julianandrews / pyeval7 Goto Github PK

Keep getting error for py3.12 when trying to pip install eval7

Hi mate,

I am always me.
I will try to explain my issue and why I edited a piece of your code, and how and why so if you need it for your AI you can reuse it.

First of all, I edited equity.pyx (I added the bold to the changes):

cdef tuple hand_vs_range_monte_carlo(unsigned long long hand, ...):
cdef unsigned int wins = 0
cdef unsigned int tie = 0

    if hero > villain:
        **wins += 1**
    elif hero == villain:
        **tie += 1**

**return <double>wins / <double>iterations, <double>tie / <double>iterations**

Let me explain, I encountered a lot of problems trying to clustering rivers combinations (hand + 5 cards) and I have tried many different solutions (your equity, machine learning tequines like k-mean, manually and a combination of all these mentioned above).
The first solution was simply creating 1002 clusters (from equity 0, ..., 0.342, 0.343, ..., 0.5, ..., 0.789, ... to 1.0) but the problem was a missing of board informations. For him a board like "sssss" (a flush on the board) with no flushes in the cards of the hand was good as a middle pair. And of course that's not good. You can't cluster these two type of hands because are too different (AI was really weak with that type of clustering), but if you offer him the percentage of tie he can detect better a lot of board patterns.

My post isn't a issue, it was only a discussion to have a comparison and get more ideas.

And thank you for all

My code is 50% C++ and 50% Python and I connect the first with the second troguht different hacks (es. I generate dinamically C++ code with Python and then I include it with #include, or I open the shell, call the python script and it reads the pipeline and so on).

One problem I had is to generate a great number of hands trought your library in a "acceptable" time and I want to share with you my three different solutions with relative benchmarks.

import time
import random
import numpy as np
import eval7 as e7

######################
bench = time.time()
for i in range(1000000):
    deck = e7.Deck()
    deck.shuffle()
    hands = deck.deal(4)
    board = deck.deal(5)
print("bench1:", time.time() - bench)

######################
bench = time.time()
deck = e7.Deck()
deck.shuffle()
cards = deck.deal(52)
for i in range(1000000):
    random.shuffle(cards)
    hands = cards[:4]
    board = cards[4:9]
print("bench2:", time.time() - bench)

######################
bench = time.time()
num_iters = 1000000
combos = np.random.randint(52, size=(num_iters, 9))
deck = e7.Deck()
cards = np.array(deck.deal(52))
for i in range(1000000):
    hands = cards[combos[i][:4]]
    board = cards[combos[i][4:]]
print("bench3:", time.time() - bench)

bench1: 177.55185198783875
bench2: 77.64555835723877
bench3: 18.77227473258972

Vanilla eval7 is the worst, the most "compatible" with simple edit of library code is the second one, and the last one is the most difficult to integrate and reuse it but the difference of performances has no rivals.

When I run something like
d = Deck()

I get the error:
E TypeError: 'cython_function_or_method' object does not support vectorcall

It works fine on Python 3.7 however.

This would be useful if I ever want a new windows distribution of FlopFerret

Hi, it looks great.

I want make easy software to train Range vs Range equity on various flops. I am not good programmer, but if I understand it, pyeval7 dont have this function.

I tried pokereval, but results are differents from results in poker softwares, is it difficult to implement? Or dont you have some advice about it?

Thanks,
Martin Psik

For some reason I can't import eval7, I've used pip install eval7 on latest windows 10, Python 3.8 (64-bit)

Right now I've only built wheels for python 3.7 on Linux

Right now I've just uploaded the python3.7 wheel, but with docker I should be able to build manylinux wheels for all python versions.

Hi mate,

I'm using your montecarlo to feeds my NN, but I noticed strange results. For example on

54sh KJT88shhhc

with a calculator for heads-up like a pokerstove I obtain 0% win an ~25% of tie.
With your code I obtain "only 12%". So, 2 questions:

1. Do you merge wins/draws? How?
2. Is it your MCC for heads-up?

Btw, I like very much your code, helped me a lot for different tasks.