insider_trading.py

import pandas as pd
import numpy as np
from datetime import date,time,timedelta
from nsepy import get_history
import copy
import requests
import datetime

#@st.cache()
def fetchInsiderData(lookback_period):
    today = datetime.datetime.now().date()
    lookback = datetime.datetime.now().date() - datetime.timedelta(days=lookback_period)
    today = today.strftime("%d-%m-%Y")
    lookback = lookback.strftime("%d-%m-%Y")

    headers = {
        'User-Agent': 'Apache/2.4.34 (Ubuntu) OpenSSL/1.1.1 (internal dummy connection)',
        'Accept': '*/*',
        'Accept-Language': 'en-US,en;q=0.5',
        'DNT': '1',
        'Connection': 'keep-alive',
        'Referer': 'https://www.nseindia.com/companies-listing/corporate-filings-insider-trading',
    }

    params = (
        ('index', 'equities'),
        ('from_date', lookback),
        ('to_date', today),
    )

    # If response recieved
    while True:
        print("Sending request")
        s = requests.Session()
        r = s.get('https://www.nseindia.com/companies-listing/corporate-filings-insider-trading',
                  headers=headers,
                  params=params)
        r = s.get('https://www.nseindia.com/api/corporates-pit',
                  headers=headers,
                  params=params,
                  cookies=s.cookies.get_dict())
        print("response received")
        if r.status_code == 200:
            data_dict = r.json()
            keys = list(data_dict.keys())
            df = pd.DataFrame(data_dict[keys[1]])
            print("Data fetched")
            break
        else:
            print("Insider Trading Data Denied")
            continue

    df = df[["symbol","company", "anex", "acqName", "personCategory", "secType", "befAcqSharesNo",
            "befAcqSharesPer", "tkdAcqm", "secAcq", "secVal","tdpTransactionType", "securitiesTypePost",
            "afterAcqSharesNo" , "afterAcqSharesPer", "acqfromDt", "acqtoDt", "intimDt", "acqMode",
            "derivativeType", "tdpDerivativeContractType", "buyValue", "buyQuantity",
            "sellValue", "sellquantity", "exchange",  "remarks", "date", "xbrl"]]

    df.columns = ['SYMBOL', 'COMPANY', 'REGULATION', 'NAME OF THE ACQUIRER/DISPOSER',
       'CATEGORY OF PERSON', 'TYPE OF SECURITY (PRIOR)',
       'NO. OF SECURITY (PRIOR)', '% SHAREHOLDING (PRIOR)',
       'TYPE OF SECURITY (ACQUIRED/DISPLOSED)',
       'NO. OF SECURITIES (ACQUIRED/DISPLOSED)',
       'VALUE OF SECURITY (ACQUIRED/DISPLOSED)',
       'ACQUISITION/DISPOSAL TRANSACTION TYPE', 'TYPE OF SECURITY (POST)',
       'NO. OF SECURITY (POST)', '% POST',
       'DATE OF ALLOTMENT/ACQUISITION FROM',
       'DATE OF ALLOTMENT/ACQUISITION TO', 'DATE OF INITMATION TO COMPANY',
       'MODE OF ACQUISITION', 'DERIVATIVE TYPE SECURITY',
       'DERIVATIVE CONTRACT SPECIFICATION', 'NOTIONAL VALUE(BUY)',
       'NUMBER OF UNITS/CONTRACT LOT SIZE (BUY)', 'NOTIONAL VALUE(SELL)',
       'NUMBER OF UNITS/CONTRACT LOT SIZE  (SELL)', 'EXCHANGE', 'REMARK',
       'BROADCASTE DATE AND TIME', 'XBRL']
    # Saving to a csv file
    # This is daily data
    df.to_csv(r"Insider Trading {today}.csv", index=False)
    print("Insider Data Saved!")

