L6

src/apiBroker.py

@@ -0,0 +1,17 @@
+import requests
+
+
+def get_all_data(trading_pair):
+    (fst, snd) = trading_pair.split("-")
+    url = f"https://min-api.cryptocompare.com/data/v2/histoday?fsym={fst}&tsym={snd}&allData=true"
+    resp = requests.get(url)
+    if resp.status_code != 200:
+        raise ConnectionError('Error during fetching the data, Error code: {}'.format(resp.status_code))
+    else:
+        return resp.json()
+
+
+def get_data(trading_pair, timestamp_from, timestamp_to):
+    data = get_all_data(trading_pair)
+    return list(
+        filter(lambda data: float(timestamp_from) <= float(data["time"]) <= float(timestamp_to), data["Data"]["Data"]))

src/graphs.py

@@ -0,0 +1,24 @@
+import apiBroker
+import matplotlib.pyplot as plt
+import pandas as pd
+
+
+def draw(data, title):
+
+    fig = plt.figure()
+    plt.style.use('seaborn')
+    ax1 = fig.add_subplot(211)
+    ax2 = fig.add_subplot(212)
+    dataset = pd.DataFrame(data)
+    dataset['time'] = pd.to_datetime(dataset['time'], unit='s')
+    ax1 = dataset.plot(kind='line' , x='time', y='open', title=title, ax=ax1)
+    ax1.set_xlabel("time")
+    ax1.set_ylabel("price")
+    ax1.legend(['price'])
+    ax2 = dataset.plot(kind="bar", x='time', y='volumefrom', stacked=True, ax=ax2)
+    ax2.legend(['volume'])
+    ax2.set_xlabel("time")
+    ax2.set_xticks([])
+    ax2.set_ylabel("volume")
+
+    plt.draw()

src/simulation.py

@@ -0,0 +1,111 @@
+import apiBroker
+import time
+import datetime
+import numpy as np
+import pandas as pd
+import graphs
+
+
+pairs = [
+    "BTC-USD",
+    "ETH-USD",
+    "LTC-USD"
+]
+
+
+# YYYY-MM_DD format
+def convert_dates(human_readable_date):
+    (yyyy, mm, dd) = (human_readable_date.split("-"))
+    date = datetime.datetime(int(yyyy), int(mm), int(dd))
+    return int(date.timestamp())
+
+
+def sim(trading_pair, starting_date, ending_date, count):
+    data_list = []
+    for _ in range(count):
+        data_list.append(sim_one(trading_pair, starting_date, ending_date))
+    return data_list
+
+
+def sim_one(trading_pair, starting_date, ending_date):
+    starting_date = convert_dates(starting_date)
+    ending_date = convert_dates(ending_date)
+    if ending_date >= time.time() or starting_date >= ending_date:
+        raise ValueError("Wrong date range")
+    if trading_pair not in pairs:
+        raise ValueError("Wrong trading pair")
+    data = pd.DataFrame(apiBroker.get_data(trading_pair, starting_date, ending_date))
+    data = data[['time', 'open', 'close', 'volumefrom']]
+    days = int((ending_date - starting_date) / (24 * 60 * 60))
+    data['months'] = int(days / 30)
+    data['months'] = data['months'].clip(lower=1)
+    for i in range(days):
+        data = data.append(generate_entry(data), ignore_index=True)
+    data = data.drop(columns=['months', 'diff'])
+    return data
+
+
+def generate_entry(data):
+    price_change = (decide_how_price_changes(data, count_rise_probability(data)))
+    new_volume = decide_volume_change(data, price_change)
+    last_entry = data.tail(1)
+    time = last_entry['time'] + (24 * 60 * 60)
+    open = last_entry['close']
+    close = open + price_change
+    diff = np.abs(price_change)
+    new_entry = pd.DataFrame({'time': time, 'open': open, 'volumefrom': new_volume, 'close': close, 'diff': diff})
+    return new_entry
+
+
+def decide_volume_change(data, price_change):
+    diff = np.abs(price_change)
+    result_index = data['diff'].sub(diff).abs().idxmin()
+    return data.iloc[result_index, :]['volumefrom']
+
+
+def count_rise_probability(data):
+    data['diff'] = data['open'] - data['close']
+    total_change_in_price = np.abs(data['diff']).sum()
+    return (data['diff'][data['diff'] > 0]).sum() / total_change_in_price
+
+
+def decide_how_price_changes(data, probability_of_rise):
+    diffs = data['diff'].abs()
+    p = np.random.uniform(diffs.min(), diffs.max()) / data.iloc[0]['months']
+    if np.random.rand() < probability_of_rise:
+        return p
+    else:
+        a = data.tail(1).iloc[0]['close'] - p
+        if a - p < 0:
+            return np.abs(a)
+        else:
+            return -p
+
+
+def count_statistics(data):
+    return {'averages': data.mean(), 'median': data.median(), 'deviation': data.std()}
+
+
+def mean_from_multiple_sims(data_list):
+    combined = pd.concat(data_list)
+    means = combined.groupby(combined.index).mean()
+    return means
+
+
+if __name__ == '__main__':
+    print('Pick a pair')
+    for i in range(3):
+        print(f"{i} - {pairs[i]}")
+    pair = input()
+    start_date = input("Enter start date(YYYY-MM-DD): ")
+    end_date = input("Enter end date(YYYY-MM-DD): ")
+    one = sim_one(pairs[int(pair)], start_date, end_date)
+    graphs.draw(one, 'Simulation')
+    stats = count_statistics(one)
+    print("Statistics from simulation: ")
+    for stat in stats:
+        print(f"{stat}:")
+        print(stats[stat])
+    data = sim(pairs[int(pair)], start_date, end_date, 100)
+    graphs.draw(mean_from_multiple_sims(data), 'Avg from 100 simulations')
+    graphs.plt.show()