Import issues due to non compatibility with python 3.13

[Lukas-09]

Hi everyone,
I just ran into some issues while upgrading python to 3.13
it seems that versions of pandas >2.0.3 are running into problems that already occurred in an older thread "TypeError: Argument 'placement' has incorrect type (expected pandas._libs.internals.BlockPlacement, got slice)" .
The actual problem is that with python 3.12 and 3.13 you can't pip install pandas 2.0.3.
Moreover current versions of numpy(2.2.3) are not compatible with pandas 2.0.3

I don't know how complicated making all of that compatible would be but it surely would be helpful.

[darkrams98]

use your own dataset "csv" and convert it into pkl
I use this structure :

import pandas as pd
import pickle
from gym_mtsim.metatrader.symbol import SymbolInfo

df = (
pd.read_csv(
'XAU_1h_data.csv',
sep=';',
parse_dates=['Date'],
index_col='Date'
)
.sort_index()
[['Open','High','Low','Close','Volume']]
.tz_localize('UTC')
)

info = object.new(SymbolInfo)
for attr, val in {
'name': 'XAUUSD',
'market': 'Forex',
'currency_margin': 'XAU',
'currency_profit': 'USD',
'currencies': ('XAU','USD'),
'trade_contract_size':100000.0,
'margin_rate': 1.0,
'volume_min': 0.01,
'volume_max': 500.0,
'volume_step': 0.01
}.items():
setattr(info, attr, val)

symbols_info = {'XAUUSD': info}

symbols_data = {'XAUUSD': df}

with open('symbols_forex_new.pkl', 'wb') as f:
pickle.dump((symbols_info, symbols_data), f)

print("Done writing symbols_forex_new.pkl")