coin_analyser.py

"""The main script that demos most of the features of the all project."""

from technical_analyser import technical_analyser
from chain_analyser import chain_analyser
from fundamental_analyser import fundamental_analyser
from macroeconomic_analyser import macroeconomic_analyser
from langchain_openai import ChatOpenAI
from langchain_openai import AzureOpenAI
from langchain_groq import ChatGroq
from langchain_google_genai import ChatGoogleGenerativeAI
import os
import time
from cachetools import TTLCache
import logging
from datetime import datetime, timedelta
import pandas

logging.basicConfig(level=logging.INFO)

# Caching 100 items for 24 hours
cache = TTLCache(maxsize=100, ttl=86400)
logging.basicConfig(level=logging.INFO)

def _get_llm(model_name:str):
    "Return one of the LangChain LLM's according to the model_name. OpenAI, AzureOpenAI, llama-3 and Gemini-pro is supported."
    try:
        if model_name == "openai":
            OPENAI_KEY = os.getenv("OPENAI_KEY")
            llm = ChatOpenAI(model="gpt-4o",temperature=0,max_tokens=None,timeout=None,max_retries=2,api_key=OPENAI_KEY)
        elif model_name == "azure_openai":
            AZURE_AD_TOKEN = os.getenv("AZURE_AD_TOKEN")
            AZURE_AD_TOKEN_PROVIDER = os.getenv("AZURE_AD_TOKEN_PROVIDER")
            AZURE_DEPLOYMENT = os.getenv("AZURE_DEPLOYMENT")
            AZURE_ENDPOINT = os.getenv("AZURE_ENDPOINT")
            llm = AzureOpenAI(azure_ad_token=AZURE_AD_TOKEN, azure_ad_token_provider=AZURE_AD_TOKEN_PROVIDER, azure_deployment=AZURE_DEPLOYMENT, azure_endpoint=AZURE_ENDPOINT, model="gpt-3.5-turbo-instruct", temperature=0,)
        elif model_name == "llama3":
            GROQ_API_KEY = os.getenv("GROQ_API_KEY")
            llm = ChatGroq(api_key=GROQ_API_KEY, model_name="llama3-70b-8192", temperature=0.0) #llama-3.1-70b-versatile
        else:
            GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
            llm = ChatGoogleGenerativeAI(google_api_key=GOOGLE_API_KEY, model="gemini-1.5-pro", temperature=0)
        return llm
    except Exception as e:
        logging.error(f"Error getting LLM: {e}")
        return None


def _get_answer_from_llm(llm: any, input:str) -> str:
    """The last llm call that gathers all analyzes and creates a one final analysis interpretion to send to the user."""
    prompt = """
    You are an experienced cryptocurrency trader and analyst who specializes in analysing blockchain projects.
    You analyse projects by doing Technical(technical analysis), Chain (chain analysis), Fundamental(fundamental analysis) and Macroeconomic(global macroeconomic financial analysis).
    You will be presented all these analysis.
    Your goal is to aggregate and summarize all these findings to the user.
    The background is a trader is asking you to anlyze a coin and by the data you will receive you will give your answer.
    Your response should be professional, technical, and easy to understand.
    ONLY ANSWER IN TURKISH.

    <Example>:
    <HUMAN>:
    
    <Technical>
    Technical Analysis UP Signals: 0
    DOWN Signals: 3 (SAR, Stochastic Oscillator, SMA)
    NEUTRAL Signals: 5 (CCI, ATR, Bollinger Bands, ADX, RSI, Previous Performance Potential)
    Given the mix of signals:
    The SAR, Stochastic Oscillator, and SMA indicate a down due to the downward trend, strong bearish trend, and price being below the Parabolic SAR, respectively.
    However, there are no strong up signals to counterbalance the bearish indicators.
    Prediction: Slight Decrease.
    The market signals indicate a bearish trend, with three indicators pointing towards a downtrend.
    Although there are neutral indicators, they are not strong enough to offset the bearish signals.
    It is important to use a risk-managed approach, such as setting tight stop-loss orders to mitigate potential losses.
    A small to moderate short position with a stop-loss above the current price and a reasonable take-profit target below the current price is suggested.
    Continuously monitor the coin's price action and adjust positions as needed.
    It is advisable to wait for clearer trend signals before making significant trades, always considering a risk-managed approach.
    
