17657445007

Committed 11 Sep 2025 09:12PM UTC coverage: 89.856% (-0.06%) from 89.917%

Build # 17657445007

Build Type

push

github

Committed by

avivajpeyi

Commit Message

Merge branch 'main' of github.com:nz-gravity/LogPSplinePSD

Run Details

179 of 190 branches covered (94.21%)

Branch coverage included in aggregate %.

77 of 94 new or added lines in 8 files covered. (81.91%)

9 existing lines in 3 files now uncovered.

1566 of 1752 relevant lines covered (89.38%)

1.79 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

76.81

/src/log_psplines/plotting/utils.py

import dataclasses

import arviz as az
import jax.numpy as jnp
import numpy as np

from ..datatypes import Periodogram
from ..psplines import LogPSplines

__all__ = ["unpack_data"]


@dataclasses.dataclass
class PlottingData:
    freqs: np.ndarray = None
    pdgrm: np.ndarray = None
    model: np.ndarray = None
    ci: np.ndarray = None

    @property
    def n(self):
        if self.freqs is not None:
            return len(self.freqs)
        elif self.pdgrm is not None:
            return len(self.pdgrm)
        elif self.model is not None:
            return len(self.model)
        else:
            raise ValueError("No data to get length from.")


def unpack_data(
    pdgrm: Periodogram = None,
    spline_model: LogPSplines = None,
    weights=None,
    yscalar=1.0,
    use_uniform_ci=True,
    use_parametric_model=True,
    freqs=None,
):
    plt_dat = PlottingData()
    if pdgrm is not None:
        plt_dat.pdgrm = np.array(pdgrm.power, dtype=np.float64) * yscalar
        plt_dat.freqs = np.array(pdgrm.freqs)

    if spline_model is not None:

        if weights is None:
            # just use the initial weights/0 weights
            ln_spline = spline_model(use_parametric_model=use_parametric_model)

        elif weights.ndim == 1:
            # only one set of weights -- no CI possible
            ln_spline = spline_model(weights, use_parametric_model)

        else:  # weights.ndim == 2
            # multiple sets of weights -- CI possible

            if weights.shape[0] > 500:
                # subsample to speed up
                idx = np.random.choice(
                    weights.shape[0], size=500, replace=False
                )
                weights = weights[idx]

            ln_splines = jnp.array(
                [spline_model(w, use_parametric_model) for w in weights]
            )

            if use_uniform_ci:
                ln_ci = _get_uni_ci(ln_splines)
            else:  # percentile
                ln_ci = jnp.percentile(
                    ln_splines, q=jnp.array([16, 50, 84]), axis=0
                )
            ln_ci = jnp.array(ln_ci)
            plt_dat.ci = np.exp(ln_ci, dtype=np.float64) * yscalar
            ln_spline = ln_ci[1]
        plt_dat.model = np.exp(ln_spline, dtype=np.float64) * yscalar

    if plt_dat.freqs is None and freqs is None:
        plt_dat.freqs = np.linspace(0, 1, plt_dat.n)
    elif freqs is not None:
        plt_dat.freqs = freqs

    return plt_dat


def _get_uni_ci(samples, alpha=0.1):
    """
    Compute a uniform (simultaneous) confidence band for a set of function samples.

    Args:
        samples (jnp.ndarray): Shape (num_samples, num_points) array of function samples.
        alpha (float): Significance level (default 0.1 for 90% CI).

    Returns:
        tuple: (lower_bound, median, upper_bound) arrays.
    """
    num_samples, num_points = samples.shape

    # Compute pointwise median and standard deviation
    median = jnp.median(samples, axis=0)
    std = jnp.std(samples, axis=0)

    # Compute the max deviation over all samples
    deviations = (samples - median[None, :]) / std[
        None, :
    ]  # Normalize deviations
    max_deviation = jnp.max(
        jnp.abs(deviations), axis=1
    )  # Max deviation per sample

    # Compute the scaling factor using the distribution of max deviations
    k_alpha = jnp.percentile(
        max_deviation, 100 * (1 - alpha)
    )  # Critical value

