Pulsar Data#

Discovery uses lightweight Pulsar objects saved as Arrow Feather files for efficient data storage and loading.

Data Format#

Pulsar data is stored in the Apache Arrow Feather format, which provides:

  • Fast read/write performance

  • Efficient columnar storage

  • Cross-language compatibility

  • Minimal memory overhead

Each Feather file contains:

  • Time-of-arrival (TOA) data

  • Observing frequencies

  • Measurement uncertainties

  • Design matrix (timing model)

  • Pulsar position and metadata

  • Optional: noise parameter defaults

Converting from Enterprise#

If you have existing Enterprise Pulsar objects, you can convert them to Discovery format:

import discovery as ds
import enterprise

# Load Enterprise pulsar
ent_psr = enterprise.Pulsar(parfile, timfile)

# Convert and save as Feather file
ds.Pulsar.save_feather(ent_psr, 'B1855+09.feather', noisedict)

The noisedict parameter is optional and allows you to include default noise parameter values (e.g., EFAC, EQUAD values) in the Feather file.

Loading Pulsar Data#

Loading a Discovery pulsar is straightforward:

import discovery as ds

# Load single pulsar
psr = ds.Pulsar.read_feather('data/v1p1_de440_pint_bipm2019-B1855+09.feather')

# Load multiple pulsars
import glob
psrs = [ds.Pulsar.read_feather(f)
        for f in glob.glob('data/v1p1_de440_pint_bipm2019-*.feather')]

Pulsar Object Attributes#

A Discovery Pulsar object contains:

TOA Data

  • psr.toas - Times of arrival (MJD, in seconds)

  • psr.residuals - Timing residuals (seconds)

  • psr.toaerrs - TOA uncertainties (seconds)

Observing Information

  • psr.freqs - Observing frequencies (MHz)

  • psr.backends - Backend/receiver identifiers

  • psr.flags - Additional observational flags

Timing Model

  • psr.Mmat - Design matrix for timing model

  • psr.pos - Pulsar sky position (unit vector)

Solar System

  • psr.planetssb - Planet positions (solar system barycenter)

  • psr.sunssb - Sun position (solar system barycenter)

Metadata

  • psr.name - Pulsar name (e.g., ‘B1855+09’)

  • psr.noisedict - Dictionary of noise parameters (if saved)

Example Datasets#

Discovery includes example datasets based on NANOGrav 15-year data in the data/ folder:

data/
├── v1p1_de440_pint_bipm2019-B1855+09.feather
├── v1p1_de440_pint_bipm2019-B1937+21.feather
├── v1p1_de440_pint_bipm2019-J0030+0451.feather
└── ...

These can be used for testing and development:

# Load example pulsar
psr = ds.Pulsar.read_feather('data/v1p1_de440_pint_bipm2019-B1855+09.feather')

print(f"Pulsar: {psr.name}")
print(f"TOAs: {len(psr.toas)}")
print(f"Span: {(psr.toas.max() - psr.toas.min()) / 365.25 / 86400:.1f} years")

Working with Pulsar Data#

Accessing residuals:

# Residuals as numpy array
res = psr.residuals

# Can be used directly in likelihood
signals = [res, ...]

See Also#