    <Chain>
    Current Price: $56,423
    Market Cap: $1,114,378,772,366 (Ranked 1st)
    24h Price Change: -$671.34 (-1.18%)
    24h Market Cap Change: -$12,618,620,188.38 (-1.12%)
    Circulating Supply: 19,719,662 BTC
    All-Time High (ATH): $73,738 (23.38% below ATH)
    All-Time Low (ATL): $67.81 (83,216.62% above ATL)
    Trading Volume: $42,537,556,960
    Exchange Inflow: 19,722,353.13 BTC
    Exchange Outflow: 55,848.73 BTC
    Price Change: 1.18% decrease (DOWN)
    Market Cap Change: 1.12% decrease (DOWN)
    Trading Volume: High (NEUTRAL)
    Significant Change from ATH: 23.38% below ATH (NEUTRAL)
    Significant Change from ATL: Significantly higher than ATL (NEUTRAL)
    High Circulating Supply: Neutral impact (NEUTRAL)
    Low Transaction Volume & Number of Transactions: Extremely low (DOWN)
    Exchange Inflows & Outflows: High inflows, low outflows (DOWN)
    Market Cap Rank: High market cap indicates stability (NEUTRAL)
    Prediction: Slightly down
    The prediction is bearish due to the negative price and market cap changes, low transaction volume, and bearish exchange flows.
    A stop-loss is recommended at $53,000, with a target price of $52,000.
    Monitor the market for changes in transaction volume and exchange flows, which could indicate a shift in market behavior.
    
    <Fundamental>
    Bitcoin is the first successful decentralized cryptocurrency enabling peer-to-peer transactions without intermediaries.
    Key Indicators:
    Block Time: 10 minutes, moderate speed (NEUTRAL).
    Hashing Algorithm: SHA-256, secure (UP).
    Web Presence: Strong with official site, whitepaper, and multiple tracking sites (UP).
    Social Media & Community: Active on major platforms, large followings (UP).
    Developer Activity: High GitHub activity, strong support (UP).
    Sentiment: 64.59% positive, mixed community trust (NEUTRAL).
    Additional Factors: 15 years old, established platform (UP).
    Conclusion:
    Bitcoin shows strong performance in security, web presence, social media, developer activity, and community support. However, moderate transaction speed and mixed sentiment result in an overall NEUTRAL recommendation.
    
    <Macroeconomic>
    Inflation: Mixed signals; normal in the US and EU, high in India, Russia, and Japan.
    Interest Rates: High in key economies, indicating tightening policies.
    GDP Growth: Positive overall, except for low/negative growth in the US and China.
    Market Indices: Downward trends, indicating bearish sentiment.
    Commodities: Precious metals uptrend, industrial commodities downtrend.
    Currency Strength: Strength in JPY and CNY, others weak or neutral.
    Elections: High volatility expected due to upcoming elections.
    Overall Guidance:
    The global economic outlook shows inflationary pressures, potential slowdowns, and high volatility due to elections and geopolitical tensions. A cautious investment approach is advised.
    Recommendations:
    Focus on safe-haven assets: Precious metals and strong currencies.
    Monitor central bank policies: Interest rate and monetary policy changes.
    Diversify investments: Across asset classes and regions.
    Watch geopolitical developments: Impact on market sentiment and volatility.
    Prepare for high volatility: Due to elections and geopolitical tensions.

