import os, json, asyncio from datetime import datetime from lxml import etree from playwright.async_api import async_playwright # Config API_URL = "https://www.nseindia.com/api/corporates-financial-results?index=equities&symbol=AAKASH&period=Quarterly" TO_DATE_THRESHOLD = datetime.strptime("31-Dec-2024", "%d-%b-%Y") OUTPUT_DIR = './json/' os.makedirs(OUTPUT_DIR, exist_ok=True) # Pre-generated quarter code mapping from 30-Sep-2022 (code=115) quarter_code_map = { "30-Sep-2022": 115, "31-Dec-2022": 116, "31-Mar-2023": 117, "30-Jun-2023": 118, "30-Sep-2023": 119, "31-Dec-2023": 120, "31-Mar-2024": 121, "30-Jun-2024": 122, "30-Sep-2024": 123, "31-Dec-2024": 124, "31-Mar-2025": 125 } def get_first(result): return result[0].text.strip() if result and result[0].text else None def parse_nse_date(raw_date): for fmt in ("%d-%b-%Y", "%Y-%m-%d"): try: return datetime.strptime(raw_date, fmt).strftime("%d-%b-%y") except: continue raise ValueError(f"Unrecognized date format: {raw_date}") def extract_data(xml_bytes): root = etree.fromstring(xml_bytes) ns = {'in-bse-fin': "http://www.bseindia.com/xbrl/fin/2020-03-31/in-bse-fin"} xp = lambda path: root.xpath(path, namespaces=ns) def gstr(path): return get_first(xp(path)) or "" def gnum(path): val = get_first(xp(path)) return str(round(float(val.replace(',', '')) / 1_000_000, 2)) if val else "0" data = { 'symbol': gstr("//in-bse-fin:Symbol[@contextRef='OneD']"), 'company_name': gstr("//in-bse-fin:NameOfTheCompany[@contextRef='OneD']"), 'Net Sales': gnum("//in-bse-fin:RevenueFromOperations[@contextRef='OneD']"), 'Other Income': gnum("//in-bse-fin:OtherIncome[@contextRef='OneD']"), 'Total Income': gnum("//in-bse-fin:Income[@contextRef='OneD']"), 'Expenditure': gnum("//in-bse-fin:Expenses[@contextRef='OneD']"), 'Cost of Materials Consumed': gnum("//in-bse-fin:CostOfMaterialsConsumed[@contextRef='OneD']"), 'Finance Costs': gnum("//in-bse-fin:FinanceCosts[@contextRef='OneD']"), 'Changes in inventories of finished goods, work-in-progress and stock-in-trade': gnum("//in-bse-fin:ChangesInInventoriesOfFinishedGoodsWorkInProgressAndStockInTrade[@contextRef='OneD']"), 'Purchases of stock-in-trade': gnum("//in-bse-fin:PurchasesOfStockInTrade[@contextRef='OneD']"), 'Employee benefit expense': gnum("//in-bse-fin:EmployeeBenefitExpense[@contextRef='OneD']"), 'Depreciation and amortisation expense': gnum("//in-bse-fin:DepreciationDepletionAndAmortisationExpense[@contextRef='OneD']"), 'Profit after Interest but before Exceptional Items': gnum("//in-bse-fin:ProfitBeforeExceptionalItemsAndTax[@contextRef='OneD']"), 'Exceptional Item': gnum("//in-bse-fin:ExceptionalItemsBeforeTax[@contextRef='OneD']"), 'Profit (+)/ Loss (-) from Ordinary Activities before Tax': gnum("//in-bse-fin:ProfitBeforeTax[@contextRef='OneD']"), 'Tax': gnum("//in-bse-fin:TaxExpense[@contextRef='OneD']"), 'Net Profit (+)/ Loss (-) from Ordinary Activities after Tax': gnum("//in-bse-fin:ProfitLossForPeriodFromContinuingOperations[@contextRef='OneD']"), 'Net Profit': gnum("//in-bse-fin:ProfitLossForPeriod[@contextRef='OneD']"), 'Current tax': gnum("//in-bse-fin:CurrentTax[@contextRef='OneD']"), 'Deferred tax': gnum("//in-bse-fin:DeferredTax[@contextRef='OneD']"), 'Other Comprehensive Income Net of