def process(lookback_input):
    today = datetime.datetime.now().date()
    l = datetime.datetime.now().date() - datetime.timedelta(days=lookback_input)
    today = today.strftime("%d-%m-%Y")
    lookback = l.strftime("%d-%m-%Y")
    end_date = date.today().strftime("%d-%m-%Y")
    fetchInsiderData(lookback_input)
    print('back in process')
    data = pd.read_csv(r"Insider Trading {today}.csv")
    new_columns = []

    for col in data.columns:
        new_columns.append(col.split(" \n")[0])

    data.columns = new_columns
    data = data[data["CATEGORY OF PERSON"].isin(["Promoters", "Promoter Group"])]
    data_last_txn = copy.deepcopy(data)
    data_promoter_sell = data[data["MODE OF ACQUISITION"] == "Market Sale"]
    data = data[data["MODE OF ACQUISITION"] == "Market Purchase"]

    data = data[["SYMBOL", "VALUE OF SECURITY (ACQUIRED/DISPLOSED)", "NO. OF SECURITIES (ACQUIRED/DISPLOSED)"]]
    data_promoter_sell = data_promoter_sell[["SYMBOL", "VALUE OF SECURITY (ACQUIRED/DISPLOSED)", "NO. OF SECURITIES (ACQUIRED/DISPLOSED)"]]
    data_last_txn = data_last_txn[["SYMBOL", "NAME OF THE ACQUIRER/DISPOSER", "VALUE OF SECURITY (ACQUIRED/DISPLOSED)",
                                                     "NO. OF SECURITIES (ACQUIRED/DISPLOSED)", "MODE OF ACQUISITION",
                                                    "DATE OF ALLOTMENT/ACQUISITION TO"]]

    data_last_txn = data_last_txn[data_last_txn["VALUE OF SECURITY (ACQUIRED/DISPLOSED)"]!='-']
    data_last_txn = data_last_txn[data_last_txn["DATE OF ALLOTMENT/ACQUISITION TO"]!='-']

    data_last_txn["DATE OF ALLOTMENT/ACQUISITION TO"].iloc[0]

    ser1 = data["VALUE OF SECURITY (ACQUIRED/DISPLOSED)"].astype("float")
    ser2 = data["NO. OF SECURITIES (ACQUIRED/DISPLOSED)"].astype("float")

    ser3 = data_promoter_sell["VALUE OF SECURITY (ACQUIRED/DISPLOSED)"].astype("float")
    ser4 = data_promoter_sell["NO. OF SECURITIES (ACQUIRED/DISPLOSED)"].astype("float")

    ser5 = data_last_txn["VALUE OF SECURITY (ACQUIRED/DISPLOSED)"].astype("float")
    ser6 = data_last_txn["NO. OF SECURITIES (ACQUIRED/DISPLOSED)"].astype("float")
    ser7 = pd.to_datetime(pd.Series(data_last_txn["DATE OF ALLOTMENT/ACQUISITION TO"]), format="%d-%b-%Y")
    # print(data_last_txn[["SYMBOL", "DATE OF ALLOTMENT/ACQUISITION TO"]])
    data_last_txn_symbolwise = data_last_txn[["SYMBOL", "DATE OF ALLOTMENT/ACQUISITION TO"]].sort_values(by = ["SYMBOL", "DATE OF ALLOTMENT/ACQUISITION TO"], ascending = [True,False])#.drop_duplicates(subset=["SYMBOL"], keep='first').sort_values(by = ["DATE OF ALLOTMENT/ACQUISITION TO"], ascending = False)
    print(data_last_txn_symbolwise)

    data["VALUE OF SECURITY (ACQUIRED/DISPLOSED)"] = ser1
    data["NO. OF SECURITIES (ACQUIRED/DISPLOSED)"] = ser2

    data_promoter_sell["VALUE OF SECURITY (ACQUIRED/DISPLOSED)"] = ser3
    data_promoter_sell["NO. OF SECURITIES (ACQUIRED/DISPLOSED)"] = ser4