    <AI>:
    Bitcoin (BTC) Kapsamlı Analizi
    Teknik Analiz:
    YÜKSELİŞ Sinyalleri: 0
    DÜŞÜŞ Sinyalleri: 3 (SAR, Stokastik Osilatör, SMA)
    NÖTR Sinyaller: 5 (CCI, ATR, Bollinger Bantları, ADX, RSI, Önceki Performans Potansiyeli)
    Özet: SAR, Stokastik Osilatör ve SMA, ayı trendleri ve fiyatın anahtar seviyelerin altında olması nedeniyle bir düşüş trendi gösteriyor. Nötr göstergeler, düşüş sinyallerine yeterli denge sağlamıyor, bu da hafif bir düşüş öngörüsüne yol açıyor. Sıkı stop-loss emirleriyle risk yönetimi önerilir.
    Tavsiye: Mevcut fiyatın üzerine stop-loss ayarlayarak ve mevcut fiyatın biraz altına take-profit hedefi belirleyerek küçük ila orta ölçekli bir short pozisyon düşünün. Coin'in fiyat hareketlerini izlemek ve pozisyonu buna göre ayarlamak önemlidir.

    Zincir Analizi:
    Mevcut Fiyat: $56,423
    Piyasa Değeri: $1,114,378,772,366 (1. sırada)
    24 Saatlik Fiyat Değişimi: -$671,34 (-%1,18)
    24 Saatlik Piyasa Değeri Değişimi: -$12,618,620,188,38 (-%1,12)
    Dolaşımdaki Arz: 19,719,662.0 BTC
    Tüm Zamanların En Yüksek Değeri (ATH): $73,738 (ATH'nin %23,38 altında)
    Tüm Zamanların En Düşük Değeri (ATL): $67,81 (ATL'nin %83,216.62 üzerinde)
    İşlem Hacmi: $42,537,556,960
    İşlem Sayısı: 0
    İşlem Hacmi: 0
    Borsaya Giriş: 19,722,353.13 BTC
    Borsadan Çıkış: 55,848.73 BTC
    Özet: Analiz, fiyat ve piyasa değeri düşüşleri, düşük ağ aktivitesi ve yüksek borsa girişleri nedeniyle potansiyel satış baskısını işaret eden kısa vadeli ayı momentumunu göstermektedir. Nötr sinyaller, ayı göstergelerini dengeleyemediği için hafif bir düşüş öngörülüyor.
    Tavsiye: Potansiyel kayıpları sınırlamak için $53,000 seviyesinde stop-loss belirleyin, mevcut ayı momentumu temel alınarak $52,000 hedef fiyat belirleyin.
    Dikkatli olun ve piyasa davranışlarındaki değişiklikler için ağ aktivitelerini ve borsa giriş-çıkışlarını izleyin.

    Temel Analiz:
    Bitcoin, güvenlik, web varlığı, sosyal medya etkinliği, geliştirici desteği ve topluluk katılımı açısından güçlü göstergelere sahiptir.
    Bununla birlikte, nötr duygu oranı ve orta derecede işlem hızına bağlı olarak genel olarak nötr bir tavsiye verilmektedir.

    Makroekonomik Analiz:
    Küresel ekonomi, enflasyonist baskılar, potansiyel ekonomik yavaşlamalar ve seçimler ile jeopolitik gerilimler nedeniyle yüksek volatilite ile karşı karşıyadır.
    Dikkatli bir yatırım yaklaşımı önerilmektedir.
    Tavsiyeler:
    Güvenli liman varlıklarına odaklanılabilir.
    Merkez bankası politikalarını izlenilebilir: Faiz oranı kararları ve para politikası değişikliklerini takipte kalınabilir.
    Yatırımları çeşitlendirilebilir: Yatırımları farklı varlık sınıflarına ve bölgelere yayarak çeşitlendirilebilir.
    Yüksek volatiliteye hazırlıklı olun: Seçim yılı ve jeopolitik gerilimler ani piyasa hareketlerine neden olabilir.

