EasyReportDataMCP/financial_analyzer.py

"""Financial Data Analysis Module"""

from EasyReportDataMCP.edgar_client import EdgarDataClient
from datetime import datetime
from functools import lru_cache
import json


class FinancialAnalyzer:
    def __init__(self, user_agent="Juntao Peng Financial Report Metrics App (jtyxabc@gmail.com)"):
        """
        Initialize financial analyzer
        
        Args:
            user_agent (str): User agent string for identifying request source
        """
        self.edgar_client = EdgarDataClient(user_agent)
        # 新增：实例级缓存，进一步提升性能
        self._search_cache = {}
        self._extract_metrics_cache = {}  # 缓存 extract_financial_metrics 结果
        
    def search_company(self, company_input):
        """
        Search company information (by name or CIK) - Optimized version
        
        Args:
            company_input (str): Company name or CIK
            
        Returns:
            dict: Company information
        """
        # 实例级缓存检查
        if company_input in self._search_cache:
            return self._search_cache[company_input]
        
        # If input is numeric, assume it's a CIK
        if company_input.isdigit() and len(company_input) >= 8:
            # Get company information from cache (will use @lru_cache)
            company_info = self.edgar_client.get_company_info(company_input)
            if company_info:
                self._search_cache[company_input] = company_info
                return company_info
            else:
                return {"error": "Company not found for specified CIK"}
        else:
            # Search company by name/ticker (uses cached company_tickers.json)
            company = self.edgar_client.search_company_by_name(company_input)
            if company:
                # ✅ OPTIMIZATION: Return basic info directly without calling get_company_info
                # search_company_by_name already returns: cik, name, ticker
                # Only call get_company_info if we need SIC code or description
                
                # For basic searches, the ticker data is sufficient
                # This eliminates the 3-5 second delay from get_company_info
                result = {
                    "cik": company['cik'],
                    "name": company['name'],
                    "tickers": [company['ticker']] if company.get('ticker') else [],
                    "_source": "company_tickers_cache"  # Debug info
                }
                self._search_cache[company_input] = result
                return result
            else:
                return {"error": "No matching company found"}
    
    def get_company_filings_list(self, cik, form_types=['10-K', '10-Q']):
        """
        Get company filings list
        
        Args:
            cik (str): Company CIK
            form_types (list): List of form types
            
        Returns:
            list: Filings list
        """
        filings = self.edgar_client.get_company_filings(cik, form_types)
        return filings
    
    def extract_financial_metrics(self, cik, years=3):
        """
        Extract financial metrics for specified number of years (optimized)
        
        Args:
            cik (str): Company CIK
            years (int): Number of years to extract, default is 3 years
            
        Returns:
            list: List of financial data
        """
        # 实例级缓存检查（避免重复计算）
        cache_key = f"{cik}_{years}"
        if cache_key in self._extract_metrics_cache:
            return self._extract_metrics_cache[cache_key]
        
        financial_data = []
        
        # Step 1: Get company facts ONCE (will be cached)
        facts = self.edgar_client.get_company_facts(cik)
        if not facts:
            return []
        
        # Step 2: Get company filings ONCE to determine available years
        # Use tuple for caching compatibility
        filings_10k = self.edgar_client.get_company_filings(cik, ('10-K',))
        filings_20f = self.edgar_client.get_company_filings(cik, ('20-F',))
        all_annual_filings = filings_10k + filings_20f
        
        if not all_annual_filings:
            return []
        
        # Detect if company is a 20-F filer (foreign company)
        is_20f_filer = len(filings_20f) > 0 and len(filings_10k) == 0
        has_quarterly = False  # 20-F filers typically don't have quarterly reports
        
        # Step 3: Extract filing years from annual reports
        filing_year_map = {}  # Map: filing_year -> list of filings
        
        for filing in all_annual_filings:
            filing_date = filing.get('filing_date', '')
            if filing_date and len(filing_date) >= 4:
                try:
                    file_year = int(filing_date[:4])
                    if file_year not in filing_year_map:
                        filing_year_map[file_year] = []
                    filing_year_map[file_year].append(filing)
                except ValueError:
                    continue
        
        if not filing_year_map:
            return []
        
        # Step 4: Sort years in descending order and take the most recent N years
        sorted_years = sorted(filing_year_map.keys(), reverse=True)
        target_years = sorted_years[:years]
        
        # Step 5: Map filing years to fiscal years using facts (already fetched)
        filing_to_fiscal_year = {}  # Map: filing_year -> fiscal_year
        
        # Try to map filing years to fiscal years using Company Facts
        for data_source in ["us-gaap", "ifrs-full"]:
            if data_source in facts.get("facts", {}):
                source_data = facts["facts"][data_source]
                
                # Look for Revenue tag to get fiscal year mapping
                revenue_tags = ["Revenues", "RevenueFromContractWithCustomerExcludingAssessedTax", 
                               "Revenue", "RevenueFromContractWithCustomer"]
                
                for tag in revenue_tags:
                    if tag in source_data:
                        units = source_data[tag].get("units", {})
                        if "USD" in units:
                            for entry in units["USD"]:
                                form = entry.get("form", "")
                                fy = entry.get("fy", 0)
                                filed = entry.get("filed", "")  # Filing date
                                fp = entry.get("fp", "")
                                
