Merge pull request #124 from hummingbot/feat/config_generator

(feat) position builder | config generator

main.py

@@ -14,6 +14,7 @@ def main_page():
  Page("pages/master_conf/app.py", "Credentials", "🗝️"),
  Page("pages/bot_orchestration/app.py", "Instances", "🦅"),
  Page("pages/file_manager/app.py", "File Explorer", "🗂"),
+ Page("pages/position_builder/app.py", "Position Builder", "🔭"),
  Section("Backtest Manager", "⚙️"),
  Page("pages/backtest_get_data/app.py", "Get Data", "💾"),
  Page("pages/backtest_create/create.py", "Create", "⚔️"),

pages/position_builder/app.py

@@ -0,0 +1,266 @@
+from math import exp
+import streamlit as st
+from plotly.subplots import make_subplots
+import plotly.graph_objects as go
+from decimal import Decimal
+import yaml
+
+from utils.st_utils import initialize_st_page
+from hummingbot.smart_components.utils.distributions import Distributions
+
+# Initialize the Streamlit page
+initialize_st_page(title="Position Generator", icon="🔭", initial_sidebar_state="collapsed")
+
+# Page content
+st.text("This tool will help you analyze and generate a position config.")
+st.write("---")
+
+# Layout in columns
+col_quote, col_tp_sl, col_levels, col_spread_dist, col_amount_dist = st.columns([1, 1, 1, 2, 2])
+
+
+def normalize(values):
+ total = sum(values)
+ return [val / total for val in values]
+
+
+def convert_to_yaml(spreads, order_amounts):
+ data = {
+ 'dca_spreads': [float(spread)/100 for spread in spreads],
+ 'dca_amounts': [float(amount) for amount in order_amounts]
+ }
+ return yaml.dump(data, default_flow_style=False)
+
+
+with col_quote:
+ total_amount_quote = st.number_input("Total amount of quote", value=1000)
+
+with col_tp_sl:
+ tp = st.number_input("Take Profit (%)", min_value=0.0, max_value=100.0, value=2.0, step=0.1)
+ sl = st.number_input("Stop Loss (%)", min_value=0.0, max_value=100.0, value=8.0, step=0.1)
+
+with col_levels:
+ n_levels = st.number_input("Number of Levels", min_value=1, value=5)
+
+
+def distribution_inputs(column, dist_type_name):
+ if dist_type_name == "Spread":
+ dist_type = column.selectbox(
+ f"Type of {dist_type_name} Distribution",
+ ("GeoCustom", "Geometric", "Fibonacci", "Manual", "Logarithmic", "Arithmetic"),
+ key=f"{dist_type_name.lower()}_dist_type",
+ # Set the default value
+ )
+ else:
+ dist_type = column.selectbox(
+ f"Type of {dist_type_name} Distribution",
+ ("Geometric", "Fibonacci", "Manual", "Logarithmic", "Arithmetic"),
+ key=f"{dist_type_name.lower()}_dist_type",
+ # Set the default value
+ )
+ base, scaling_factor, step, ratio, manual_values = None, None, None, None, None
+
+ if dist_type != "Manual":
+ start = column.number_input(f"{dist_type_name} Start Value", value=1.0, key=f"{dist_type_name.lower()}_start")
+ if dist_type == "Logarithmic":
+ base = column.number_input(f"{dist_type_name} Log Base", value=exp(1), key=f"{dist_type_name.lower()}_base")
+ scaling_factor = column.number_input(f"{dist_type_name} Scaling Factor", value=2.0, key=f"{dist_type_name.lower()}_scaling")
+ elif dist_type == "Arithmetic":
+ step = column.number_input(f"{dist_type_name} Step", value=0.1, key=f"{dist_type_name.lower()}_step")
+ elif dist_type == "Geometric":
+ ratio = column.number_input(f"{dist_type_name} Ratio", value=2.0, key=f"{dist_type_name.lower()}_ratio")
+ elif dist_type == "GeoCustom":
+ ratio = column.number_input(f"{dist_type_name} Ratio", value=2.0, key=f"{dist_type_name.lower()}_ratio")
+ else:
+ manual_values = [column.number_input(f"{dist_type_name} for level {i+1}", value=1.0, key=f"{dist_type_name.lower()}_{i}") for i in range(n_levels)]
+ start = None # As start is not relevant for Manual type
+
+ return dist_type, start, base, scaling_factor, step, ratio, manual_values
+
+
+# Spread and Amount Distributions
+spread_dist_type, spread_start, spread_base, spread_scaling, spread_step, spread_ratio, manual_spreads = distribution_inputs(col_spread_dist, "Spread")
+amount_dist_type, amount_start, amount_base, amount_scaling, amount_step, amount_ratio, manual_amounts = distribution_inputs(col_amount_dist, "Amount")
