virgil.pipeline

Bases: _Pipeline

Search for a companion, set contrast limits, fit and sample a binary.

Stages: load, overview, search (point-companion grid, detection_statistics), limits (absil_limits), fit, posterior (NUTS, same prior as the fit) and quicklook. res.summary["companion"] has sep_mas and pa_deg for either template. Settings: pipeline settings.

Parameters:

Examples:

>>> import virgil as vg
>>> from virgil.pipeline import BinaryPipeline
>>> res = BinaryPipeline(vg.OIData("hd1234.oifits"), output="runs/binary").run()
>>> res.summary["companion"]["sep_mas"]

Bases: _Pipeline

Measure a star's angular diameter, with and without limb darkening.

Stages: load, overview, fit (a scan of the uniform-disk diameter, then fit of each other model from the scanned diameter), posterior (NUTS) and quicklook. The default models are compared on −2 log L against a BIC penalty. res.summary["star"]["diam_mas"] is the preferred model's. All settings: pipeline settings.

Parameters:

Examples:

>>> import virgil as vg
>>> from virgil.pipeline import StarPipeline
>>> res = StarPipeline(vg.OIData("star.oifits"), output="runs/star").run()
>>> res.summary["star"]["diam_mas"]

Bases: _Pipeline

Fit a disk, ring or halo with a MultipoleDisk.

The star is a grey point source unless star_diameter_mas, star_temperature_k or star_index says otherwise; the disk's spectral index is relative to it. Stages: load (wavel_range [min, max] in metres; error_floor a dict of absolute floors), overview (plots; sets the half-fringe λ/2B of the longest baseline), geometry (χ² grid over cos inc and pa, then a single-ring fit), profile, compare, posterior (NUTS, with posterior=True) and quicklook. model="gp" fits a RadialGP profile at each length scale and keeps the largest Laplace evidence; "rings" adds Gaussian rings until the BIC stops improving. run() returns a Result: res.summary["disk"] (n_rings and rings for "rings"), res.model() (the best fit), res.checks. All settings, with defaults: pipeline settings.

Parameters:

Examples:

>>> import virgil as vg
>>> from virgil.pipeline import DiskPipeline
>>> res = DiskPipeline(vg.OIData("disk.oifits"), "rings", output="runs/disk").run()
>>> res.summary["disk"]["n_rings"], res.model()

Bases: _Pipeline

Fit the orbit of a binary seen on several nights, blind to its period.

The data must be prepared already (the instrument's channel, band and noise handling). Stages: load (group into nights), epochs (epoch_positions), search (fit_orbit_aliases with no published period, window or prior), modes (rank the bands; compare with a single star), posterior (NUTS per significant band) and quicklook. summary["modes"]["winner"] is the winning alias band, ["delta_log_z"] its margin (nats) over the runner-up. All settings: pipeline settings.

Parameters:

Examples:

>>> from virgil.pipeline import OrbitPipeline
>>> res = OrbitPipeline.from_oifits(files, output="runs/orbit").run()
>>> res.summary["modes"]["winner"]["period"]

Bases: _Pipeline

Image the source from V², closure phases or complex visibilities.

A Gaussian-process image on the quoted errors; a good fit does not promise that the morphology is right. Multi-channel data give one grey image. Stages: load, overview (the grid), star (with or without a point source, by log evidence), hyperparameters (σ and ℓ by Laplace evidence), members (members fits at those shared hyperparameters), combine (images/mean.fits and sigma.fits) and quicklook. To spread the members over tasks, call pipeline.run(through="hyperparameters"), then pipeline.run_member(k) for k in range(members), then pipeline.run(), which returns the Result (res.image()). All settings, with defaults: pipeline settings.

Parameters:

Examples:

>>> from virgil.pipeline import ImagingPipeline
>>> res = ImagingPipeline(data, output="runs/image").run()
>>> res.image().pixels.shape

An image from Result.image: pixels (row 0 North, column 0 East), the centre offset dra, ddec (mas) and the FITS header (STARFRAC: star's flux share).

A model DiskPipeline can fit, registered under a short name.

Entries of MODELS are parallel options, not a chain: each runs the profile and compare stages on its own settings and checks. Adding an entry (say a cone) is adding a row to MODELS, with no new stage.

Attributes:

The outputs of a pipeline run, reloaded from its folder.

Nothing is recomputed: every accessor reads the files the run wrote, and returns plain numbers, numpy arrays or real virgil objects that can be refined by hand.

Attributes:

One quality check of a pipeline run.

Attributes:

Bases: ValueError

Resuming a run whose stored configuration differs from this one.