def setUp(self):
self.G = augmented_covariance_matrix(np.array([1.1, 0.9]),
random_cov_matrix(2))
self.G3 = augmented_covariance_matrix(np.array([1.1, 0.9, 1.05]),
random_cov_matrix(3))
def setUp(self):
self.mean = np.random.randn(5)
self.cov = random_cov_matrix(5)
.. note:: This example and features in this module are currently incomplete and a work in progress.
"""
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from pyrolite.comp.impute import EMCOMP
from pyrolite.util.synthetic import random_composition, random_cov_matrix
from pyrolite.plot import pyroplot
np.random.seed(41)
########################################################################################
sample_data = random_composition(
size=1000,
D=4,
cov=random_cov_matrix(3, sigmas=[0.1, 0.3, 0.15]),
propnan=0.1,
missing="MNAR",
missingcols=3,
)
########################################################################################
imputed_data, p0, niter = EMCOMP(sample_data,
threshold=np.nanpercentile(sample_data,
10,
axis=0),
tol=0.01)
imputed_data = pd.DataFrame(imputed_data, columns=["A", "B", "C", "D"])
########################################################################################
fig, ax = plt.subplots(1, 3, sharex=True, sharey=True, figsize=(12, 5))
ax[0].set_title("Original Data")
the compositional space. The commonly used log-transformations include the
Additive Log-Ratio (:func:`~pyrolite.comp.pyrocomp.ALR`), Centred Log-Ratio
(:func:`~pyrolite.comp.pyrocomp.CLR`), and Isometric Log-Ratio
(:func:`~pyrolite.comp.pyrocomp.ILR`) [#ref_1]_ [#ref_2]_.
This example will show you how to access and use some of these functions in pyrolite.
"""
########################################################################################
# First let's create some example data:
#
from pyrolite.util.synthetic import normal_frame, random_cov_matrix
df = normal_frame(
size=100,
cov=random_cov_matrix(sigmas=[0.1, 0.05, 0.3, 0.6], dim=4, seed=32),
seed=32,
)
df.describe()
########################################################################################
# Let's have a look at some of the log-transforms, which can be accessed directly from
# your dataframes (via :class:`pyrolite.comp.pyrocomp`), after you've imported
# :mod:`pyrolite.comp`. Note that the transformations will return *new* dataframes,
# rather than modify their inputs. For example:
#
import pyrolite.comp
lr_df = df.pyrocomp.CLR() # using a centred log-ratio transformation
########################################################################################
# The transformations are implemented such that the column names generally make it
# evident which transformations have been applied (here using default simple labelling;
cs = cs.reshape(cs.shape[0], *c1.shape)
# generate samples from each
samples = np.vstack(
[
np.random.multivariate_normal(m.flatten(), cs[ix], size=size // resolution)
for ix, m in enumerate(ms)
]
)
# combine together.
return inverse_ilr(samples)
########################################################################################
# First we create an array of compositions which represent a trend.
#
m1, m2 = np.array([[0.3, 0.1, 2.1]]), np.array([[0.5, 2.5, 0.05]])
c1, c2 = (
random_cov_matrix(2, sigmas=[0.15, 0.05]),
random_cov_matrix(2, sigmas=[0.05, 0.2]),
)
trend = pd.DataFrame(
random_compositional_trend(m1, m2, c1, c2, resolution=100, size=5000),
columns=["A", "B", "C"],
)
########################################################################################
# We can visualise this compositional trend with a density plot.
#
ax = trend.pyroplot.density(mode="density", bins=100)
plt.show()
########################################################################################
# First we can see where the geometric mean would fall:
#
represents the best of both worlds. A version inspired by similar existing
visualisations is implemented with :func:`~pyrolite.plot.pyroplot.heatscatter`.
"""
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from pyrolite.plot import pyroplot
np.random.seed(12)
########################################################################################
# First we'll create some example data
from pyrolite.util.synthetic import normal_frame, random_cov_matrix
df = normal_frame(
size=1000,
cov=random_cov_matrix(sigmas=np.random.rand(4) * 2, dim=4, seed=12),
seed=12,
)
########################################################################################
# We can compare a minimal :func:`~pyrolite.plot.pyroplot.heatscatter` plot to other
# visualisations for the same data:
#
from pyrolite.util.plot.axes import share_axes
fig, ax = plt.subplots(3, 4, figsize=(12, 9))
ax = ax.flat
share_axes(ax[:4], which="xy")
share_axes(ax[4:8], which="xy")
share_axes(ax[8:], which="xy")
Break down a list of phases into oxide components.
"""
atoms = set()
# get atoms which are not O
for p in phases:
atoms |= set(get_cations(p))
print(atoms)
oxides = [simple_oxides(a)[1] for a in atoms]
return oxides
from pyrolite.util.synthetic import test_df, random_cov_matrix
X = test_df(
cols=["cpx", "olivine", "plag"],
cov=random_cov_matrix(dim=2, sigmas=np.array([0.5, 0.7])),
)
component_matrix(components=X.columns,
phases=["CaAl2Si2O8", "Mg2Si2O4", "CaMgSi2O6"])
mixed = X @ components.pyrochem.to_weight().pyrocomp.renormalise(scale=1)
mixed
def unmix_compositions(X, components, bdl=10 * -5):
x = X.copy()
c = components.copy()
c = c.pyrochem.to_weight().pyrocomp.renormalise(scale=1)
# c[c == 0] = bdl; clr
