Skip to content

Ecallisto NG is a Python package tailored for interacting with Ecallisto data.

License

Notifications You must be signed in to change notification settings

i4Ds/ecallisto_ng

Repository files navigation

Ecallisto NG

Ecallisto NG is a comprehensive Python package tailored for interacting with Ecallisto data. It focuses on facilitating efficient data manipulation, processing, and visualization, leveraging the power of Pandas for data handling and PyTorch for advanced computations. The package is particularly optimized for dealing with large datasets, providing tools for slicing, filtering, and resampling data to make spectogram plotting more manageable.

Table of Contents

Background

The package is built with Python 3.9<3.12 and primarily uses the Pandas library for data manipulation. It's core functionality is centered around efficient data processing and visualization. The data provided by e-Callisto is stored in a pandas Dataframe, where the index is the time and the column names are the observed frequencies.

PyPI

Available on PyPI: https://pypi.org/project/ecallisto-ng/

Installation

To install this package, use pip for installation. Execute the following command in your terminal with basic functionality, such as plotting and data fetching: pip install ecallisto-ng If you want to use the virtual antenna: pip install ecallisto-ng[va] If you want to use the detection of flares, powered by ML: pip install ecallisto-ng[ml]

To use the notebook, please install the notebook dependencies.

Dev Installation

Of course, you can also install the package from source. To do so, clone the repository and install the package in editable mode: pip install -e . pip install -e .[va] pip install -e .[ml]

Pandas

Pandas is an open-source data analysis and manipulation tool, pivotal to Ecallisto NG. Learning Pandas is essential for effectively using Ecallisto NG, as it allows for sophisticated data slicing, filtering, and aggregation. More on Pandas: https://pandas.pydata.org/docs/

Examples and Usage

Please have a look at the jupyter notebook under example.

Getting data

get_ecallisto_data Function

This function fetches e-Callisto data within a specified date range and optional instrument regex pattern. It's suitable for smaller datasets. For larger datasets, consider using the get_ecallisto_data_generator function.

Parameters

  • start_datetime (datetime-like): The start date for the file search.
  • end_datetime (datetime-like): The end date for the file search.
  • instrument_string (str, List[str] or None): Instrument name(s) for file URL matching. If None, all files are considered.
  • freq_start (float or None): The start frequency for filtering.
  • freq_end (float or None): The end frequency for filtering.
  • base_url (str): Base URL of the remote file directory.

Returns

  • (dict of str: pandas.DataFrame) or pandas.DataFrame: A dictionary of instrument names and their corresponding dataframes. If only one instrument is found, it returns a single dataframe.

Example

from ecallisto_ng.data_download.downloader import (
    get_ecallisto_data,
)
from datetime import datetime

start = datetime(2021, 5, 7, 0, 00, 0)
end = datetime(2021, 5, 7, 23, 59, 0)
instrument_name = "Australia-ASSA_01"

dfs = get_ecallisto_data(start, end, instrument_name)
df = dfs[instrument_name] # Returns a dict of pd.Dataframes because instrument_name can also be a substring, e.g. "ASSA".

Getting data via a generator

get_ecallisto_data_generator Function

A generator function that yields e-Callisto data one file at a time within a date range. It's ideal for handling large datasets or when working with limited memory.

Parameters

  • start_datetime (datetime-like): The start date for the file search.
  • end_datetime (datetime-like): The end date for the file search.
  • instrument_name (List[str], str, or None): Instrument name(s) for file URL matching. If None, all files are considered.
  • freq_start (float or None): The start frequency for filtering.
  • freq_end (float or None): The end frequency for filtering.
  • base_url (str): Base URL of the remote file directory.

Yields

  • (str, pandas.DataFrame): A tuple containing the instrument name and its corresponding DataFrame.

Example

from ecallisto_ng.data_download.downloader import (
    get_ecallisto_data_generator,
)
from datetime import datetime

start = datetime(2021, 5, 7, 0, 00, 0)
end = datetime(2021, 5, 7, 23, 59, 0)
instrument_name = ["Australia-ASSA_01", "Australia-ASSA_02"]

data_generator = get_ecallisto_data_generator(start, end, instrument_name)
for instrument_name, data_frame in data_generator:
    print(instrument_name)
    print(f"{df.shape=}")

Plotting

Ecallisto NG provides basic plotting capabilities. Here's an example of how to generate a spectogram (make sure that df is defined):

from ecallisto_ng.plotting.plotting import plot_spectogram

plot_spectogram(df)

Make use of .resample to reduce the size of the data. Alternatively, you can pass a resolution parameter to the plot_spectogram. Here's an example:

plot_spectogram(df.resample("1min").max())
plot_spectogram(df, resolution=720) # Pixels

For more documentation on resample, please refer to the Pandas documentation.

Spectogram editing

We also provide some basic functionalities to edit the spectogram. Here's how you can do it (make sure that df is defined):

from ecallisto_ng.data_processing.utils import elimwrongchannels, subtract_constant_background, subtract_low_signal_noise_background
from datetime import datetime
from ecallisto_ng.plotting.plotting import plot_spectogram

# Filter keep frequencies only between 40 and 70 MHz
df = df.loc[:, 20:80]

# Select specific time
start = datetime(2021, 5, 7, 3, 39, 0)
end = datetime(2021, 5, 7, 3, 42, 0)
df = df.loc[start:end]

# Edit data
df = elimwrongchannels(df)
df = subtract_low_signal_noise_background(df)
df = subtract_constant_background(df)

plot_spectogram(df)

Additional Information

Note on .attrs and FITS Header

The function utilizes DataFrames with the .attrs attribute to store FITS header information. This attribute is a dictionary-like object and contains metadata about the FITS file, including header details. Accessing .attrs is essential for understanding the context of the data:

print(df.attrs)

These simple commands allow you to easily manipulate spectogram data, enabling effective use of the Ecallisto API for your needs.

Contributing

Contributions to Ecallisto NG are very welcome! If you have an idea for an improvement or have found a bug, please feel free to contribute. The preferred way to contribute is by submitting a Pull Request (PR) or creating an issue on our GitHub repository. This way, we can discuss potential changes or fixes and maintain the quality of the project.