L6

.gitignore

@@ -0,0 +1,3 @@
+venv
+__pycache__
+node_modules

src/app/config.py

@@ -0,0 +1,13 @@
+from datetime import datetime
+
+MARKET_1 = 'BTC'
+MARKET_2 = 'USD'
+
+START_TIME = datetime(2020, 4, 1)
+END_TIME = datetime(2020, 5, 21)
+
+PAST_DAYS_USED_TO_PREDICT = 7
+DAYS_TO_PREDICT = 4
+
+ITERATIONS = 100
+MUTATION_WEIGHT = 0.6

src/app/cryptocompare.py

@@ -0,0 +1,7 @@
+from app.utils import get_json_data
+
+def get_data(timestamp_from, timestamp_to, market_1, market_2):
+    url = f'https://min-api.cryptocompare.com/data/v2/histoday?fsym={market_1}&tsym={market_2}&allData=true'
+    data = get_json_data(url)['Data']['Data']
+    filtered_data = [d for d in data if d['time'] >= timestamp_from and d['time'] <= timestamp_to]
+    return filtered_data

src/app/plot.py

@@ -0,0 +1,47 @@
+import pandas as pd
+from datetime import datetime, timedelta
+import plotly.graph_objects as go
+from plotly.subplots import make_subplots
+from app.config import MARKET_1, MARKET_2, END_TIME
+
+def plot(df):
+    df['time'] = df['time'].apply(lambda x: datetime.fromtimestamp(x))
+
+    fig = make_subplots(rows=2, cols=2,
+                        subplot_titles=('Stock prices', 'Percentage change', 'Volume'))
+    fig.add_trace(
+        go.Candlestick(
+            name='Stock prices',
+            x=df['time'],
+            open=df['open'],
+            high=df['high'],
+            low=df['low'],
+            close=df['close'],
+        ),
+        row=1, col=1,
+    )
+    fig.add_trace(
+        go.Bar(
+            name='Percentage change',
+            x=df['time'],
+            y=df['per_ch'],
+        ),
+        row=1, col=2,
+    )
+    fig.add_trace(
+        go.Bar(
+            name='Volume',
+            x=df['time'],
+            y=df['volume'],
+        ),
+        row=2, col=1,
+    )
+
+    # Add line separating predictions from real values
+    line_date = END_TIME - timedelta(hours=12)
+    fig.update_layout(title_text=f'{MARKET_1} {MARKET_2} predictions',
+                      shapes = [dict(
+                          x0=line_date, x1=line_date, y0=0, y1=1, xref='x', yref='paper', line_width=2,
+                      )])
+
+    fig.show()

src/app/simulator.py

@@ -0,0 +1,61 @@
+import random
+import pandas as pd
+import numpy as np
+from datetime import datetime, timedelta
+
+from app import cryptocompare
+from app.config import PAST_DAYS_USED_TO_PREDICT, DAYS_TO_PREDICT, ITERATIONS, MUTATION_WEIGHT, END_TIME
+
+def decision(probability):
+    return random.random() < probability
+def get_adjusted_data(from_date, to_date, market_1, market_2):
+    data = cryptocompare.get_data(
+        datetime.timestamp(from_date),
+        datetime.timestamp(to_date),
+        market_1,
+        market_2,
+    )
+
+    for d in data:
+        d['per_ch'] = (d['close'] - d['open']) / d['open'] * 100.0
+        d['volume'] = (d['volumeto'] - d['volumefrom']) / 2
+        to_remove = ['volumefrom', 'volumeto', 'conversionType', 'conversionSymbol']
+        for r in to_remove:
+            del d[r]
+
+    return data
+
+def simulate(from_date, to_date, market_1, market_2):
+    data = get_adjusted_data(from_date, to_date, market_1, market_2)
+    df = pd.DataFrame(data)
+
+    day = END_TIME
+    for days in range(DAYS_TO_PREDICT):
+        print(f'Simulating: {day}')
+        analysis_data = df.tail(PAST_DAYS_USED_TO_PREDICT).drop(columns=['time'])
+        analysis_data_means = analysis_data.mean()
+
+        percentage_change = np.mean(np.absolute(analysis_data['per_ch']))
+        increase_probability = np.mean([1 if x > 0 else 0 for x in analysis_data['per_ch']])
+
+        simulation_data = analysis_data.iloc[0:0,:].copy()
+        for iteration in range(ITERATIONS):
+            should_increase = decision(increase_probability)
+            mutation = (1 if should_increase else -1) * random.random() * MUTATION_WEIGHT
+            predicted_values = analysis_data_means * (1 + percentage_change * mutation / 100)
+
+            predicted_values['per_ch'] = (predicted_values['close'] - predicted_values['open']) / predicted_values['open'] * 100.0
+
+            simulation_data = simulation_data.append(predicted_values, ignore_index = True)
+        print('Mean:')
+        print(simulation_data.mean())
+        print('Median:')
+        median = simulation_data.median()
+        print(median)
+        print('Deviation:')
+        print(simulation_data.std())
+
+        median['time'] = datetime.timestamp(day)
+        df = df.append(median, ignore_index = True)
+        day += timedelta(days=1)
+    return df

src/app/start_app.py

@@ -0,0 +1,7 @@
+from app import plot
+from app.config import START_TIME, END_TIME, MARKET_1, MARKET_2
+from app import simulator
+
+if __name__ == "__main__":
+    df = simulator.simulate(START_TIME, END_TIME, MARKET_1, MARKET_2)
+    plot.plot(df)

src/app/utils.py

@@ -0,0 +1,5 @@
+import urllib.request, json
+
+def get_json_data(url):
+    with urllib.request.urlopen(url) as u:
+        return json.loads(u.read().decode())

src/requirements.txt

@@ -0,0 +1,3 @@
+pandas==1.0.3
+plotly==4.8.0
+numpy==1.18.3

src/run.sh

@@ -0,0 +1,4 @@
+#!/bin/bash -e
+cd "$(dirname "$0")"
+
+python -m app.start_app

src/setupenv.sh

@@ -0,0 +1,8 @@
+#!/bin/bash -e
+cd "$(dirname "$0")"
+
+python3 -m venv venv &&
+  source venv/bin/activate &&
+  pip install --upgrade pip &&
+  pip install -r requirements.txt &&
+  deactivate