Taxes': gnum("//in-bse-fin:OtherComprehensiveIncomeNetOfTaxes[@contextRef='OneD']"), 'Total Comprehensive Income for the Period': gnum("//in-bse-fin:ComprehensiveIncomeForThePeriod[@contextRef='OneD']"), 'Share of profit(loss) of associates and joint ventures': gnum("//in-bse-fin:ShareOfProfitLossOfAssociatesAndJointVenturesAccountedForUsingEquityMethod[@contextRef='OneD']"), 'Basic EPS for continuing operation': gstr("//in-bse-fin:BasicEarningsLossPerShareFromContinuingOperations[@contextRef='OneD']"), 'Diluted EPS for continuing operation': gstr("//in-bse-fin:DilutedEarningsLossPerShareFromContinuingOperations[@contextRef='OneD']"), 'Basic for discontinued & continuing operation': gstr("//in-bse-fin:BasicEarningsLossPerShareFromContinuingAndDiscontinuedOperations[@contextRef='OneD']"), 'Diluted for discontinued & continuing operation': gstr("//in-bse-fin:DilutedEarningsLossPerShareFromContinuingAndDiscontinuedOperations[@contextRef='OneD']"), 'Data Type': gstr("//in-bse-fin:NatureOfReportStandaloneConsolidated[@contextRef='OneD']") } for i in range(1, 50): ctx = f"OneOperatingExpenses0{i}D" desc = gstr(f"//in-bse-fin:DescriptionOfOtherExpenses[@contextRef='{ctx}']") if desc.lower() in ["other expenses", "other expense", "others"]: data["Other Expenses"] = gnum(f"//in-bse-fin:OtherExpenses[@contextRef='{ctx}']") break else: data["Other Expenses"] = gnum("//in-bse-fin:OtherExpenses[@contextRef='OneD']") return data async def fetch_symbol_quarters(): async with async_playwright() as p: browser = await p.firefox.launch(headless=True) context = await browser.new_context() page = await context.new_page() print("🌐 Visiting NSE homepage...") await page.goto("https://www.nseindia.com", timeout=60000) await page.wait_for_timeout(3000) print("📦 Fetching Financial Results for symbol...") response = await context.request.get(API_URL, headers={ "Accept": "application/json", "Referer": "https://www.nseindia.com/" }) json_data = await response.json() print(f"✅ Total docs: {len(json_data)}") for item in json_data: to_date_str = item.get("toDate", "") to_date = datetime.strptime(to_date_str, "%d-%b-%Y") if to_date < TO_DATE_THRESHOLD: continue qtr_code = quarter_code_map.get(to_date_str) if not qtr_code: print(f"⚠️ Skipping unknown quarter: {to_date_str}") continue xbrl_url = item["xbrl"].split("
")[0] print(f"➡️ Fetching XBRL for {item['symbol']} | {to_date_str}") try: res = await page.goto(xbrl_url, timeout=60000) if not res or not res.ok: print(f"❌ Failed to fetch XML: {xbrl_url}") continue xml = await res.text() xml_bytes = xml.encode("utf-8") data = extract_data(xml_bytes) data.update({ "isin_number": item["isin"], "Type": item["audited"], "Date Begin": parse_nse_date(item["fromDate"]), "Date End": parse_nse_date(item["toDate"]), "Description": "Amount (Rs. million)", "qtr_code": qtr_code, "image_id": item.get("image_id"), "image_url": item.get("image_url", ""), "BSE/NSE": "NSE" }) filename = os.path.join(OUTPUT_DIR, f"{item['symbol']}_{to_date_str.replace('-', '')}_{data['Data Type']}.json") with open(filename, "w") as f: json.dump(data, f, indent=2) print(f"✅ Saved {filename}") except Exception as e: print(f"❌ Error for {item['symbol']}: {e}") continue await browser.close() print("🏁 Done.") if __name__ == "__main__": asyncio.run(fetch_symbol_quarters())