prediction-market-agent-tooling 0.33.0__py3-none-any.whl → 0.35.0__py3-none-any.whl

prediction_market_agent_tooling/markets/manifold/data_models.py

@@ -79,7 +79,7 @@ class ManifoldMarket(BaseModel):
         """
         Clip the timestamp to the maximum valid timestamp.
         """
-        max_timestamp = (datetime.max - timedelta(hours=1)).timestamp()
+        max_timestamp = (datetime.max - timedelta(days=1)).timestamp()
         value = int(min(value / 1000, max_timestamp))
         return datetime.fromtimestamp(value)
 

prediction_market_agent_tooling/markets/omen/omen.py

@@ -10,6 +10,7 @@ from prediction_market_agent_tooling.gtypes import (
     ChecksumAddress,
     HexAddress,
     HexStr,
+    OmenOutcomeToken,
     OutcomeStr,
     Probability,
     Wei,
@@ -54,7 +55,10 @@ from prediction_market_agent_tooling.markets.omen.omen_subgraph_handler import (
 )
 from prediction_market_agent_tooling.tools.balances import get_balances
 from prediction_market_agent_tooling.tools.hexbytes_custom import HexBytes
-from prediction_market_agent_tooling.tools.utils import check_not_none
+from prediction_market_agent_tooling.tools.utils import (
+    calculate_sell_amount_in_collateral,
+    check_not_none,
+)
 from prediction_market_agent_tooling.tools.web3_utils import (
     add_fraction,
     remove_fraction,
@@ -80,6 +84,8 @@ class OmenAgentMarket(AgentMarket):
     finalized_time: datetime | None
     created_time: datetime
     close_time: datetime
+    outcome_token_amounts: list[OmenOutcomeToken]
+    fee: float  # proportion, from 0 to 1
 
     INVALID_MARKET_ANSWER: HexStr = HexStr(
         "0xFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF"
@@ -155,19 +161,50 @@ class OmenAgentMarket(AgentMarket):
             web3=web3,
         )
 
+    def calculate_sell_amount_in_collateral(
+        self, amount: TokenAmount, outcome: bool
+    ) -> xDai:
+        if len(self.outcome_token_amounts) != 2:
+            raise ValueError(
+                f"Market {self.id} has {len(self.outcome_token_amounts)} "
+                f"outcomes. This method only supports binary markets."
+            )
+        sell_index = self.yes_index if outcome else self.no_index
+        other_index = self.no_index if outcome else self.yes_index
+        collateral = calculate_sell_amount_in_collateral(
+            shares_to_sell=amount.amount,
+            holdings=wei_to_xdai(Wei(self.outcome_token_amounts[sell_index])),
+            other_holdings=wei_to_xdai(Wei(self.outcome_token_amounts[other_index])),
+            fee=self.fee,
+        )
+        return xDai(collateral)
+
     def sell_tokens(
-        self, outcome: bool, amount: TokenAmount, auto_withdraw: bool = False
+        self,
+        outcome: bool,
+        amount: TokenAmount,
+        auto_withdraw: bool = False,
+        api_keys: APIKeys | None = None,
+        web3: Web3 | None = None,
     ) -> None:
         if not self.can_be_traded():
             raise ValueError(
                 f"Market {self.id} is not open for trading. Cannot sell tokens."
             )
+
+        # Convert from token (i.e. share) number to xDai value of tokens, as
+        # this is the expected unit of the argument in the smart contract.
+        collateral = self.calculate_sell_amount_in_collateral(
+            amount=amount,
+            outcome=outcome,
+        )
         binary_omen_sell_outcome_tx(
-            api_keys=APIKeys(),
-            amount=xDai(amount.amount),
+            amount=collateral,
+            api_keys=api_keys if api_keys is not None else APIKeys(),
             market=self,
             binary_outcome=outcome,
             auto_withdraw=auto_withdraw,
+            web3=web3,
         )
 
     def was_any_bet_outcome_correct(
@@ -245,6 +282,8 @@ class OmenAgentMarket(AgentMarket):
             url=model.url,
             volume=wei_to_xdai(model.collateralVolume),
             close_time=model.close_time,
+            outcome_token_amounts=model.outcomeTokenAmounts,
+            fee=float(wei_to_xdai(model.fee)) if model.fee is not None else 0.0,
         )
 
