Cyptknow / app.py
MacDash's picture
Update app.py
1e2c87b verified
import gradio as gr
import pandas as pd
import numpy as np
from prophet import Prophet
import plotly.graph_objs as go
import requests
from sklearn.ensemble import RandomForestClassifier
from textblob import TextBlob
import yfinance as yf
import re
# --- Constants ---
CRYPTO_SYMBOLS = ["BTC-USD", "ETH-USD", "LTC-USD", "XRP-USD"]
STOCK_SYMBOLS = ["AAPL", "MSFT", "GOOGL", "AMZN"]
INTERVAL_OPTIONS = ["1h", "1d", "1wk"]
DEFAULT_FORECAST_STEPS = 24
DEFAULT_DAILY_SEASONALITY = True
DEFAULT_WEEKLY_SEASONALITY = True
DEFAULT_YEARLY_SEASONALITY = False
DEFAULT_SEASONALITY_MODE = "additive"
DEFAULT_CHANGEPOINT_PRIOR_SCALE = 0.05
RANDOM_FOREST_PARAMS = {
"n_estimators": 100,
"max_depth": 10,
"random_state": 42
}
# --- Data Fetching Functions ---
def fetch_crypto_data(symbol, interval="1h", limit=100):
try:
ticker = yf.Ticker(symbol)
if interval == "1h":
period = "1d"
df = ticker.history(period=period, interval="1h")
elif interval == "1d":
df = ticker.history(period="1y", interval=interval)
elif interval == "1wk":
df = ticker.history(period="5y", interval=interval)
else:
raise ValueError("Invalid interval for yfinance.")
if df.empty:
raise Exception("No data returned from yfinance.")
df.reset_index(inplace=True)
df.rename(columns={"Datetime": "timestamp", "Open": "open", "High": "high", "Low": "low", "Close": "close", "Volume": "volume"}, inplace=True)
df = df[["timestamp", "open", "high", "low", "close", "volume"]]
return df.dropna()
except Exception as e:
raise Exception(f"Error fetching crypto data from yfinance: {e}")
def fetch_stock_data(symbol, interval="1d"):
try:
ticker = yf.Ticker(symbol)
df = ticker.history(period="1y", interval=interval)
if df.empty:
raise Exception("No data returned from yfinance.")
df.reset_index(inplace=True)
df.rename(columns={"Date": "timestamp", "Open": "open", "High": "high", "Low": "low", "Close": "close", "Volume": "volume"}, inplace=True)
df = df[["timestamp", "open", "high", "low", "close", "volume"]]
return df.dropna()
except Exception as e:
raise Exception(f"Error fetching stock data from yfinance: {e}")
def fetch_sentiment_data(keyword):
try:
tweets = [
f"{keyword} is going to moon!",
f"I hate {keyword}, it's trash!",
f"{keyword} is amazing!"
]
sentiments = [TextBlob(tweet).sentiment.polarity for tweet in tweets]
return sum(sentiments) / len(sentiments) if sentiments else 0
except Exception as e:
print(f"Sentiment analysis error: {e}")
return 0
# --- Technical Analysis Functions ---
def calculate_technical_indicators(df):
if df.empty:
return df
delta = df['close'].diff()
gain = (delta.where(delta > 0, 0)).rolling(window=14).mean()
loss = (-delta.where(delta < 0, 0)).rolling(window=14).mean()
rs = gain / loss
df['RSI'] = 100 - (100 / (1 + rs))
exp1 = df['close'].ewm(span=12, adjust=False).mean()
exp2 = df['close'].ewm(span=26, adjust=False).mean()
df['MACD'] = exp1 - exp2
df['Signal_Line'] = df['MACD'].ewm(span=9, adjust=False).mean()
df['MA20'] = df['close'].rolling(window=20).mean()
df['BB_upper'] = df['MA20'] + 2 * df['close'].rolling(window=20).std()
df['BB_lower'] = df['MA20'] - 2 * df['close'].rolling(window=20).std()
return df
def create_technical_charts(df):
if df.empty:
return None, None, None
fig1 = go.Figure()
fig1.add_trace(go.Candlestick(
x=df['timestamp'],
open=df['open'],
high=df['high'],
low=df['low'],
close=df['close'],
name='Price'
))
fig1.add_trace(go.Scatter(x=df['timestamp'], y=df['BB_upper'], name='Upper BB', line=dict(color='gray', dash='dash')))
fig1.add_trace(go.Scatter(x=df['timestamp'], y=df['BB_lower'], name='Lower BB', line=dict(color='gray', dash='dash')))
fig1.update_layout(title='Price and Bollinger Bands', xaxis_title='Date', yaxis_title='Price')
fig2 = go.Figure()
fig2.add_trace(go.Scatter(x=df['timestamp'], y=df['RSI'], name='RSI'))
fig2.add_hline(y=70, line_dash="dash", line_color="red")
fig2.add_hline(y=30, line_dash="dash", line_color="green")
fig2.update_layout(title='RSI Indicator', xaxis_title='Date', yaxis_title='RSI')
fig3 = go.Figure()
fig3.add_trace(go.Scatter(x=df['timestamp'], y=df['MACD'], name='MACD'))
fig3.add_trace(go.Scatter(x=df['timestamp'], y=df['Signal_Line'], name='Signal Line'))
fig3.update_layout(title='MACD', xaxis_title='Date', yaxis_title='Value')
return fig1, fig2, fig3
# --- Prophet Forecasting Functions ---
def prepare_data_for_prophet(df):
if df.empty:
return pd.DataFrame(columns=["ds", "y"])
df_prophet = df.rename(columns={"timestamp": "ds", "close": "y"})
return df_prophet[["ds", "y"]]
def prophet_forecast(df_prophet, periods=10, freq="h", daily_seasonality=False, weekly_seasonality=False, yearly_seasonality=False, seasonality_mode="additive", changepoint_prior_scale=0.05):
if df_prophet.empty:
return pd.DataFrame(), "No data for Prophet."
