produce_figures.py

"""
(Re)produce the plots from the figures!
"""
from pyolin import FIGDIR
import pyolin.figures as figures

import pandas
import os
import sys
from shutil import which

from pyolin.dataframe import GateData
from pyolin.analysis import filter_valid

fig_dir = 'notebooks/results/figure_data'


def plot_system_call(filepath, *gnuplotargs):
    minus_extension = filepath.rpartition('.')[0]
    directory = filepath.rpartition('/')[0]
    os.system(f"gnuplot -c {filepath} {' '.join(gnuplotargs)}")
    os.system(f"latex -output-directory={directory} {minus_extension}.tex")
    os.system(f"dvips -E -o {minus_extension}.ps {minus_extension}.dvi")

    
if __name__ == '__main__':
    # check the required utilities are available
    gnuplot = which("gnuplot")
    latex = which("latex")
    dvips = which("dvips")
    if latex is None or gnuplot is None or dvips is None:
        print("ERR: Plotting requires gnuplot, latex and dvips are installed.")
        sys.exit(1)

    convert = which("convert")
    if convert is None:
        print("WARN: convert not found, skipping conversion to png.")

    # reading and renaming some columns
    raw = pandas.read_csv('cyto2func/standardised_cheeky.csv')
    raw = raw.rename(columns={'rrpu': 'decomp_flor'})
    raw = raw.rename(columns={'newstandard': 'rrpu'})

    data = GateData(raw)
    inverter_names = data.gates[3:]

    # Figure 2
    # this generates the comparison plot for ALL pairs of gates. The
    # produced files weigh quite a lot. But you could also get rid of
    # the for loops and do:
    # figures.figure2(data[::<GATE_NAME>], data[::<OTHER_NAME>], fig_dir + '/Figure2')
    # obviously replacing the names between <>
    for a in inverter_names:
        for b in inverter_names:
            figures.figure2(data[::a], data[::b])

    os.system(f"gnuplot '{FIGDIR / 'Figure2'}/figure2_with_Psra_r1_vs_Phif_p2.gnuplot'")

    # Figure 1
    os.system(f"gnuplot {FIGDIR/'Figure1/figure.gp'}")

    # Figure 3
    for a in inverter_names:
        figures.similaritymap(data[::a], FIGDIR / 'Figure3/B')
    os.system(f"gnuplot {FIGDIR / 'Figure3/B/similarity_matrix_Phif_p1.gnuplot'}")
    
    gates = filter_valid(data['DH5alpha':'pAN':])
    figures.compatmap(gates, 'DH5alpha_+_pAN_Context', FIGDIR / 'Figure3/C')
    gates = filter_valid(data['DH5alpha'::])
    figures.compatmap(gates, 'DH5alpha_Strain', FIGDIR / 'Figure3/C')
    gates = filter_valid(data[::])
    figures.compatmap(gates, 'All_Strains', FIGDIR / 'Figure3/C')
    os.system(f"gnuplot {FIGDIR/'Figure3/C/compat_map_DH5alpha_+_pAN_Context.gnuplot'}")
    os.system(f"gnuplot {FIGDIR/'Figure3/C/compat_map_DH5alpha_Strain.gnuplot'}")
    os.system(f"gnuplot {FIGDIR/'Figure3/C/compat_map_All_Strains.gnuplot'}")

    # Figure 4
    os.system(f"gnuplot '{FIGDIR}/Figure4/panelI.gp'")
    os.system(f"gnuplot '{FIGDIR}/Figure4/panelII.gp'")
    os.system(f"gnuplot '{FIGDIR}/Figure4/panelIII.gp'")

    # Supplementary
    # S2
    for strain in data.strains:
        figures.compatmap(data[strain::], f"{strain}_Strain", FIGDIR/'SuppFigure2')
        os.system(f"gnuplot {FIGDIR}/SuppFigure2/compat_map_{strain}_Strain.gnuplot")

    # S3
    for s, b in data.contexts:
        figures.compatmap(data[s:b:], f"{b}_in_{s}", FIGDIR/'SuppFigure3')
        os.system(f"gnuplot {FIGDIR}/SuppFigure3/compat_map_{b}_in_{s}.gnuplot")

    # S4
    for inverter in inverter_names:
        figures.similaritymap(data[::inverter], FIGDIR/'SuppFigure4')
        os.system(f"gnuplot {FIGDIR}/SuppFigure4/similarity_matrix_{inverter}.gnuplot")