     @staticmethod

prediction_market_agent_tooling/monitor/monitor.py

@@ -185,7 +185,7 @@ def monitor_agent(agent: DeployedAgent) -> None:
     col3.markdown(f"Public ID: `{agent.public_id}`")
 
     show_agent_bets = st.checkbox(
-        "Show resolved bets", value=False, key=f"{agent.name}_show_bets"
+        "Show resolved bets statistics", value=False, key=f"{agent.name}_show_bets"
     )
     if not show_agent_bets:
         return
@@ -203,16 +203,43 @@ def monitor_agent(agent: DeployedAgent) -> None:
         "Is Correct": [bet.is_correct for bet in agent_bets],
         "Profit": [round(bet.profit.amount, 2) for bet in agent_bets],
     }
-    bets_df = pd.DataFrame(bets_info).sort_values(by="Resolved Time")
+
+    # Time column to use for x-axes and sorting
+    x_axis_column = st.selectbox(
+        "X-axis column",
+        ["Created Time", "Resolved Time"],
+        key=f"{agent.name}_x_axis_column",
+    )
+
+    bets_df = pd.DataFrame(bets_info).sort_values(by=x_axis_column, ascending=False)
+    bets_df["x-axis-day"] = bets_df[x_axis_column].dt.date
 
     # Metrics
     col1, col2 = st.columns(2)
     col1.metric(label="Number of bets", value=f"{len(agent_bets)}")
     col2.metric(label="% Correct", value=f"{100 * bets_df['Is Correct'].mean():.2f}%")
 
+    # Chart of grouped accuracy per day
+    per_day_accuracy = bets_df.groupby("x-axis-day")["Is Correct"].mean()
+    per_day_accuracy_chart = (
+        alt.Chart(per_day_accuracy.reset_index())
+        .encode(
+            x=alt.X("x-axis-day", axis=alt.Axis(format="%Y-%m-%d"), title=None),
+            y=alt.Y("Is Correct", axis=alt.Axis(format=".2f")),
+        )
+        .interactive()
+    )
+    st.altair_chart(
+        per_day_accuracy_chart.mark_line()
+        + per_day_accuracy_chart.transform_loess("x-axis-day", "Is Correct").mark_line(
+            color="red", strokeDash=[5, 5]
+        ),
+        use_container_width=True,
+    )
+
     # Chart of cumulative profit per day
     profit_info = {
-        "Time": bets_df["Resolved Time"],
+        "Time": bets_df[x_axis_column],
         "Cumulative Profit": bets_df["Profit"].astype(float),
     }
     profit_df = pd.DataFrame(profit_info)
@@ -233,8 +260,12 @@ def monitor_agent(agent: DeployedAgent) -> None:
     )
 
     # Table of resolved bets
-    st.subheader("Resolved Bet History")
-    st.table(bets_df)
+    show_bet_history = st.checkbox(
+        "Show resolved bet history", value=False, key=f"{agent.name}_show_bet_history"
+    )
+    if show_bet_history:
+        st.subheader("Resolved Bet History")
+        st.table(bets_df.drop(columns=["x-axis-day"]))
 
 
 def monitor_market(

prediction_market_agent_tooling/tools/utils.py

@@ -7,6 +7,7 @@ from typing import Any, NoReturn, Optional, Type, TypeVar, cast
 import pytz
 import requests
 from pydantic import BaseModel, ValidationError
+from scipy.optimize import newton
 from scipy.stats import entropy
 
 from prediction_market_agent_tooling.gtypes import (
@@ -173,3 +174,35 @@ def prob_uncertainty(prob: Probability) -> float:
         - Market's probability is 0.95, so the market is quite certain about YES: prob_uncertainty(0.95) == 0.286
     """
     return float(entropy([prob, 1 - prob], base=2))
+
+
+def calculate_sell_amount_in_collateral(
+    shares_to_sell: float,
+    holdings: float,
+    other_holdings: float,
+    fee: float,
+) -> float:
+    """
+    Computes the amount of collateral that needs to be sold to get `shares`
+    amount of shares. Returns None if the amount can't be computed.
+
+    Taken from https://github.com/protofire/omen-exchange/blob/29d0ab16bdafa5cc0d37933c1c7608a055400c73/app/src/util/tools/fpmm/trading/index.ts#L99
+    Simplified for binary markets.
+    """
+
+    if not (0 <= fee < 1.0):
+        raise ValueError("Fee must be between 0 and 1")
+
+    for v in [shares_to_sell, holdings, other_holdings]:
+        if v <= 0:
+            raise ValueError("All share args must be greater than 0")
+
+    def f(r: float) -> float:
+        R = r / (1 - fee)
+        first_term = other_holdings - R
+        second_term = holdings + shares_to_sell - R
+        third_term = holdings * other_holdings
+        return (first_term * second_term) - third_term
+
+    amount_to_sell = newton(f, 0)
+    return float(amount_to_sell) * 0.999999  # Avoid rounding errors