    data_last_txn["VALUE OF SECURITY (ACQUIRED/DISPLOSED)"] = ser5
    data_last_txn["NO. OF SECURITIES (ACQUIRED/DISPLOSED)"] = ser6
    data_last_txn["DATE OF ALLOTMENT/ACQUISITION TO"] = ser7

    data = data.groupby('SYMBOL')[['VALUE OF SECURITY (ACQUIRED/DISPLOSED)','NO. OF SECURITIES (ACQUIRED/DISPLOSED)']].sum().reset_index()
    data_promoter_sell = data_promoter_sell.groupby('SYMBOL')[['VALUE OF SECURITY (ACQUIRED/DISPLOSED)','NO. OF SECURITIES (ACQUIRED/DISPLOSED)']].sum().reset_index()
    data.sort_values(by = "VALUE OF SECURITY (ACQUIRED/DISPLOSED)", ascending=False, inplace = True)

    data_promoter_sell.sort_values(by = "VALUE OF SECURITY (ACQUIRED/DISPLOSED)", ascending=False, inplace = True)

    data.rename(columns = {"VALUE OF SECURITY (ACQUIRED/DISPLOSED)" : "Net Acquired Value", "NO. OF SECURITIES (ACQUIRED/DISPLOSED)": "Net Acquired Qty"}, inplace=True)

    data_promoter_sell.rename(columns = {"VALUE OF SECURITY (ACQUIRED/DISPLOSED)" : "Net Disposed Value", "NO. OF SECURITIES (ACQUIRED/DISPLOSED)": "Net Disposed Qty"}, inplace = True)

    merged_data = pd.merge(data,data_promoter_sell, how = "left").fillna(0)


    merged_data["Net Value Acquired"] = merged_data["Net Acquired Value"] - merged_data["Net Disposed Value"]

    merged_data["Net Qty Acquired"] = merged_data["Net Acquired Qty"] - merged_data["Net Disposed Qty"]

    merged_data.sort_values(by = "Net Value Acquired", ascending = False, inplace = True)

    merged_data["Avg buy price"] = round(merged_data["Net Value Acquired"]/merged_data["Net Qty Acquired"],2)

    merged_data = merged_data[["SYMBOL", "Net Value Acquired","Net Qty Acquired", "Avg buy price"]]

    merged_data = merged_data[merged_data["Net Value Acquired"] >= 10000000]

   # ltp = []
   # i = 0
   # for sym in merged_data["SYMBOL"]:
   #     try:
   #         print('getting history for ',sym)
   #         ltp.append(get_history(sym,date.today() - timedelta(5), date.today())["Close"][-1:][0])
   #     except:
   #         print("ERROR FOR ", sym)
   #         ltp.append(0)
   #     print(i, sym)
   #     i+=1
   # merged_data["LTP"] = ltp

   # merged_data["% diff"] = round((merged_data["LTP"] - merged_data["Avg buy price"])*100/merged_data["Avg buy price"],2)

   # merged_data = merged_data[merged_data["LTP"]!=0]


    merged_data.sort_values(by = ["Net Value Acquired"], ascending=[True], inplace = True)

    txns = {}
    for i in range(len(merged_data)):
        try:
            sym = merged_data.iloc[i]["SYMBOL"]
            curr_txns = data_last_txn[data_last_txn["SYMBOL"] == sym]
            txns[sym] = curr_txns
            txns[sym]["Avg. txn price"] = round(txns[sym]["VALUE OF SECURITY (ACQUIRED/DISPLOSED)"]/txns[sym]["NO. OF SECURITIES (ACQUIRED/DISPLOSED)"],3)
        except:
            print("txns error for ", merged_data.iloc[i]["SYMBOL"])
   # merged_data = merged_data[merged_data["% diff"] <= 5].sort_values(by = "Net Value Acquired", ascending=False)

    print(merged_data)
    return merged_data,txns


lookback = 90
df,txns = process(int(lookback))
print(df)
print(txns)