                                # Map filing year to fiscal year
                                if form in ["10-K", "20-F"] and fy > 0 and filed and (fp == "FY" or not fp):
                                    if len(filed) >= 10:  # Format: YYYY-MM-DD
                                        try:
                                            file_year = int(filed[:4])
                                            # Store the mapping: filing_year -> fiscal_year
                                            if file_year not in filing_to_fiscal_year:
                                                filing_to_fiscal_year[file_year] = fy
                                        except ValueError:
                                            continue
                break  # Found revenue tag, no need to check more
        
        # Step 6: Generate period list for target years
        # For each year: FY -> Q4 -> Q3 -> Q2 -> Q1 (descending order)
        # For 20-F filers: only FY (no quarterly data)
        periods = []
        for file_year in target_years:
            # Try to get fiscal year from mapping, otherwise use filing year
            fiscal_year = filing_to_fiscal_year.get(file_year, file_year)
            
            # First add annual data for this fiscal year
            periods.append({
                'period': str(fiscal_year),
                'type': 'annual',
                'fiscal_year': fiscal_year,
                'filing_year': file_year
            })
            
            # Only add quarterly data for 10-K filers (not for 20-F filers)
            if not is_20f_filer:
                # Then add quarterly data in descending order: Q4, Q3, Q2, Q1
                for quarter in range(4, 0, -1):
                    periods.append({
                        'period': f"{fiscal_year}Q{quarter}",
                        'type': 'quarterly',
                        'fiscal_year': fiscal_year,
                        'filing_year': file_year
                    })
        
        # Step 7: Get financial data for each period
        for idx, period_info in enumerate(periods):
            period = period_info['period']
            fiscal_year = period_info['fiscal_year']
            
            data = self.edgar_client.get_financial_data_for_period(cik, period)
            
            if data and "period" in data:
                # Add fiscal year prefix for annual data
                if period_info['type'] == 'annual':
                    data["period"] = f"FY{fiscal_year}"
                
                # Add sequence number to maintain order
                data["_sequence"] = idx
                
                financial_data.append(data)
        
        # 缓存结果
        if financial_data:
            self._extract_metrics_cache[cache_key] = financial_data
        
        return financial_data
    
    def get_latest_financial_data(self, cik):
        """
        Get latest financial data (optimized)
        
        Args:
            cik (str): Company CIK
            
        Returns:
            dict: Latest financial data
        """
        # Get latest filing year (supports 10-K and 20-F)
        # Use tuple for caching
        filings_10k = self.edgar_client.get_company_filings(cik, ('10-K',))
        filings_20f = self.edgar_client.get_company_filings(cik, ('20-F',))
        filings = filings_10k + filings_20f
        
        if not filings:
            return {}
        
        # Get latest filing year
        latest_filing_year = None
        for filing in filings:
            if 'filing_date' in filing and filing['filing_date']:
                try:
                    filing_year = int(filing['filing_date'][:4])
                    if latest_filing_year is None or filing_year > latest_filing_year:
                        latest_filing_year = filing_year
                except ValueError:
                    continue
        
        if latest_filing_year is None:
            return {}
        
        # Get financial data for latest year
        return self.edgar_client.get_financial_data_for_period(cik, str(latest_filing_year))
    
    def format_financial_data(self, financial_data):
        """
        Format financial data for display
        
        Args:
            financial_data (dict or list): Financial data
            
        Returns:
            dict or list: Formatted financial data
        """
        if isinstance(financial_data, list):
            # Sort by _sequence to maintain correct order (FY -> Q4 -> Q3 -> Q2 -> Q1)
            sorted_data = sorted(financial_data, key=lambda x: x.get("_sequence", 999))
            formatted_data = []
            for data in sorted_data:
                formatted_data.append(self._format_single_financial_data(data))
            return formatted_data
        else:
            return self._format_single_financial_data(financial_data)
    
    def _format_single_financial_data(self, data):
        """
        Format single financial data entry - optimized structure
        
        Args:
            data (dict): Financial data with new optimized structure
            
        Returns:
            dict: Formatted financial data
        """
        formatted = {
            "period": data.get("period"),
            "_sequence": data.get("_sequence")
        }
        
        # Handle new optimized structure with metrics
        if "metrics" in data and isinstance(data["metrics"], dict):
            # Extract metrics to top level for backward compatibility
            for metric_key, metric_data in data["metrics"].items():
                if isinstance(metric_data, dict):
                    formatted[metric_key] = metric_data.get("value")
                else:
                    # Fallback for old format
                    formatted[metric_key] = metric_data
            
            # Add metadata to top level
            if "_metadata" in data:
                metadata = data["_metadata"]
                formatted["source_url"] = metadata.get("source_url")
                formatted["source_form"] = metadata.get("form")
                formatted["data_source"] = metadata.get("data_source")
        else:
            # Fallback: old format compatibility
            formatted.update(data)
        
        # Ensure all key fields exist, even if None
        key_fields = ['total_revenue', 'net_income', 'earnings_per_share', 
                     'operating_expenses', 'operating_cash_flow', 'source_url', 'source_form']
        for key in key_fields:
            if key not in formatted:
                formatted[key] = None
        
        # Format EPS, keep two decimal places
        if 'earnings_per_share' in formatted and isinstance(formatted['earnings_per_share'], (int, float)):
            formatted['earnings_per_share'] = round(formatted['earnings_per_share'], 2)
            
        return formatted