{prediction_market_agent_tooling-0.33.0.dist-info → prediction_market_agent_tooling-0.35.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: prediction-market-agent-tooling
-Version: 0.33.0
+Version: 0.35.0
 Summary: Tools to benchmark, deploy and monitor prediction market agents.
 Author: Gnosis
 Requires-Python: >=3.10,<3.12
@@ -35,7 +35,7 @@ Requires-Dist: streamlit (>=1.31.0,<2.0.0)
 Requires-Dist: subgrounds (>=1.8.1,<2.0.0)
 Requires-Dist: tabulate (>=0.9.0,<0.10.0)
 Requires-Dist: tqdm (>=4.66.2,<5.0.0)
-Requires-Dist: typer (>=0.9.0,<0.10.0)
+Requires-Dist: typer (>=0.12.3,<0.13.0)
 Requires-Dist: types-pytz (>=2024.1.0.20240203,<2025.0.0.0)
 Requires-Dist: types-requests (>=2.31.0.0,<3.0.0.0)
 Requires-Dist: web3 (>=6.15.1,<7.0.0)

{prediction_market_agent_tooling-0.33.0.dist-info → prediction_market_agent_tooling-0.35.0.dist-info}/RECORD

@@ -25,13 +25,13 @@ prediction_market_agent_tooling/markets/categorize.py,sha256=yTd-lDMPW4ESDSzmxeL
 prediction_market_agent_tooling/markets/data_models.py,sha256=uODY3aoFp8YYeLAUcrzMk1yU8pIKsTLobB9xIEGTmKs,1170
 prediction_market_agent_tooling/markets/manifold/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 prediction_market_agent_tooling/markets/manifold/api.py,sha256=m6qOzDiyQfxj62Eo_SzzQLkX4ijpi8KtQKGd4CpKAsk,7307
-prediction_market_agent_tooling/markets/manifold/data_models.py,sha256=2DZIxwtDp-PH0UWTGiktMFIGAAQoVutI07UsxjNyTyE,5296
+prediction_market_agent_tooling/markets/manifold/data_models.py,sha256=Iw-ajDPYhtfcTfN45nJN_kZShfWcWe9mgzexSWnQCCQ,5295
 prediction_market_agent_tooling/markets/manifold/manifold.py,sha256=EwRL06E2Y_ZAzr8efwS5yD6p6rnykrcBhqmNDUGZ8Aw,4075
 prediction_market_agent_tooling/markets/manifold/utils.py,sha256=cPPFWXm3vCYH1jy7_ctJZuQH9ZDaPL4_AgAYzGWkoow,513
 prediction_market_agent_tooling/markets/markets.py,sha256=w05Oo7yCA2axpCw69Q9y4i9Gcdpht0u5bZGbWqld3rU,2964
 prediction_market_agent_tooling/markets/omen/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 prediction_market_agent_tooling/markets/omen/data_models.py,sha256=EXtjmcujx68Xu50BVkYXvLuf_Asx5o65RvFS3ZS6HGs,14405
-prediction_market_agent_tooling/markets/omen/omen.py,sha256=9le-tvs-zclHtzpvK7bMVqg2lZV-VZh1c5OB_uAdSiY,32042
+prediction_market_agent_tooling/markets/omen/omen.py,sha256=aJ5CCste61B7q0OD3EjLP8H7AjD_RaBBNyUfnZGh5KM,33597
 prediction_market_agent_tooling/markets/omen/omen_contracts.py,sha256=eDS8vN4Klv_-Y1wwfIeLDt3twhl9U_AJjPQov0JScb0,19888
 prediction_market_agent_tooling/markets/omen/omen_resolving.py,sha256=g77QsQ5WnSI2rzBlX87L_EhWMwobkyXyfRhHQmpAdzo,9012
 prediction_market_agent_tooling/markets/omen/omen_subgraph_handler.py,sha256=RlaQMDF75lo2fos8LUb_OAdrw_ILLOXQLdejzE7tOmA,23053
@@ -44,7 +44,7 @@ prediction_market_agent_tooling/monitor/langfuse/langfuse_wrapper.py,sha256=b6T6
 prediction_market_agent_tooling/monitor/markets/manifold.py,sha256=GdYpgRX1GahDi-75Mr53jgtEg6nWcs_rHDUkg4o_7dQ,3352
 prediction_market_agent_tooling/monitor/markets/omen.py,sha256=jOLPnIbDU9syjnYtHfOb2xa6-Ize3vbplgh-8WWkuT4,3323
 prediction_market_agent_tooling/monitor/markets/polymarket.py,sha256=I9z9aO1wncyGI3a09ihrw17JkeBKjAuMmC0I9pl_9o4,1781
-prediction_market_agent_tooling/monitor/monitor.py,sha256=elvCqCASCd3fOnM7f2kajb_Oy3RMUmGh88jzpworrEU,13387
+prediction_market_agent_tooling/monitor/monitor.py,sha256=uJdaEpGWp4VcejoQLF0uOxhV75AhrXNyWPVG38gBeOo,14505
 prediction_market_agent_tooling/monitor/monitor_app.py,sha256=rDxgdDJqSK0ARx0TJFMkS76npkHZJz0__Rp0xTpiRok,4611
 prediction_market_agent_tooling/monitor/monitor_settings.py,sha256=Xiozs3AsufuJ04JOe1vjUri-IAMWHjjmc2ugGGiHNH4,947
 prediction_market_agent_tooling/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -64,10 +64,10 @@ prediction_market_agent_tooling/tools/parallelism.py,sha256=8mgkF5sBwFGS5GMvlpza
 prediction_market_agent_tooling/tools/safe.py,sha256=h0xOO0eNtitClf0fPkn-0oTc6A_bflDTee98V_aiV-A,5195
 prediction_market_agent_tooling/tools/singleton.py,sha256=CiIELUiI-OeS7U7eeHEt0rnVhtQGzwoUdAgn_7u_GBM,729
 prediction_market_agent_tooling/tools/streamlit_user_login.py,sha256=NXEqfjT9Lc9QtliwSGRASIz1opjQ7Btme43H4qJbzgE,3010
-prediction_market_agent_tooling/tools/utils.py,sha256=zkmwPi3YisgZDPCeNwaRbL8sInOYOkvFgFY4Q8PbEo4,5077
+prediction_market_agent_tooling/tools/utils.py,sha256=-G22UEbCRm59bm1RWFdeF55hRsaxgwZVAHvK32-Ye1g,6190
 prediction_market_agent_tooling/tools/web3_utils.py,sha256=cboATXNmEdn5RmPbVblHOwOdUMKBYrUK3GiI6i6Vzxo,9855
-prediction_market_agent_tooling-0.33.0.dist-info/LICENSE,sha256=6or154nLLU6bELzjh0mCreFjt0m2v72zLi3yHE0QbeE,7650
-prediction_market_agent_tooling-0.33.0.dist-info/METADATA,sha256=m8GpvhogLrDzkiUsfmt5CV1YBhIdTrgn8bRaVb98kqE,7514
-prediction_market_agent_tooling-0.33.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-prediction_market_agent_tooling-0.33.0.dist-info/entry_points.txt,sha256=m8PukHbeH5g0IAAmOf_1Ahm-sGAMdhSSRQmwtpmi2s8,81
-prediction_market_agent_tooling-0.33.0.dist-info/RECORD,,
+prediction_market_agent_tooling-0.35.0.dist-info/LICENSE,sha256=6or154nLLU6bELzjh0mCreFjt0m2v72zLi3yHE0QbeE,7650
+prediction_market_agent_tooling-0.35.0.dist-info/METADATA,sha256=PGIW673s30z0PdKLKcyhroDIjsBJWwzuZ7S-2bqIzyU,7515
+prediction_market_agent_tooling-0.35.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+prediction_market_agent_tooling-0.35.0.dist-info/entry_points.txt,sha256=m8PukHbeH5g0IAAmOf_1Ahm-sGAMdhSSRQmwtpmi2s8,81
+prediction_market_agent_tooling-0.35.0.dist-info/RECORD,,