pyrolite

Install

pip install pyrolite

Build Status

master	develop

Maintainer: Morgan Williams (morgan.williams at csiro.au)

Usage Examples

Note: Examples for compositional data yet to come.

Elements and Oxides

Index Generators

All Elements up to U

>>> import pyrolite.geochem.common_elements as ce
>>> ce()  # string return
['H', 'He', 'Li', 'Be', ...,  'Th', 'Pa', 'U']
>>> ce(output='formula')  # periodictable.core.Element return
[H, He, Li, Be, ...,  Th, Pa, U]

Oxides for Elements with Positive Charges (up to U)

>>> import pyrolite.geochem.common_oxides as co
>>> co()  # string return
['H2O', 'He2O', 'HeO', 'Li2O', 'Be2O', 'BeO', 'B2O', 'BO', 'B2O3', ...,
'U2O', 'UO', 'U2O3', 'UO2', 'U2O5', 'UO3']
>>> co()  # periodictable.formulas.Formula return
[H, He, Li, Be, ...,  Th, Pa, U]

REE Elements

>>> from pyrolite.geochem import REE
>>> REE()
['La', 'Ce', 'Pr', 'Nd', 'Pm', ..., 'Dy', 'Ho', 'Er', 'Tm', 'Yb', 'Lu']

Data Cleaning

Some simple utilities for cleaning up data tables are included. Assuming you're importing data into pd.DataFrame:

import pandas as pd
df = pd.DataFrame({'label':'basalt', 'ID': 19076,
                   'mgo':20.0, 'SIO2':30.0, 'cs':5.0, 'TiO2':2.0},
                  index=[0])
>>> df.columns
Index(['label', 'ID', 'mgo', 'SIO2', 'cs', 'TiO2'], dtype='object')

from pyrolite.util.text import titlecase
from pyrolite.geochem import tochem

>>> df.columns = [titlecase(h, abbrv=['ID']) for h in df.columns]
Index(['Label', 'ID', 'Mgo', 'Sio2', 'Cs', 'Tio2'], dtype='object')
>>> df.columns = tochem(df.columns)
Index(['Label', 'ID', 'MgO', 'SiO2', 'Cs', 'TiO2'], dtype='object')

Normalisation

A selection of reference compositions are included:

>>> from pyrolite.normalisation import ReferenceCompositions
>>> refcomp = ReferenceCompositions()
{
'Chondrite_PON': Model of Chondrite (Palme2014),
'D-DMM_WH': Model of DepletedDepletedMORBMantle (Workman2005),
'DMM_WH': Model of DepletedMORBMantle (Workman2005),
'DM_SS': Model of DepletedMantle (Salters2004),
'E-DMM_WH': Model of EnrichedDepletedMORBMantle (Workman2005),
'PM_PON': Model of PrimitiveMantle (Palme2014)
}

>>> CH = refcomp['Chondrite_PON']
>>> PM = refcomp['PM_PON']
>>> CH[REE()]
      value  unc_2sigma units
var                           
La    0.2414    0.014484   ppm
Ce    0.6194    0.037164   ppm
...
Tm   0.02609    0.001565   ppm
Yb    0.1687    0.010122   ppm
Lu   0.02503    0.001502   ppm

The normalize method can be used to normalise dataframes to a given reference (e.g. for spiderplots):

>>> from pyrolite.plot import spiderplot
>>> refcomp = ReferenceCompositions()
>>> CH = refcomp['Chondrite_PON']
>>> DMM = refcomp['DMM_WH']
>>>
>>> df = DMM.data.loc[REE(), ['value']]
>>> spiderplot(CH.normalize(df), label=f'{DMM.Reference}')

More reference compositions will soon be included (e.g. Sun and McDonough, 1989).

Data Density Plots

Log-spaced data density plots can be useful to visualise geochemical data density:

>>> from pyrolite.plot import densityplot
>>> # with a dataframe <df> containing columns Nb/Yb and Th/Yb
>>> densityplot(df, components=['Nb/Yb', 'Th/Yb'], bins=100, logspace=True)

Below is an example of ocean island basalt data (GEOROC compilation), plotted in a 'Pearce' discrimination diagram:

More on these discrimination diagrams: Pearce, J.A., 2008. Geochemical fingerprinting of oceanic basalts with applications to ophiolite classification and the search for Archean oceanic crust. Lithos 100, 14–48.

Dimensional Reduction using Orthagonal Polynomials ('Lambdas')

Derivation of weight values for deconstructing a smooth function into orthagonal polynomial components (e.g. for the REE):

>>> from pyrolite.geochem import lambda_lnREE
>>> refc = 'Chondrite_PON'
>>> # with a dataframe <df> containing REE data in columns La, ..., Lu
>>> lambdas = lambda_lnREE(df, exclude=['Pm'], norm_to=refc)


For more on using orthagonal polynomials to describe geochemical pattern data, see: O’Neill, H.S.C., 2016. The Smoothness and Shapes of Chondrite-normalized Rare Earth Element Patterns in Basalts. J Petrology 57, 1463–1508.

Classification

Some simple discrimination methods are implemented, including the Total Alkali-Silica (TAS) classification:

>>> from pyrolite.classification import Geochemistry
>>>
>>> cm = Geochemistry.TAS()
>>> df.TotalAlkali = df.Na2O + df.K2O
>>> df['TAS'] = cm.classify(df)

This classifier can be quickly added to a bivariate plot, assuming you have data in a pandas DataFrame:

>>> import pandas as pd
>>> import matplotlib.pyplot as plt
>>>
>>> df['TotalAlkali'] = df['Na2O'] + df['K2O']
>>>
>>> fig, ax = plt.subplots(1, figsize=(6, 4))
>>> cm.add_to_axes(ax, facecolor='0.9', edgecolor='k',
>>>                linewidth=0.5, zorder=-1)
>>> classnames = cm.clsf.fclasses + ['none']
>>> df['TAScolors'] = df['TAS'].map(lambda x: classnames.index(x))
>>> ax.scatter(df.SiO2, df.TotalAlkali, c=df.TAScolors,
>>>            alpha=0.5, marker='D', s=8, cmap='tab20c')