    Genel Sonuç:
    Teknik, zincir, temel ve makroekonomik perspektiflerden yapılan birleşik analiz, Bitcoin (BTC) konusunda temkinli bir duruşu işaret etmektedir.
    Teknik ve zincir analizlerinden gelen ayı sinyalleri, temkinli makroekonomik görünümle uyumludur.
    Temel analiz, uzun vadede güçlü destek sağlamakla birlikte, kısa vadeli riskleri ortadan kaldırmamaktadır.
    Ticaret stratejilerini buna göre ayarlayın, risk yönetimine odaklanın ve piyasa gelişmelerini takip ederek bilgi sahibi olun.

    Now it's your turn:
    """
    final_answer = llm.invoke(prompt + input)
    final_answer = final_answer.content
    return final_answer


def cached(func):
    """Caches input output pair."""
    def wrapper(*args, **kwargs):
        cache_key = (args, tuple(sorted(kwargs.items())))
        if cache_key in cache:
            return cache[cache_key]
        result = func(*args, **kwargs)
        cache[cache_key] = result
        return result
    return wrapper


@cached
def coin_analyser(coin_name:str, start_date:str, end_date:str, visualization:bool=True) -> dict[str,str]:
    """The main method of the main script that showcase projects features by calling all analysers and sending the all the outputs with a last llm call to get it summarized."""
    logging.info("Technical analyser...")
    technical_analysis, technical_analysis_figure = technical_analyser(coin_name=coin_name, visualization=visualization) # longer than 2 weeks is not that meaningful
    logging.info(technical_analysis)
    
    time.sleep(20) # To pass rate limits

    logging.info("Chain analyser...")
    chain_analysis, chain_analysis_figure = chain_analyser(coin_name=coin_name, visualization=visualization) # does not require time
    logging.info(chain_analysis)

    logging.info("Chain analyser...")
    fundamental_analysis, fundamental_analysis_figure = fundamental_analyser(coin_name=coin_name, visualization=visualization) # does not require time
    logging.info(fundamental_analysis)

    time.sleep(20) # To pass rate limits

    logging.info("Macroeconomic analyser...")
    macroeconomic_analysis, macroeconomic_analysis_figure = macroeconomic_analyser(start_date=start_date,end_date=end_date, visualization=visualization) # looks at macroeconomical state of the world
    logging.info(macroeconomic_analysis)

    final_request = technical_analysis + chain_analysis + fundamental_analysis + macroeconomic_analysis

    llm = _get_llm("google")
    total_analysis = _get_answer_from_llm(llm=llm, input=final_request)

    coin_analysis = {
        'time': str(datetime.now()),
        'coin_name': coin_name,
        'start_date': start_date,
        'end_date': end_date,
        'total_analysis': total_analysis,
        'technical_analysis': technical_analysis,
        'chain_analysis': chain_analysis,
        'fundamental_analysis': fundamental_analysis,
        'macroeconomic_analysis':macroeconomic_analysis,
        'technical_analysis_figure': technical_analysis_figure,
        'chain_analysis_figure': chain_analysis_figure,
        'fundamental_analysis_figure': fundamental_analysis_figure,
        'macroeconomic_analysis_figure':macroeconomic_analysis_figure
    }
    
    coin_analysis_dataframe = pandas.DataFrame(coin_analysis, index=[0])
    coin_analysis_dataframe.to_csv("coin_analysis.csv", mode='a', index=True)

    return coin_analysis

if __name__ == '__main__':
    # Get today's date
    today = datetime.today()

    # Set start_date to 2 days ago
    start_date = (today - timedelta(days=2)).strftime('%Y-%m-%d')

    # Set end_date to yesterday
    end_date = (today - timedelta(days=1)).strftime('%Y-%m-%d')
    
    #tickers = ["btc", "eth", "ham", "pepe24478", "floki", "fet"]
    #coin_comparator(tickers=tickers, start_date=start_date, end_date=end_date)
    
    coin_name="bitcoin"
    coin_analysis = coin_analyser(coin_name=coin_name, start_date=start_date, end_date=end_date, visualization=False)
    print("\nRESULT\n:")
    print(coin_analysis["total_analysis"])