Skip to content

Stocks module

The Stock object is the core class handling the data read/write. Multiple Stocks object can be combined into one StockDict object which is kind of a dataholder.


flowchart LR

st1[Stock]
dhm[(Local <br> database)]

st1 --data for stock from <br>*t1* to *t2* timestamp--> dhm
dhm --OHLCV dataframe--> st1

Symbol([Symbol]) & e(["`Exchange<br>(NSE)`"]) & et(["`Exchange<br>type<br>(Cash)`"]) --> st1

st1 --> sd1@{ shape: procs, label: "StockDict"}

The Stock object takes in the input of Exchange and Exchange_type, which tells the type of the security. The Stock object handles the reading and writing of the data to the local database. The Stock class uses duckdb for managing database with .parquet files. This enables quicker read/write.

Saving the data

The local_data_foldpath should be the path to the parent folder. The Stock object takes care of making a subdirectory within this foldpath.

s1.save_historical_data(data=data, interval=interval, local_data_foldpath=local_data_foldpath, overwrite=False)
The data is saved as month wise files.

Loading the data from local storage

You don't need to log-in the client to read from local data. In a way, this enables us to download the data independently of the strategies. The strategy only needs to interact with the local data.

start_date, end_date = "2025-01-01", "2026-01-14"
interval = fm.constants.INTERVAL.one_day

data = s1.load_historical_data(start=start_date, end=end_date, interval=interval, local_data_foldpath=local_data_foldpath)
data.info()

<class 'pandas.core.frame.DataFrame'>
DatetimeIndex: 258 entries, 2025-01-01 to 2026-01-13
Data columns (total 5 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   Open    258 non-null    float64
 1   High    258 non-null    float64
 2   Low     258 non-null    float64
 3   Close   258 non-null    float64
 4   Volume  258 non-null    int64  
dtypes: float64(4), int64(1)
memory usage: 12.1 KB