+
+
+def get_distribution(dist_type, n_levels, start, base=None, scaling_factor=None, step=None, ratio=None, manual_values=None):
+ if dist_type == "Manual":
+ return manual_values
+ elif dist_type == "Linear":
+ return Distributions.linear(n_levels, start, start + tp)
+ elif dist_type == "Fibonacci":
+ return Distributions.fibonacci(n_levels, start)
+ elif dist_type == "Logarithmic":
+ return Distributions.logarithmic(n_levels, base, scaling_factor, start)
+ elif dist_type == "Arithmetic":
+ return Distributions.arithmetic(n_levels, start, step)
+ elif dist_type == "Geometric":
+ return Distributions.geometric(n_levels, start, ratio)
+ elif dist_type == "GeoCustom":
+ return [Decimal("0")] + Distributions.geometric(n_levels - 1, start, ratio)
+
+spread_distribution = get_distribution(spread_dist_type, n_levels, spread_start, spread_base, spread_scaling, spread_step, spread_ratio, manual_spreads)
+amount_distribution = normalize(get_distribution(amount_dist_type, n_levels, amount_start, amount_base, amount_scaling, amount_step, amount_ratio, manual_amounts))
+order_amounts = [Decimal(amount_dist * total_amount_quote) for amount_dist in amount_distribution]
+spreads = [Decimal(spread - spread_distribution[0]) for spread in spread_distribution]
+
+
+# Export Button
+if st.button('Export as YAML'):
+ yaml_data = convert_to_yaml(spreads, order_amounts)
+ st.download_button(
+ label="Download YAML",
+ data=yaml_data,
+ file_name='config.yaml',
+ mime='text/yaml'
+ )
+
+break_even_values = []
+take_profit_values = []
+for level in range(n_levels):
+ spreads_normalized = [Decimal(spread) + Decimal(0.01) for spread in spreads[:level+1]]
+ amounts = order_amounts[:level+1]
+ break_even = (sum([spread * amount for spread, amount in zip(spreads_normalized, amounts)]) / sum(amounts)) - Decimal(0.01)
+ break_even_values.append(break_even)
+ take_profit_values.append(break_even - Decimal(tp))
+
+accumulated_amount = [sum(order_amounts[:i+1]) for i in range(len(order_amounts))]
+
+
+def calculate_unrealized_pnl(spreads, break_even_values, accumulated_amount):
+ unrealized_pnl = []
+ for i in range(len(spreads)):
+ distance = abs(spreads[i] - break_even_values[i])
+ pnl = accumulated_amount[i] * distance / 100 # PNL calculation
+ unrealized_pnl.append(pnl)
+ return unrealized_pnl
+
+# Calculate unrealized PNL
+cum_unrealized_pnl = calculate_unrealized_pnl(spreads, break_even_values, accumulated_amount)
+
+
+tech_colors = {
+ 'spread': '#00BFFF', # Deep Sky Blue
+ 'break_even': '#FFD700', # Gold
+ 'take_profit': '#32CD32', # Green
+ 'order_amount': '#1E90FF', # Dodger Blue
+ 'cum_amount': '#4682B4', # Steel Blue
+ 'stop_loss': '#FF0000', # Red
+}
+
+# Create Plotly figure with secondary y-axis and a dark theme
+fig = make_subplots(specs=[[{"secondary_y": True}]])
+fig.update_layout(template="plotly_dark")
+
+# Update the Scatter Plots and Horizontal Lines
+fig.add_trace(go.Scatter(x=list(range(len(spreads))), y=spreads, name='Spread (%)', mode='lines+markers', line=dict(width=3, color=tech_colors['spread'])), secondary_y=False)
+fig.add_trace(go.Scatter(x=list(range(len(break_even_values))), y=break_even_values, name='Break Even (%)', mode='lines+markers', line=dict(width=3, color=tech_colors['break_even'])), secondary_y=False)
+fig.add_trace(go.Scatter(x=list(range(len(take_profit_values))), y=take_profit_values, name='Take Profit (%)', mode='lines+markers', line=dict(width=3, color=tech_colors['take_profit'])), secondary_y=False)
+
+# Add the new Bar Plot for Cumulative Unrealized PNL
+fig.add_trace(go.Bar(
+ x=list(range(len(cum_unrealized_pnl))),
+ y=cum_unrealized_pnl,
+ text=[f"{pnl:.2f}" for pnl in cum_unrealized_pnl],
+ textposition='auto',
+ textfont=dict(color='white', size=12),
+ name='Cum Unrealized PNL',
+ marker=dict(color='#FFA07A', opacity=0.6) # Light Salmon color, adjust as needed
+), secondary_y=True)
+
+fig.add_trace(go.Bar(
+ x=list(range(len(order_amounts))),
+ y=order_amounts,
+ text=[f"{amt:.2f}" for amt in order_amounts], # List comprehension to format text labels