    # Compute uniform confidence bands
    lower_bound = median - k_alpha * std
    upper_bound = median + k_alpha * std

    return lower_bound, median, upper_bound

1	import dataclasses	2✔
2
3	import arviz as az	2✔
4	import jax.numpy as jnp	2✔
5	import numpy as np	2✔
6
7	from ..datatypes import Periodogram	2✔
8	from ..psplines import LogPSplines	2✔
9
10	__all__ = ["unpack_data"]	2✔
11
12
13	@dataclasses.dataclass	2✔
14	class PlottingData:	2✔
15	freqs: np.ndarray = None	2✔
16	pdgrm: np.ndarray = None	2✔
17	model: np.ndarray = None	2✔
18	ci: np.ndarray = None	2✔
19
20	@property	2✔
21	def n(self):	2✔
22	if self.freqs is not None:	×
23	return len(self.freqs)	×
24	elif self.pdgrm is not None:	×
25	return len(self.pdgrm)	×
26	elif self.model is not None:	×
27	return len(self.model)	×
28	else:
29	raise ValueError("No data to get length from.")	×
30
31
32	def unpack_data(	2✔
33	pdgrm: Periodogram = None,
34	spline_model: LogPSplines = None,
35	weights=None,
36	yscalar=1.0,
37	use_uniform_ci=True,
38	use_parametric_model=True,
39	freqs=None,
40	):
41	plt_dat = PlottingData()	2✔
42	if pdgrm is not None:	2✔
43	plt_dat.pdgrm = np.array(pdgrm.power, dtype=np.float64) * yscalar	2✔
44	plt_dat.freqs = np.array(pdgrm.freqs)	2✔
45
46	if spline_model is not None:	2✔
47
48	if weights is None:	2✔
49	# just use the initial weights/0 weights
50	ln_spline = spline_model(use_parametric_model=use_parametric_model)	2✔
51
52	elif weights.ndim == 1:	2✔
53	# only one set of weights -- no CI possible
54	ln_spline = spline_model(weights, use_parametric_model)	×
55
56	else: # weights.ndim == 2
57	# multiple sets of weights -- CI possible
58
59	if weights.shape[0] > 500:	2✔
60	# subsample to speed up
NEW 61	idx = np.random.choice(	×
62	weights.shape[0], size=500, replace=False
63	)
NEW 64	weights = weights[idx]	×
65
66	ln_splines = jnp.array(	2✔
67	[spline_model(w, use_parametric_model) for w in weights]
68	)
69
70	if use_uniform_ci:	2✔
71	ln_ci = _get_uni_ci(ln_splines)	2✔
72	else: # percentile
UNCOV 73	ln_ci = jnp.percentile(	×
74	ln_splines, q=jnp.array([16, 50, 84]), axis=0
75	)
76	ln_ci = jnp.array(ln_ci)	2✔
77	plt_dat.ci = np.exp(ln_ci, dtype=np.float64) * yscalar	2✔
78	ln_spline = ln_ci[1]	2✔
79	plt_dat.model = np.exp(ln_spline, dtype=np.float64) * yscalar	2✔
80
81	if plt_dat.freqs is None and freqs is None:	2✔
UNCOV 82	plt_dat.freqs = np.linspace(0, 1, plt_dat.n)	×
83	elif freqs is not None:	2✔
UNCOV 84	plt_dat.freqs = freqs	×
85
86	return plt_dat	2✔
87
88
89	def _get_uni_ci(samples, alpha=0.1):	2✔
90	"""
91	Compute a uniform (simultaneous) confidence band for a set of function samples.
92
93	Args:
94	samples (jnp.ndarray): Shape (num_samples, num_points) array of function samples.
95	alpha (float): Significance level (default 0.1 for 90% CI).
96
97	Returns:
98	tuple: (lower_bound, median, upper_bound) arrays.
99	"""
100	num_samples, num_points = samples.shape	2✔
101
102	# Compute pointwise median and standard deviation
103	median = jnp.median(samples, axis=0)	2✔
104	std = jnp.std(samples, axis=0)	2✔
105
106	# Compute the max deviation over all samples
107	deviations = (samples - median[None, :]) / std[	2✔
108	None, :
109	] # Normalize deviations
110	max_deviation = jnp.max(	2✔
111	jnp.abs(deviations), axis=1
112	) # Max deviation per sample
113
114	# Compute the scaling factor using the distribution of max deviations
115	k_alpha = jnp.percentile(	2✔
116	max_deviation, 100 * (1 - alpha)
117	) # Critical value
118
119	# Compute uniform confidence bands
120	lower_bound = median - k_alpha * std	2✔
121	upper_bound = median + k_alpha * std	2✔
122
123	return lower_bound, median, upper_bound	2✔

nz-gravity / LogPSplinePSD / 17657445007

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous