Seer's `get_liquidity_for_outcome` should be `get_liquidity_for_outcome_and_type`

[kongzii]

Right now, we check if some market's outcome is tradeable in

prediction-market-agent-tooling/prediction_market_agent_tooling/markets/seer/seer.py

Line 673 in 3dad902

if not self.has_liquidity():

.

This checks if all outcomes have some liquidity.

In there, we consider both buying and selling:

prediction-market-agent-tooling/prediction_market_agent_tooling/markets/seer/seer.py

Line 566 in 3dad902

for token_address in [pool.token0.id, pool.token1.id]:

, and we succeed if any of them is available.

However, we should check only liquidity for the type of trade we are going to do. E.g. if we are going to be buying, we need to check liquidity for buying. Otherwise it will error out, as it just happened with market https://app.seer.pm/markets/100/will-elon-musk-be-ceo-of-tesla-on-1-september-2025/?outcome=Yes

(credits for graphics to @gabrielfior 😄 )

[gabrielfior]

In my head, the easiest implementation is to simulate the transaction we would do when swapping, exact_input_single (see below - contract link)

/// @inheritdoc ISwapRouter
    function exactInputSingle(ExactInputSingleParams calldata params)
        external
        payable
        override
        checkDeadline(params.deadline)
        returns (uint256 amountOut)
    {
        amountOut = exactInputInternal(
            params.amountIn,
            params.recipient,
            params.limitSqrtPrice,
            SwapCallbackData({path: abi.encodePacked(params.tokenIn, params.tokenOut), payer: msg.sender})
        );
        require(amountOut >= params.amountOutMinimum, 'Too little received');
    }

hence, if we simulate, we get the amountOut we would get in return for the transaction.
Now, if

1. transaction throws errors -> don't trade
2. transaction returns amountOut < amount of sDAI transferred in -> log and don't trade
3. else, do transfer

I expect the above to cover this liquidity gap and avoid errors.