+ textposition='auto',
+ textfont=dict(
+ color='white',
+ size=12
+ ),
+ name='Order Amount',
+ marker=dict(color=tech_colors['order_amount'], opacity=0.5),
+), secondary_y=True)
+
+# Modify the Bar Plot for Accumulated Amount
+fig.add_trace(go.Bar(
+ x=list(range(len(accumulated_amount))),
+ y=accumulated_amount,
+ text=[f"{amt:.2f}" for amt in accumulated_amount], # List comprehension to format text labels
+ textposition='auto',
+ textfont=dict(
+ color='white',
+ size=12
+ ),
+ name='Cum Amount',
+ marker=dict(color=tech_colors['cum_amount'], opacity=0.5),
+), secondary_y=True)
+
+
+# Add Horizontal Lines for Last Breakeven Price and Stop Loss Level
+last_break_even = break_even_values[-1]
+stop_loss_value = last_break_even + Decimal(sl)
+# Horizontal Lines for Last Breakeven and Stop Loss
+fig.add_hline(y=last_break_even, line_dash="dash", annotation_text=f"Global Break Even: {last_break_even:.2f} (%)", annotation_position="top left", line_color=tech_colors['break_even'])
+fig.add_hline(y=stop_loss_value, line_dash="dash", annotation_text=f"Stop Loss: {stop_loss_value:.2f} (%)", annotation_position="bottom right", line_color=tech_colors['stop_loss'])
+
+# Update Annotations for Spread and Break Even
+for i, (spread, be_value, tp_value) in enumerate(zip(spreads, break_even_values, take_profit_values)):
+ fig.add_annotation(x=i, y=spread, text=f"{spread:.2f}%", showarrow=True, arrowhead=1, yshift=10, xshift=-2, font=dict(color=tech_colors['spread']))
+ fig.add_annotation(x=i, y=be_value, text=f"{be_value:.2f}%", showarrow=True, arrowhead=1, yshift=5, xshift=-2, font=dict(color=tech_colors['break_even']))
+ fig.add_annotation(x=i, y=tp_value, text=f"{tp_value:.2f}%", showarrow=True, arrowhead=1, yshift=10, xshift=-2, font=dict(color=tech_colors['take_profit']))
+# Update Layout with a Dark Theme
+fig.update_layout(
+ title="Spread, Accumulated Amount, Break Even, and Take Profit by Order Level",
+ xaxis_title="Order Level",
+ yaxis_title="Spread (%)",
+ yaxis2_title="Amount (Quote)",
+ height=800,
+ width=1800,
+ plot_bgcolor='rgba(0, 0, 0, 0)', # Transparent background
+ paper_bgcolor='rgba(0, 0, 0, 0.1)', # Lighter shade for the paper
+ font=dict(color='white') # Font color
+)
+
+# Calculate metrics
+max_loss = total_amount_quote * Decimal(sl / 100)
+profit_per_level = [cum_amount * Decimal(tp / 100) for cum_amount in accumulated_amount]
+loots_to_recover = [max_loss / profit for profit in profit_per_level]
+
+# Define a consistent annotation size and maximum value for the secondary y-axis
+circle_text = "●" # Unicode character for a circle
+max_secondary_value = max(max(accumulated_amount), max(order_amounts), max(cum_unrealized_pnl)) # Adjust based on your secondary y-axis data
+
+# Determine an appropriate y-offset for annotations
+y_offset_secondary = max_secondary_value * Decimal(0.1) # Adjusts the height relative to the maximum value on the secondary y-axis
+
+# Add annotations to the Plotly figure for the secondary y-axis
+for i, loot in enumerate(loots_to_recover):
+ fig.add_annotation(
+ x=i,
+ y=max_secondary_value + y_offset_secondary, # Position above the maximum value using the offset
+ text=f"{circle_text}<br>LTR: {round(loot, 2)}", # Circle symbol and loot value in separate lines
+ showarrow=False,
+ font=dict(size=16, color='purple'),
+ xanchor="center", # Centers the text above the x coordinate
+ yanchor="bottom", # Anchors the text at its bottom to avoid overlapping
+ align="center",
+ yref="y2" # Reference the secondary y-axis
+ )
+# Add Max Loss Metric as an Annotation
+max_loss_annotation_text = f"Max Loss (Quote): {max_loss:.2f}"
+fig.add_annotation(
+ x=max(len(spreads), len(break_even_values)) - 1, # Positioning the annotation to the right
+ text=max_loss_annotation_text,
+ showarrow=False,
+ font=dict(size=20, color='white'),
+ bgcolor='red', # Red background for emphasis
+ xanchor="left",
+ yanchor="top",
+ yref="y2" # Reference the secondary y-axis
+)
+
+st.write("---")
+
+# Display in Streamlit
+st.plotly_chart(fig)
+