from prediction_market_agent.tools.web_search.tavily import web_search
 from prediction_market_agent.utils import APIKeys, completion_str_to_json
 
 
@@ -220,7 +220,7 @@ def get_known_outcome(model: str, question: str, max_tries: int) -> KnownOutcome
             ).content
         ).strip('"')
         logger.debug(f"Searching web for the search query '{search_query}'")
-        search_results = web_search(query=search_query, max_results=5)
+        search_results = tavily_search(query=search_query, max_results=5).results
         if not search_results:
             raise ValueError("No search results found.")
 

prediction_market_agent/agents/microchain_agent/market_functions.py

@@ -16,7 +16,6 @@
 from prediction_market_agent_tooling.tools.betting_strategies.kelly_criterion import (
     get_kelly_bet_simplified,
 )
-from prediction_market_agent_tooling.tools.tavily.tavily_storage import TavilyStorage
 from prediction_market_agent_tooling.tools.utils import utcnow
 
 from prediction_market_agent.agents.microchain_agent.utils import (
@@ -139,7 +138,6 @@ def __call__(self, market_id: str) -> str:
             model=self.model,
             openai_api_key=self.keys.openai_api_key,
             tavily_api_key=self.keys.tavily_api_key,
-            tavily_storage=TavilyStorage(agent_id=self.__class__.__name__),
         )
         prediction = prophet_make_prediction(
             market_question=question,

prediction_market_agent/agents/prophet_agent/deploy.py

@@ -6,7 +6,6 @@
 from prediction_market_agent_tooling.loggers import logger
 from prediction_market_agent_tooling.markets.agent_market import AgentMarket
 from prediction_market_agent_tooling.markets.data_models import ProbabilisticAnswer
-from prediction_market_agent_tooling.tools.tavily.tavily_storage import TavilyStorage
 from prediction_prophet.benchmark.agents import (
     EmbeddingModel,
     OlasAgent,
@@ -53,7 +52,6 @@ def load(self) -> None:
         self.agent = PredictionProphetAgent(
             model="gpt-4o-2024-08-06",
             include_reasoning=True,
-            tavily_storage=TavilyStorage(agent_id=self.__class__.__name__),
             logger=logger,
         )
 
@@ -74,7 +72,6 @@ def load(self) -> None:
         self.agent = PredictionProphetAgent(
             model="gpt-4-0125-preview",
             include_reasoning=True,
-            tavily_storage=TavilyStorage(agent_id=self.__class__.__name__),
             logger=logger,
         )
 
@@ -95,7 +92,6 @@ def load(self) -> None:
         self.agent = PredictionProphetAgent(
             model="gpt-4-turbo-2024-04-09",
             include_reasoning=True,
-            tavily_storage=TavilyStorage(agent_id=self.__class__.__name__),
             logger=logger,
         )
 
@@ -138,7 +134,6 @@ def load(self) -> None:
             research_temperature=1.0,
             prediction_temperature=1.0,
             include_reasoning=True,
-            tavily_storage=TavilyStorage(agent_id=self.__class__.__name__),
             logger=logger,
         )
 
@@ -162,6 +157,5 @@ def load(self) -> None:
             research_temperature=1.0,
             prediction_temperature=1.0,
             include_reasoning=True,
-            tavily_storage=TavilyStorage(agent_id=self.__class__.__name__),
             logger=logger,
         )

prediction_market_agent/agents/think_thoroughly_agent/think_thoroughly_agent.py

@@ -21,7 +21,6 @@
 )
 from prediction_market_agent_tooling.tools.langfuse_ import langfuse_context, observe
 from prediction_market_agent_tooling.tools.parallelism import par_generator, par_map
-from prediction_market_agent_tooling.tools.tavily.tavily_storage import TavilyStorage
 from prediction_market_agent_tooling.tools.utils import (
     LLM_SUPER_LOW_TEMPERATURE,
     DatetimeUTC,
@@ -467,9 +466,6 @@ def generate_prediction_for_one_outcome(
             model=model,
             openai_api_key=api_keys.openai_api_key,
             tavily_api_key=api_keys.tavily_api_key,
-            tavily_storage=TavilyStorage(
-                agent_id=ThinkThoroughlyWithPredictionProphetResearch.identifier
-            ),
         )
         prediction = prophet_make_prediction(
             market_question=scenario,
@@ -509,7 +505,6 @@ def generate_final_decision(
                 model=self.model,
                 openai_api_key=api_keys.openai_api_key,
                 tavily_api_key=api_keys.tavily_api_key,
-                tavily_storage=TavilyStorage(agent_id=self.identifier),
             ).report
         )
         return super().generate_final_decision(

prediction_market_agent/tools/prediction_prophet/research.py

@@ -1,5 +1,4 @@
 from prediction_market_agent_tooling.loggers import logger
-from prediction_market_agent_tooling.tools.tavily.tavily_storage import TavilyStorage
 from prediction_prophet.benchmark.agents import (  # noqa: F401 # Just to make it available for the user of research.
     _make_prediction as prophet_make_prediction,
 )
@@ -13,7 +12,6 @@ def prophet_research(
     model: str,
     openai_api_key: SecretStr,
     tavily_api_key: SecretStr,
-    tavily_storage: TavilyStorage | None,
     initial_subqueries_limit: int = 20,
     subqueries_limit: int = 4,
     max_results_per_search: int = 5,
@@ -41,6 +39,5 @@ def prophet_research(
         min_scraped_sites=min_scraped_sites,
         openai_api_key=openai_api_key,
         tavily_api_key=tavily_api_key,
-        tavily_storage=tavily_storage,
         logger=logger,
     )

prediction_market_agent/tools/web_scrape/markdown.py

@@ -1,17 +1,18 @@
+from datetime import timedelta
+
 import requests
 import tenacity
 from bs4 import BeautifulSoup
 from markdownify import markdownify
 from prediction_market_agent_tooling.loggers import logger
-from prediction_market_agent_tooling.tools.cache import persistent_inmemory_cache
+from prediction_market_agent_tooling.tools.caches.db_cache import db_cache
 from prediction_market_agent_tooling.tools.langfuse_ import observe
 from requests import Response
 
