Squashed 'utils_v2/' content from commit d593c1ec

git-subtree-dir: utils_v2
git-subtree-split: d593c1ec43b43c6881c40cbef6ae7c170dc76776
This commit is contained in:
2025-01-07 10:13:06 +05:30
commit 6d1f731ff7
185 changed files with 141092 additions and 0 deletions
@@ -0,0 +1,332 @@
"""
AUTHOR:
Khushal P Soonderji
DATE:
Thursday, 5th Dec., 2024
OBJECTIVE:
To provide a way to retrieve the list of holidays from NSE and to check if today is a holiday.
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.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
# 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
# For debugging:
import inspect
# *****************************************************************************************************************
# ***** ****
# *** MACROS / ONE-TIME INIT ***
# ***** ****
# *****************************************************************************************************************
# --- Nothing Yet
# *****************************************************************************************************************
# ***** ****
# *** VARIABLES ***
# ***** ****
# *****************************************************************************************************************
# --- Nothing Yet
# *****************************************************************************************************************
# ***** ****
# *** FUNCTIONS ***
# ***** ****
# *****************************************************************************************************************
class NSETradingHolidayCalendar(AsyncNSEBase):
def __init__(
self,
http_client: httpx.AsyncClient,
cookies_refresh_interval: int | float = 300,
debug = True,
debug_prefix = "NSE (Hol) | ",
debug_only_errors = True
):
# Pass on the initialization to the parent:
super().__init__(
base_url = r"https://www.nseindia.com/resources/exchange-communication-holidays",
data_url = r"https://www.nseindia.com/api/holiday-master?type=trading",
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 holiday 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 format_data(
raw_json: dict,
timestamp: datetime.datetime = None,
raise_exception: bool = False
) -> List[dict] | 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:
# Start by creating an empty dict:
formatted_data = {}
# Process each category in the original data
for category, events in raw_json.items():
for event in events:
trading_date = event["tradingDate"]
# If the tradingDate is not in the dictionary, initialize it
if trading_date not in formatted_data:
formatted_data[trading_date] = {
"scrapeTs": timestamp,
"date": date_time.to_timezone(
date_time.as_if_timezone(
datetime.datetime.strptime(trading_date, "%d-%b-%Y"),
timezone = date_time.TIMEZONE_IST
),
timezone = date_time.TIMEZONE_UTC
),
"weekDay": event["weekDay"],
"event": event["description"],
"morningSession": event["morning_session"],
"eveningSession": event["evening_session"],
"holidayType": []
}
# Add the current category to the "type" list
formatted_data[trading_date]["holidayType"].append(category)
# Now we get rid of the unnecessary keys and make it a list:
formatted_data = [v for v in formatted_data.values()]
# If something goes wrong:
except Exception as exception:
formatted_data = None
if raise_exception: raise
# Done here:
return formatted_data
async def is_holiday(
self,
timestamp: datetime.datetime
) -> NSEApiResponse:
"""
Checks if a given date is a trading holiday, or not.
:param timestamp: The datetime instance that you want to check for it being a holiday. Preferably make it an
aware instance. If a naive instance is passed, it will be assumed to be in UTC.
:return: True if it is a holiday, False if it isn't. Can be None if something goes wrong in fetching the data.
The value will be in the 'data' field of the response model.
"""
# We first prepare UTC and IST versions of the incoming datetime:
timestamp_utc = date_time.to_timezone(timestamp, timezone = date_time.TIMEZONE_UTC)
timestamp_ist = date_time.to_timezone(timestamp_utc, timezone = date_time.TIMEZONE_IST)
# If this day is a weekend, it is a holiday by default:
ist_dow = timestamp_ist.weekday()
if ist_dow in [5, 6]: return NSEApiResponse(
action = inspect.stack()[0].function,
url = None,
method = None,
success = True,
data = {
"isHoliday": True,
"event": "Saturday" if ist_dow == 5 else "Sunday"
},
message = "Holiday due to weekend."
)
# First we fetch the data from NSE:
api_response = await self.get_data(return_raw = False)
api_response.action = inspect.stack()[0].function
if not api_response.success: return api_response
# Now we check if our date to check is in the list:
is_holiday = False
event = None
for h in api_response.data:
if h["date"].date() == timestamp_utc.date():
is_holiday = True
event = h["event"]
break
api_response.data = {
"isHoliday": is_holiday,
"event": event
}
# Done here:
return api_response
# *****************************************************************************************************************
# ***** ****
# *** 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 = NSETradingHolidayCalendar(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("HOLIDAY CALENDAR:", json.to_string(api_response.data, default = str))
print("\n---\n\n")
# Check for a holiday:
date_to_check = date_time.parse_date_time(
input_value = "2024-12-25 00:00:00",
date_formats = ["%Y-%m-%d %H:%M:%S"],
timezone = date_time.TIMEZONE_IST
)
api_response = await my_nse.is_holiday(date_to_check)
print("SUMMARY:", api_response.to_markdown(), "\n---\n\n")
print("DATETIME:", date_to_check)
print("IS HOLIDAY:", api_response.data)
asyncio.run(main())