foo.py

#!/usr/bin/python
# -*- coding: utf-8 -*-
"""
Use data already downloaded to recover the bid ask prices of differet
instruments

@author: ucaiado

Created on 05/02/2018
"""

from market_gym.lob import matching_engine, book
from market_gym.envs import make
import market_gym.utils.di_utilities as di_utilities


'''
Begin help functions
'''


'''
End help functions
'''


def collect_book_info(l_files, l_instrument, s_main_intrument, l_du, l_pu,
                      l_price_adj, f_time=30):
    '''
    Collect book information in 30 seconds intervals and return the environment
    simulated

    :param l_files: list. Name of the files of each instrument
    :param l_instrument: list. list of intruments in the simulation
    :param s_main_intrument: string. main instrument of simulation
    :param l_du: list. time to maturity, in days
    :param l_pu: list. settlement prices
    :param l_price_adj: list. rates from settlement prices
    :param *f_time: float. Time in seconds to collect the book
    '''
    d_rtn = {'time': [], 'TOB': []}
    # set up the stop time object
    l_h = [9, 10, 11, 12, 13, 14, 15, 16]
    NextStopTime = matching_engine.NextStopTime(l_h, i_milis=500.)
    obj_env = make('YieldCurve')
    e = obj_env(l_fname=l_files,
                l_instrument=l_instrument,
                NextStopTime=NextStopTime,
                s_main_intrument=s_main_intrument,
                l_du=l_du,
                l_pu=l_pu,
                l_price_adj=l_price_adj)
    f_next_check = e.order_matching.f_time + f_time
    b_quit = False
    while True:
        try:
            e.step()
            if e.order_matching.f_time > f_next_check:
                f_next_check = e.order_matching.f_time + f_time
                try:
                    d_book = {}
                    for s_instr in e.l_instrument:
                        # get best prices

                        # AttributeError: 'YieldCrvEnv' object has no attribute 'book'
                        testBook = e.book.get_n_top_prices(10, True)
                        ##############################################################
                        book_aux = e.order_matching.get_order_book_obj(s_instr)
                        d_book[s_instr] = {'qBid': book_aux.best_bid[1],
                                           'Bid': book_aux.best_bid[0],
                                           'qAsk': book_aux.best_ask[1],
                                           'Ask': book_aux.best_ask[0]}
                        print(testBook[0])
                    d_rtn['time'].append(e.order_matching.f_time)
                    d_rtn['TOB'].append(d_book)
                except TypeError:
                    pass
        except StopIteration:
            b_quit = True
        finally:

            if e.done or b_quit:
                break

    return e, d_rtn


if __name__ == '__main__':
    l_instrument=['DI1F18']
    s_main_intrument='DI1F18'
    s_root = 'data/preprocessed/'
    s_date = '20170201'
    l_files = [s_root + s_date + '/' + s_date + '_{}_{}_new.zip']
    l_du = []
    l_pu = []
    l_price_adj = []
    l_du, l_pu, l_price_adj = di_utilities.get_param_to_env(s_date, l_instrument)

    collect_book_info(l_files, l_instrument, s_main_intrument, l_du, l_pu, l_price_adj)

Opa. Não é o Environment que tem este método, é um book específico (https://github.com/ucaiado/rl_trading/blob/master/market_gym/lob/book.py#L672). Então, neste caso:

    # testBook = e.book.get_n_top_prices(10, True)
    book_aux = e.order_matching.get_order_book_obj(s_instr)
    df = book_aux.get_n_top_prices(10, True)
    print(df)