khushalps
aa4fdc1087
git-subtree-dir: utils_v2 git-subtree-split: 76554fbaab5e1629c9d1918d8dcf3187d3dd4866
285 lines
12 KiB
Python
285 lines
12 KiB
Python
"""
|
|
|
|
AUTHOR:
|
|
|
|
Khushal P Soonderji
|
|
|
|
DATE:
|
|
|
|
Thursday, 28th Nov., 2024
|
|
|
|
OBJECTIVE:
|
|
|
|
The "Index Master" contains information about just the indices, and not the component symbols of those indices.
|
|
This script provides a way to get the data that is available on the screen on the following URL:
|
|
https://www.nseindia.com/market-data/live-market-indices
|
|
|
|
NOTE: This method involves web scraping. It is good for proof-of-concept development, but it is recommended that
|
|
more professional data-sources be used when the product starts becoming mature.
|
|
|
|
REFERENCES:
|
|
|
|
N/A
|
|
|
|
DOWNLOADS:
|
|
|
|
N/A
|
|
|
|
"""
|
|
|
|
|
|
# *****************************************************************************************************************
|
|
# ***** ****
|
|
# *** IMPORT ***
|
|
# ***** ****
|
|
# *****************************************************************************************************************
|
|
|
|
|
|
# To make sibling directories accessible for imports:
|
|
import sys
|
|
sys.path.append(".")
|
|
sys.path.append("..")
|
|
|
|
# System-level activities:
|
|
import io
|
|
|
|
# My utils:
|
|
from utils_v2.string import json
|
|
from utils_v2.string import regex
|
|
from utils_v2.date_time import date_time
|
|
|
|
# NSE-related utils:
|
|
from utils_v2.nse.controllers.base import AsyncNSEBase
|
|
from utils_v2.nse.models.api_call import NSEApiResponse
|
|
from utils_v2.nse.models.indices import NSEIndexInfo
|
|
|
|
# To make REST-ful API calls:
|
|
import httpx
|
|
|
|
# To work with date and time:
|
|
import datetime
|
|
|
|
# To work with datatypes:
|
|
from typing import Any, List
|
|
|
|
# For asynchronous activities:
|
|
import asyncio
|
|
|
|
|
|
# *****************************************************************************************************************
|
|
# ***** ****
|
|
# *** MACROS / ONE-TIME INIT ***
|
|
# ***** ****
|
|
# *****************************************************************************************************************
|
|
|
|
|
|
# --- Nothing Yet
|
|
|
|
|
|
# *****************************************************************************************************************
|
|
# ***** ****
|
|
# *** VARIABLES ***
|
|
# ***** ****
|
|
# *****************************************************************************************************************
|
|
|
|
|
|
# --- Nothing Yet
|
|
|
|
|
|
# *****************************************************************************************************************
|
|
# ***** ****
|
|
# *** FUNCTIONS ***
|
|
# ***** ****
|
|
# *****************************************************************************************************************
|
|
|
|
|
|
class NSEIndexMaster(AsyncNSEBase):
|
|
|
|
# Index Types:
|
|
INDEX_TYPE_BROAD_MARKET = "BROAD MARKET INDICES"
|
|
INDEX_TYPE_SECTORAL = "SECTORAL INDICES"
|
|
INDEX_TYPE_STRATEGY = "STRATEGY INDICES"
|
|
INDEX_TYPE_THEMATIC = "THEMATIC INDICES"
|
|
INDEX_TYPE_FIXED_INCOME = "FIXED INCOME INDICES"
|
|
|
|
def __init__(
|
|
self,
|
|
http_client: httpx.AsyncClient,
|
|
cookies_refresh_interval: int | float = 300,
|
|
debug = True,
|
|
debug_prefix = "NSE (IdxMstr) | ",
|
|
debug_only_errors = True
|
|
):
|
|
|
|
# Pass on the initialization to the parent:
|
|
super().__init__(
|
|
base_url = r"https://www.nseindia.com/market-data/live-market-indices",
|
|
data_url = r"https://www.nseindia.com/api/allIndices",
|
|
http_client = http_client,
|
|
cookies_refresh_interval = cookies_refresh_interval,
|
|
debug = debug,
|
|
debug_prefix = debug_prefix,
|
|
debug_only_errors = debug_only_errors
|
|
)
|
|
|
|
async def get_data(
|
|
self,
|
|
return_raw: bool = False,
|
|
refresh_cookies: bool = True,
|
|
force_refresh_cookies: bool = False,
|
|
retry_count: int = 1,
|
|
backoff_seconds: int | float = 0.5,
|
|
backoff_multiplier: float = 1.1
|
|
) -> NSEApiResponse:
|
|
|
|
"""
|
|
To get the data of the corporate event calendar.
|
|
:param return_raw: Whether you want the raw JSON from NSE or you want it formatted.
|
|
:param refresh_cookies: Whether, or not, you would like to refresh the cookies.
|
|
:param force_refresh_cookies: If set to True, cookies will be refreshed even if not timed out. If set to False,
|
|
cookies will be refreshed only when the interval passed to the constructor has elapsed since the last
|
|
successful refresh.
|
|
:param retry_count: The no. of times to try to get the data from the API.
|
|
:param backoff_seconds: The delay between unsuccessful API calls.
|
|
:param backoff_multiplier: The multiplier to add to the delay to change delay.
|
|
:return: The raw or formatted event calendar data in the 'data' field of the response model.
