# Create project directory
mkdir ai-bitcoin-analyzer
cd ai-bitcoin-analyzer

# Create and activate virtual environment
python -m venv venv

# On Windows:
venv\Scripts\activate

# On Mac/Linux:
source venv/bin/activate

# Install required packages
pip install streamlit requests plotly python-dotenv transformers

streamlit==1.32.0
requests==2.31.0
plotly==5.18.0
python-dotenv==1.0.0
transformers==4.38.0

import requests
from datetime import datetime

def get_bitcoin_price():
    """Fetch current Bitcoin price from CoinGecko free API"""
    url = "https://api.coingecko.com/api/v3/simple/price"
    params = {
        "ids": "bitcoin",
        "vs_currencies": "usd",
        "include_24hr_change": "true",
        "include_market_cap": "true",
        "include_24hr_vol": "true"
    }
    response = requests.get(url, params=params)
    data = response.json()
    return {
        "price": data["bitcoin"]["usd"],
        "change_24h": data["bitcoin"]["usd_24h_change"],
        "market_cap": data["bitcoin"]["usd_market_cap"],
        "volume_24h": data["bitcoin"]["usd_24h_vol"]
    }

def get_price_history(days=30):
    """Get historical price data for chart"""
    url = "https://api.coingecko.com/api/v3/coins/bitcoin/market_chart"
    params = {"vs_currency": "usd", "days": days, "interval": "daily"}
    response = requests.get(url, params=params)
    data = response.json()
    prices = [p[1] for p in data["prices"]]
    timestamps = [datetime.fromtimestamp(p[0]/1000).strftime("%Y-%m-%d")
                  for p in data["prices"]]
    return timestamps, prices

def get_fear_greed_index():
    """Fetch Fear & Greed Index from alternative.me"""
    url = "https://api.alternative.me/fng/"
    response = requests.get(url)
    data = response.json()
    return {
        "value": int(data["data"][0]["value"]),
        "classification": data["data"][0]["value_classification"]
    }

import requests
import os

# Free HuggingFace Inference API — get your token at:
# https://huggingface.co/settings/tokens
HF_TOKEN = os.getenv("HF_TOKEN", "")

def analyze_sentiment(headlines: list) -> dict:
    """
    Analyze a list of news headlines using FinBERT.
    Returns overall sentiment score and per-headline breakdown.
    """
    API_URL = "https://api-inference.huggingface.co/models/ProsusAI/finbert"
    headers = {"Authorization": f"Bearer {HF_TOKEN}"}
    results = {"positive": 0, "negative": 0, "neutral": 0}
    detailed = []

    for headline in headlines:
        response = requests.post(
            API_URL, headers=headers,
            json={"inputs": headline}
        )
        if response.status_code == 200:
            scores = response.json()[0]
            top = max(scores, key=lambda x: x["score"])
            label = top["label"].lower()
            results[label] += 1
            detailed.append({
                "headline": headline[:80] + "...",
                "sentiment": label,
                "confidence": round(top["score"] * 100, 1)
            })

    total = max(sum(results.values()), 1)
    return {
        "overall": ("BULLISH" if results["positive"] > results["negative"]
                    else "BEARISH" if results["negative"] > results["positive"]
                    else "NEUTRAL"),
        "scores": results,
        "bull_pct": round(results["positive"] / total * 100),
        "bear_pct": round(results["negative"] / total * 100),
        "detailed": detailed
    }

# Sample headlines — replace with live news API in production
SAMPLE_HEADLINES = [
    "Bitcoin falls below $63,000 amid extreme investor fear",
    "Crypto ETF sees $203 million in outflows this week",
    "CZ predicts Bitcoin could reach $200,000",
    "Bitcoin RSI more oversold than during 2020 COVID crash",
    "Institutional investors quietly accumulating Bitcoin",
    "U.S. tariff proposals spark crypto selloff",
]

import streamlit as st
import plotly.graph_objects as go
from data_fetcher import get_bitcoin_price, get_price_history, get_fear_greed_index
from sentiment_analyzer import analyze_sentiment, SAMPLE_HEADLINES

st.set_page_config(
    page_title="AI Bitcoin Monitor | AIByTec",
    page_icon="₿",
    layout="wide"
)

st.title("₿ AI Bitcoin Market Analyzer")
st.caption("Powered by FinBERT + CoinGecko API | Built with AIByTec")
st.divider()

# Fetch live data
with st.spinner("Fetching live Bitcoin data..."):
    btc = get_bitcoin_price()
    fng = get_fear_greed_index()
    dates, prices = get_price_history(30)

# Metrics row
col1, col2, col3, col4 = st.columns(4)
with col1:
    st.metric("Bitcoin Price", f"${btc['price']:,.0f}", f"{btc['change_24h']:.2f}%")
with col2:
    st.metric("Market Cap", f"${btc['market_cap']/1e9:.1f}B")
with col3:
    st.metric("24h Volume", f"${btc['volume_24h']/1e9:.1f}B")
with col4:
    icon = "🔴" if fng['value'] < 25 else "🟡" if fng['value'] < 50 else "🟢"
    st.metric("Fear & Greed", f"{icon} {fng['value']}/100", fng['classification'])

# Price chart
st.subheader("📈 30-Day Price History")
fig = go.Figure()
fig.add_trace(go.Scatter(
    x=dates, y=prices, fill="tozeroy",
    line=dict(color="#6366f1", width=2),
    fillcolor="rgba(99,102,241,0.1)", name="BTC/USD"
))
fig.update_layout(
    paper_bgcolor="rgba(0,0,0,0)", plot_bgcolor="rgba(0,0,0,0)",
    height=350, yaxis=dict(tickprefix="$")
)
st.plotly_chart(fig, use_container_width=True)

# AI Sentiment
st.subheader("🧠 AI Sentiment Analysis (FinBERT)")
if st.button("Run AI Sentiment Analysis", type="primary"):
    with st.spinner("Analyzing headlines with AI..."):
        sentiment = analyze_sentiment(SAMPLE_HEADLINES)
    st.metric("Overall Sentiment", sentiment["overall"],
              f"Bull: {sentiment['bull_pct']}% | Bear: {sentiment['bear_pct']}%")
    for item in sentiment["detailed"]:
        icon = "🟢" if item['sentiment']=='positive' else "🔴" if item['sentiment']=='negative' else "⚪"
        st.write(f"{icon} **{item['label'].upper()}** ({item['confidence']}%) — {item['headline']}")

st.caption("⚠️ Educational tool only. Not financial advice. AIByTec.com")

# Set your HuggingFace token (free at huggingface.co/settings/tokens)
# Windows:
set HF_TOKEN=hf_yourTokenHere

# Mac/Linux:
export HF_TOKEN=hf_yourTokenHere

# Run the Streamlit app
streamlit run app.py

# Browser opens automatically at: http://localhost:8501