try:
model = Prophet(
daily_seasonality=daily_seasonality,
weekly_seasonality=weekly_seasonality,
yearly_seasonality=yearly_seasonality,
seasonality_mode=seasonality_mode,
changepoint_prior_scale=changepoint_prior_scale,
)
model.fit(df_prophet)
future = model.make_future_dataframe(periods=periods, freq=freq)
forecast = model.predict(future)
return forecast, ""
except Exception as e:
return pd.DataFrame(), f"Forecast error: {e}"
def prophet_wrapper(df_prophet, forecast_steps, freq, daily_seasonality, weekly_seasonality, yearly_seasonality, seasonality_mode, changepoint_prior_scale):
if len(df_prophet) < 10:
return pd.DataFrame(), "Not enough data for forecasting (need >=10 rows)."
full_forecast, err = prophet_forecast(
df_prophet,
forecast_steps,
freq,
daily_seasonality,
weekly_seasonality,
yearly_seasonality,
seasonality_mode,
changepoint_prior_scale,
)
if err:
return pd.DataFrame(), err
future_only = full_forecast.loc[len(df_prophet):, ["ds", "yhat", "yhat_lower", "yhat_upper"]]
return future_only, ""
def create_forecast_plot(forecast_df):
if forecast_df.empty:
return go.Figure()
fig = go.Figure()
fig.add_trace(go.Scatter(
x=forecast_df["ds"],
y=forecast_df["yhat"],
mode="lines",
name="Forecast",
line=dict(color="blue", width=2)
))
fig.add_trace(go.Scatter(
x=forecast_df["ds"],
y=forecast_df["yhat_lower"],
fill=None,
mode="lines",
line=dict(width=0),
showlegend=True,
name="Lower Bound"
))
fig.add_trace(go.Scatter(
x=forecast_df["ds"],
y=forecast_df["yhat_upper"],
fill="tonexty",
mode="lines",
line=dict(width=0),
name="Upper Bound"
))
fig.update_layout(
title="Price Forecast",
xaxis_title="Time",
yaxis_title="Price",
hovermode="x unified",
template="plotly_white",
)
return fig
# --- Model Training and Prediction ---
model = RandomForestClassifier(**RANDOM_FOREST_PARAMS)
def train_model(df):
if df.empty:
return
df["target"] = (df["close"].pct_change() > 0.05).astype(int)
features = df[["close", "volume"]].dropna()
target = df["target"].dropna()
if not features.empty and not target.empty:
model.fit(features, target)
else:
print("Not enough data for model training.")
def predict_growth(latest_data):
if not hasattr(model, 'estimators_') or len(model.estimators_) == 0:
return [0]
try:
prediction = model.predict(latest_data.reshape(1, -1))
return prediction
except Exception as e:
print(f"Prediction error: {e}")
return [0]
# --- Main Prediction and Display Function ---
def analyze_market(market_type, symbol, interval, forecast_steps, daily_seasonality, weekly_seasonality, yearly_seasonality, seasonality_mode, changepoint_prior_scale, sentiment_keyword=""):
df = pd.DataFrame()
error_message = ""
sentiment_score = 0
try:
if market_type == "Crypto":
df = fetch_crypto_data(symbol, interval=interval)
elif market_type == "Stock":
df = fetch_stock_data(symbol, interval=interval)
else:
error_message = "Invalid market type selected."
return None, None, None, None, None, "", error_message, 0
if sentiment_keyword:
sentiment_score = fetch_sentiment_data(sentiment_keyword)
except Exception as e:
error_message = f"Data Fetching Error: {e}"
return None, None, None, None, None, "", error_message, 0
if df.empty:
error_message = "No data fetched."