|
|
"""
|
|
|
|
# Make the API call:
|
|
api_response = await self.get(
|
|
refresh_cookies = refresh_cookies,
|
|
force_refresh_cookies = force_refresh_cookies,
|
|
retry_count = retry_count,
|
|
backoff_seconds = backoff_seconds,
|
|
backoff_multiplier = backoff_multiplier
|
|
)
|
|
|
|
# If the API call was successful:
|
|
if api_response.httpCode in [200]:
|
|
api_response.success = True
|
|
if return_raw: api_response.data = await api_response.get_json()
|
|
else:
|
|
try: api_response.data = self.format_data(
|
|
raw_json = await api_response.get_json(),
|
|
timestamp = date_time.get_current_utc_date_time(as_string = False),
|
|
raise_exception = True
|
|
)
|
|
except Exception as exception:
|
|
api_response.exception = exception
|
|
api_response.success = False
|
|
|
|
# Done here:
|
|
return api_response
|
|
|
|
@staticmethod
|
|
def parse_float(value: str | float | int) -> float | None:
|
|
|
|
"""
|
|
Parses a value as a floating point figure.
|
|
:param value: Any input value.
|
|
:return: A floating point value if parsed successfully, else None.
|
|
"""
|
|
|
|
if value is None: return None
|
|
if isinstance(value, float): return value
|
|
if isinstance(value, int): return float(value)
|
|
if isinstance(value, str):
|
|
value = regex.find_first(text = value.strip(), pattern = r"[\d,]+\.?[\d,]*")
|
|
if value: return float(value.replace(",", ""))
|
|
else: return None
|
|
|
|
def format_data(
|
|
self,
|
|
raw_json: dict,
|
|
timestamp: datetime.datetime = None,
|
|
raise_exception: bool = False
|
|
) -> List[NSEIndexInfo] | None:
|
|
|
|
"""
|
|
We format the data here to be able to retrieve it properly later.
|
|
:param raw_json: The raw data as scraped from NSE.
|
|
:param timestamp: The timestamp at which the data was scraped. This shall be useful for data retrieval from the
|
|
database, later.
|
|
:param raise_exception: If set to True, any exception will be propagated. If set to False, any exception will be
|
|
suppressed internally.
|
|
:return: The formatted data if successful, else None.
|
|
"""
|
|
|
|
# Can't do anything if the chain itself is null:
|
|
if raw_json is None: return raw_json
|
|
|
|
# Start by assuming failure:
|
|
formatted_data = None
|
|
|
|
# Ensure that we've got a proper timestamp:
|
|
if timestamp is None: timestamp = date_time.get_current_utc_date_time(as_string = False)
|
|
|
|
try:
|
|
|
|
# Format the data:
|
|
formatted_data = [
|
|
NSEIndexInfo(
|
|
scrapeTs = timestamp,
|
|
kind = idx["key"],
|
|
name = idx["index"],
|
|
symbol = idx["indexSymbol"],
|
|
open = self.parse_float(idx["open"]),
|
|
high = self.parse_float(idx["high"]),
|
|
low = self.parse_float(idx["low"]),
|
|
close = self.parse_float(idx["last"]),
|
|
prevClose = self.parse_float(idx["previousClose"]),
|
|
pChg = self.parse_float(idx["percentChange"]),
|
|
yearHigh = self.parse_float(idx["yearHigh"]),
|
|
yearLow = self.parse_float(idx["yearLow"]),
|
|
advances = self.parse_float(idx.get("advances")),
|
|
declines = self.parse_float(idx.get("declines")),
|
|
unchanged = self.parse_float(idx.get("unchanged")),
|
|
pChg30d = self.parse_float(idx["perChange30d"]),
|
|
pChg365d = self.parse_float(idx["perChange365d"])
|
|
) for idx in raw_json["data"]
|
|
]
|
|
|
|
# If something goes wrong:
|
|
except Exception as exception:
|
|
formatted_data = None
|
|
if raise_exception: raise
|
|
|
|
# Done here:
|
|
return formatted_data
|
|
|
|
|
|
# *****************************************************************************************************************
|
|
# ***** ****
|
|
# *** MAIN PROGRAM ***
|
|
# ***** ****
|
|
# *****************************************************************************************************************
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
async def main():
|
|
|
|
# Create an HTTP client:
|
|
test_client = httpx.AsyncClient(
|
|
limits = httpx.Limits(
|
|
max_connections = 100, # ............ Maximum number of connections allowed in the pool.
|
|
max_keepalive_connections = 50, # ... Maximum number of connections that can be kept alive.
|
|
),
|
|
timeout = httpx.Timeout(
|
|
pool = 120.0, # .... Time to wait for a free connection from the pool.
|
|
connect = 2.5, # ... Time to wait for establishing a connection to the server.
|
|
write = 10.0, # .... Time to wait for sending data.
|
|
read = 2.5 # ....... Time to wait for receiving data.
|
|
)
|
|
)
|
|
|
|
# Create an instance of the scraper, and refresh its cookies:
|
|
my_nse = NSEIndexMaster(http_client = test_client)
|
|
|
|
# Get and show the data:
|
|
api_response = await my_nse.get_data(return_raw = False)
|
|
print("SUMMARY:", api_response.to_markdown(), "\n---\n\n")
|
|
if api_response.success: print("INDEX MASTER:", json.to_string(api_response.data, default = str))
|
|
|
|
asyncio.run(main())
|