pandas_market_calendars | Exchange calendars to use with pandas for trading

I'm writing a trading algorithm that incorporates multiple timeframe analysis by comparing yearly OHLC to quarterly OHLC to monthly OHLC to weekly OHLC to daily OHLC.

For each of these different timeframe, I need the start date for the previous 4 periods of OHLC data, which includes the current period.

Examples, relative to 2022-01-09, the date of this post:

Yearly timeframe 4 year start date: 2019-01-02

Quarterly timeframe 4 quarter start date: 2021-03-01

Monthly timeframe 4 month start date: 2021-10-01

Weekly timeframe 4 week start date: 2021-12-13

Daily timeframe 4 day start date: 2022-01-04

Is there a finance markets-centric module for python that allows easy referencing to these kinds of date stamps that I can ultimately output as strings in the format of the aforementioned examples to insert into my polygon REST calls?

I have looked at various datetime and dateutil classes and methods, but they're all rather complex, notwithstanding they don't appear to account for nuances like days that the North American equities markets are open.

(It'd be super cool if something existed that could understand English language expression; something like: 'first market day of the year four years ago', 'first market day of the quarter, four quarters ago', 'first market day of the month, four months ago', 'first market day of the week, four weeks ago', 'day of the week, four market days ago').

Solution

pandas has a very comprehensive list of features that provides what I was looking for. Here's what I put together based on those findings:

With the start_of_day() function below, I'm attempting to return timestamps of the NYSE market open date and time for the current week. I am not using this in a pandas dataframe, I'm just using pandas time functions because they're some of the only functions that I've found that are specific to market dates and times that incorporate holidays.

In this particular case, I'm trying to return timestamps of:

I have the following code creating a table of business days in NYSE during the past 5 years:

I need to get the most recent date that there was trading on the NYSE exchange. For example, there is no trading on 11/25/2021, on the NYSE exchange. Therefore on November 26, the most recent trading day should be November 24. This is what I have so far but it's not working.

I'm trying things out on windows vs linux where I have this working in 3.8 and 3.9.5, but not on windows using anaconda

I have this df

I've got an OHLCV financial dataset I'm working with that's approx 1.5M rows. It's for a single security back to 2000 at a 1min resolution. The dataset had all the zero volume timestamps removed so has gaps in the trading day which I don't want for my purposes.

Raw data (df1) looks like:

Timestamp, Open, High, Low, Close, Volume

To fill in the all the zero volume timestamps I created an empty trading calendar (df2) using (pandas_market_calendars which is an amazing time saver) and I've then used df3 = merge_asof(df2, df1, on='Timestamp', direction = nearest) to fill all the timestamps. This is the behaviour I wanted for the price data (OHLC) but not for volume. I need all the 'filled' timestamps to show zero volume so I figured a lambda function would suit (as below) to check if each timestamp was in the original dataframe (df1) or not.