def suggest(table):
suggestion_log.info('Buff买Steam卖：\n')
for info, column in standard.items():
# buff往steam卖，steam - buff越大越好，所以最大的在前
sort_by_column(table, info, column, ascending=False)
suggestion_log.info('Steam买Buff卖：\n')
for info, column in standard.items():
# steam往buff卖，steam - buff越小越好，最好是负的，所以最小的在前
sort_by_column(table, info, column, ascending=True)
def suggest(table):
suggestion_log.info('buff买steam卖：\n')
for info, column in buff_to_steam_suggestions.items():
# buff往steam卖，steam - buff越大越好，所以最大的在前
sort_by_column(table, info, column, ascending=False)
suggestion_log.info('steam买buff卖：\n')
for info, column in steam_to_buff_suggestions.items():
# steam往buff卖，steam - buff越小越好，最好是负的，所以最小的在前
sort_by_column(table, info, column, ascending=True)
def filter_table(table):
"""
对table进行修正,去除交易量过小的和差价过大(通常是玄学货)的item
"""
table = table[table['gap_percent'] <= MAX_GAP_PERCENTAGE]
table = table[table['history_sold'] >= MIN_SOLD_THRESHOLD]
suggestion_log.info("经过对交易量和异常交易价格的过滤: \n{}\n".format(
MIN_SOLD_THRESHOLD, table.describe()))
return table
def filter_table(table):
# due to steam average history price is used!
# table = table[table['gap_percent'] <= MAX_GAP_PERCENTAGE]
# suggestion_log.info("After threshold(gap_percent <= {}) filtered: \n{}".format(MAX_GAP_PERCENTAGE, table.describe()))
table = table[table['history_sold'] >= MIN_SOLD_THRESHOLD]
suggestion_log.info(
"After threshold(history_sold >= {}) filtered: \n{}\n".format(
MIN_SOLD_THRESHOLD, table.describe()))
return table
def sort_by_column(table, suggestion, column, ascending=True):
suggestion_log.info(suggestion + "\n")
# filter
filtered_table = filter_table(table)
if ascending:
top = filtered_table.nsmallest(TOP_N, column)
# top = higher_price.sort_values(by=column, ascending=ascending).head(TOP_N)
else:
top = filtered_table.nlargest(TOP_N, column)
suggestion_log.info('收益降序：')
for item in converter.df_to_list(top):
suggestion_log.info(item.detail())
suggestion_log.info('\n')
def sort_by_column(table, suggestion, column, ascending=True):
filtered_table = filter_table(table)
if ascending:
top = heapq.nsmallest(config.TOP_N,
table,
key=lambda s: getattr(s, column))
else:
top = heapq.nlargest(config.TOP_N,
table,
key=lambda s: getattr(s, column))
suggestion_log.info(suggestion + '降序：')
for item in top:
suggestion_log.info(item.detail())
suggestion_log.info('\n')