 
 @tenacity.retry(
     stop=tenacity.stop_after_attempt(3), wait=tenacity.wait_fixed(1), reraise=True
 )
-@persistent_inmemory_cache
 def fetch_html(url: str, timeout: int) -> Response:
     headers = {
         "User-Agent": "Mozilla/5.0 (X11; Linux x86_64; rv:107.0) Gecko/20100101 Firefox/107.0"
@@ -21,6 +22,7 @@ def fetch_html(url: str, timeout: int) -> Response:
 
 
 @observe()
+@db_cache(max_age=timedelta(days=1))
 def web_scrape(url: str, timeout: int = 10) -> str:
     """
     Taken from agentcoinorg/predictionprophet

prediction_market_agent/tools/web_search/google.py

@@ -1,23 +1,9 @@
-import serpapi
+from prediction_market_agent_tooling.tools.google import search_google
 
-from prediction_market_agent import utils
-
-
-def google_search(query: str) -> list[str]:
-    params = {
-        "q": query,
-        "api_key": utils.APIKeys().serp_api_key.get_secret_value(),
-        "num": 4,
-    }
-    search = serpapi.GoogleSearch(params)
-    urls = [result["link"] for result in search.get_dict()["organic_results"]]
-    return urls
-
-
-google_search_schema = {
+search_google_schema = {
     "type": "function",
     "function": {
-        "name": "google_search",
+        "name": "search_google",
         "parameters": {
             "type": "object",
             "properties": {
@@ -35,5 +21,5 @@ def google_search(query: str) -> list[str]:
 
 class GoogleSearchTool:
     def __init__(self) -> None:
-        self.fn = google_search
-        self.schema = google_search_schema
+        self.fn = search_google
+        self.schema = search_google_schema

prediction_market_agent/utils.py

@@ -23,7 +23,6 @@ class DBKeys(BaseSettings):
 
 class APIKeys(APIKeysBase):
     # Don't get fooled! Serper and Serp are two different services.
-    SERP_API_KEY: t.Optional[SecretStr] = None
     SERPER_API_KEY: t.Optional[SecretStr] = None
     OPENAI_API_KEY: t.Optional[SecretStr] = None
     REPLICATE_API_KEY: t.Optional[SecretStr] = None
@@ -33,12 +32,6 @@ class APIKeys(APIKeysBase):
     PINATA_API_SECRET: t.Optional[SecretStr] = None
     TELEGRAM_BOT_KEY: t.Optional[SecretStr] = None
 
-    @property
-    def serp_api_key(self) -> SecretStr:
-        return check_not_none(
-            self.SERP_API_KEY, "SERP_API_KEY missing in the environment."
-        )
-
     @property
     def serper_api_key(self) -> SecretStr:
         return check_not_none(

pyproject.toml

@@ -17,7 +17,6 @@ numexpr = "*"
 pyautogen = "^0.2.32"
 flaml = "^2.2.0,<2.3.0" # Bug in newer version, because they started importing xgboost, without requiring it as a dependency.
 beautifulsoup4 = "*"
-serpapi = "*"
 google-search-results = "*"
 pytest = "*"
 llama-index = "~0.9.0"
@@ -35,7 +34,7 @@ poetry = "^1.7.1"
 poetry-plugin-export = "^1.6.0"
 functions-framework = "^3.5.0"
 cron-validator = "^1.0.8"
-prediction-market-agent-tooling = { version = "^0.55.1", extras = ["langchain", "google"] }
+prediction-market-agent-tooling = { version = "^0.56.0.dev112", extras = ["langchain", "google"] }
 pydantic-settings = "^2.1.0"
 autoflake = "^2.2.1"
 isort = "^5.13.2"
@@ -52,7 +51,7 @@ tweepy = "^4.14.0"
 langchain-pinecone = "^0.1.1"
 pinatapy-vourhey = "^0.2.0"
 plotly = "^5.22.0"
-prediction-prophet = { git = "https://github.com/agentcoinorg/predictionprophet.git", branch = "main" }
+prediction-prophet = { git = "https://github.com/agentcoinorg/predictionprophet.git", branch = "peter/latest-pmat-cache" }
 transformers = "^4.43.3"
 openfactverification-kongzii = "^0.2.0"
 

scripts/benchmark_ofv_resolver.py

@@ -1,6 +1,8 @@
 import pandas as pd
 import typer
-from prediction_market_agent_tooling.tools.cache import persistent_inmemory_cache
+from prediction_market_agent_tooling.tools.caches.inmemory_cache import (
+    persistent_inmemory_cache,
+)
 
 from prediction_market_agent.agents.ofvchallenger_agent.ofv_resolver import (
     ofv_answer_binary_question,

scripts/llm_randomness.py

@@ -4,7 +4,9 @@
 from langchain.prompts import ChatPromptTemplate
 from langchain_openai import ChatOpenAI
 from prediction_market_agent_tooling.config import APIKeys
-from prediction_market_agent_tooling.tools.cache import persistent_inmemory_cache
+from prediction_market_agent_tooling.tools.caches.inmemory_cache import (
+    persistent_inmemory_cache,
+)
 from prediction_market_agent_tooling.tools.utils import LLM_SUPER_LOW_TEMPERATURE