return None, None, None, None, None, "", error_message, 0
df["timestamp"] = pd.to_datetime(df["timestamp"])
numeric_cols = ["open", "high", "low", "close", "volume"]
df[numeric_cols] = df[numeric_cols].astype(float)
df = calculate_technical_indicators(df)
df_prophet = prepare_data_for_prophet(df)
freq = "h" if interval == "1h" or interval == "60min" else "d"
forecast_df, prophet_error = prophet_wrapper(
df_prophet,
forecast_steps,
freq,
daily_seasonality,
weekly_seasonality,
yearly_seasonality,
seasonality_mode,
changepoint_prior_scale,
)
if prophet_error:
error_message = f"Prophet Error: {prophet_error}"
return None, None, None, None, None, "", error_message, sentiment_score
forecast_plot = create_forecast_plot(forecast_df)
tech_plot, rsi_plot, macd_plot = create_technical_charts(df)
try:
train_model(df.copy())
if not df.empty:
latest_data = df[["close", "volume"]].iloc[-1].values
growth_prediction = predict_growth(latest_data)
growth_label = "Yes" if growth_prediction[0] == 1 else "No"
else:
growth_label = "N/A: Insufficient Data"
except Exception as e:
error_message = f"Model Error: {e}"
growth_label = "N/A"
forecast_df_display = forecast_df.loc[:, ["ds", "yhat", "yhat_lower", "yhat_upper"]].copy()
forecast_df_display.rename(columns={"ds": "Date", "yhat": "Forecast", "yhat_lower": "Lower Bound", "yhat_upper": "Upper Bound"}, inplace=True)
return forecast_plot, tech_plot, rsi_plot, macd_plot, forecast_df_display, growth_label, error_message, sentiment_score
def chatbot_response(message, history):
market_type = "Crypto" # Default market type
symbol = "BTC-USD" # Default symbol
interval = "1h" # Default interval
forecast_steps = DEFAULT_FORECAST_STEPS
daily_seasonality = DEFAULT_DAILY_SEASONALITY
weekly_seasonality = DEFAULT_WEEKLY_SEASONALITY
yearly_seasonality = DEFAULT_YEARLY_SEASONALITY
seasonality_mode = DEFAULT_SEASONALITY_MODE
changepoint_prior_scale = DEFAULT_CHANGEPOINT_PRIOR_SCALE
sentiment_keyword = ""
# Simple keyword parsing - improve this for more robust parsing
message_lower = message.lower()
if "stock" in message_lower:
market_type = "Stock"
symbol = "AAPL" # Default stock symbol
elif "crypto" in message_lower:
market_type = "Crypto"
symbol = "BTC-USD" # Default crypto symbol
for crypto_sym in CRYPTO_SYMBOLS:
if crypto_sym.lower() in message_lower:
symbol = crypto_sym
market_type = "Crypto"
break
for stock_sym in STOCK_SYMBOLS:
if stock_sym.lower() in message_lower:
symbol = stock_sym
market_type = "Stock"
break
for intv in INTERVAL_OPTIONS:
if intv in message_lower:
interval = intv
break
forecast_steps_match = re.search(r'forecast\s*(\d+)\s*steps', message_lower)
if forecast_steps_match:
forecast_steps = int(forecast_steps_match.group(1))
sentiment_match = re.search(r'sentiment\s*(.+)', message_lower)
if sentiment_match:
sentiment_keyword = sentiment_match.group(1).strip()
plots, tech_plot, rsi_plot, macd_plot, forecast_df, growth_label, error_message, sentiment_score = analyze_market(
market_type, symbol, interval, forecast_steps, daily_seasonality, weekly_seasonality, yearly_seasonality, seasonality_mode, changepoint_prior_scale, sentiment_keyword
)
response = ""
if error_message:
response += f"Error: {error_message}\n\n"
if plots and not error_message:
response += "Here is the price forecast plot.\n\n" # In a real chatbot, you might provide a link or embed the plot if possible.
else:
response += "Could not generate forecast plot.\n\n"
if tech_plot and rsi_plot and macd_plot and not error_message:
response += "Technical analysis plots (Bollinger Bands, RSI, MACD) are generated.\n\n" # Again, link or embed plots in a real chatbot
else:
response += "Could not generate technical analysis plots.\n\n"
if not error_message:
response += f"Explosive Growth Prediction: {growth_label}\n"
response += f"Sentiment Score (for keyword '{sentiment_keyword}'): {sentiment_score:.2f}\n"
if not forecast_df.empty:
# Summarize forecast data instead of displaying the full dataframe in text
forecast_summary = forecast_df.tail().to_string() # Just showing last few rows as summary
response += "\nForecast Data Summary (last few points):\n" + forecast_summary + "\n"
else:
response += "\nNo forecast data available.\n"
return response
with gr.ChatInterface(
chatbot_response,
title="Market Analysis Chatbot",
description="Ask me about crypto or stock market analysis. For example, try: 'Analyze crypto BTC-USD 1d forecast 30 steps sentiment Bitcoin' or 'Stock AAPL 1h analysis'.",
examples=[
"Analyze crypto ETH-USD 1h",
"Stock MSFT 1d forecast 10 steps",
"Crypto LTC-USD 1wk sentiment Litecoin",
"Analyze stock GOOGL",
"What about crypto XRP-USD?",
],
theme=gr.themes.Base()
) as demo:
demo.launch()