P-unit_model/cell_overview.py

from CellData import icelldata_of_dir, CellData
from Baseline import BaselineCellData
from FiCurve import FICurveCellData
from DataParserFactory import DatParser
import os
import numpy as np
import matplotlib.pyplot as plt


def main():

    # plot_visualizations("cells/")
    # full_overview("cells/master_table.csv", "cells/")

    # recalculate_saved_preanalysis("data/final/")
    metadata_analysis("data/final/")
    pass


def metadata_analysis(data_folder, filter_double_fish=False):
    gender = {}
    species = {}
    sampling_interval = {}
    eod_freqs = []
    sizes = []
    dates = {}
    fi_curve_stimulus = []
    fi_curve_contrasts = []
    fi_curve_c_trials = []
    for cell in os.listdir(data_folder):
        if filter_double_fish and cell[:10] in dates.keys():
            continue
        dates[cell[:10]] = 1
        cell_path = data_folder + cell
        parser = DatParser(cell_path)
        cell_data = CellData(cell_path)

        species = count_with_dict(species, parser.get_species())
        gender = count_with_dict(gender, parser.get_gender())
        sampling_interval = count_with_dict(sampling_interval, parser.get_sampling_interval())
        sizes.append(float(parser.get_fish_size()))
        # eod_freqs.append(cell_data.get_eod_frequency())

        fi_curve_stimulus.append(cell_data.get_recording_times())
        contrasts_with_trials = cell_data.get_fi_curve_contrasts_with_trial_number()
        fi_curve_contrasts.append([c[0] for c in contrasts_with_trials if c[1] >= 3])
        fi_curve_c_trials.append([c[1] for c in contrasts_with_trials if c[1] >= 3])

    for k in sampling_interval.keys():
        print("sampling:", np.rint(1.0/float(k)), "count:", sampling_interval[k])
    print("# of Fish (by dates):", len(dates.keys()))
    print("Fish genders:", gender)
    # print("EOD-freq: min {:.2f}, mean {:.2f}, max {:.2f}, std {:.2f}".format(min(eod_freqs), np.mean(eod_freqs), max(eod_freqs), np.std(eod_freqs)))
    print("Sizes: min {:.2f}, mean {:.2f}, max {:.2f}, std {:.2f}".format(min(sizes), np.mean(sizes), max(sizes), np.std(sizes)))

    print("\n\nFi-Curve Stimulus:")
    print("Delay:", np.unique([r[0] for r in fi_curve_stimulus]))
    print("starts:", np.unique([r[1] for r in fi_curve_stimulus]))
    # print("ends:", np.unique([r[0] for r in fi_curve_stimulus]))
    print("duration:", np.unique([r[2] for r in fi_curve_stimulus], return_counts=True))
    print("after:", np.unique([r[3] for r in fi_curve_stimulus]))

    # print("Contrasts:")
    # for c in fi_curve_contrasts:
    #     print("min: {:.1f}, max: {:.1f}, count: {},".format(c[0], c[-1], len(c)))
    # bins = np.arange(-1, 1.01, 0.05)
    # all_contrasts = []
    # for c in fi_curve_contrasts:
    #     all_contrasts.extend(c)
    # plt.hist([c[0] for c in fi_curve_contrasts], bins=bins, label="mins", color="blue", alpha=0.5)
    # plt.hist([c[-1] for c in fi_curve_contrasts], bins=bins, label="maxs", color="red", alpha=0.5)
    # plt.hist(all_contrasts, bins=bins, label="all", color="black", alpha=0.5)
    # plt.show()

    print("done")


def count_with_dict(dictionary, key):
    if key not in dictionary:
        dictionary[key] = 0

    dictionary[key] += 1

    return dictionary



def recalculate_saved_preanalysis(data_folder):
    for cell_data in icelldata_of_dir(data_folder, test_for_v1_trace=True):
        print(cell_data.get_data_path())
        baseline = BaselineCellData(cell_data)
        baseline.save_values(cell_data.get_data_path())

        contrasts = cell_data.get_fi_contrasts()
        fi_curve = FICurveCellData(cell_data, contrasts)

        if fi_curve.get_f_inf_slope() < 0:
            contrasts = np.array(cell_data.get_fi_contrasts()) * -1
        print(contrasts)
        fi_curve = FICurveCellData(cell_data, contrasts, save_dir=cell_data.get_data_path(), recalculate=True)
        # fi_curve.plot_fi_curve()


def move_rejected_cell_data():
    count = 0
    jump_to = 0
    negative_contrast_rel = 0
    cell_list = []
    for d in icelldata_of_dir("invivo_data/"):
        count += 1
        if count < jump_to:
            continue
        print(d.get_data_path())
        base = BaselineCellData(d)
        base.load_values(d.get_data_path())
        ficurve = FICurveCellData(d, d.get_fi_contrasts(), d.get_data_path())

        if ficurve.get_f_inf_slope() < 0:
            negative_contrast_rel += 1

            print("negative f_inf slope")
            cell_list.append(os.path.abspath(d.get_data_path()))

    for c in cell_list:
        if os.path.exists(c):
            print("Source: ", c)
            destination = os.path.abspath("rejected_cells/negative_slope_f_inf/" + os.path.basename(c))
            print("destination: ", destination)
            print()
            os.rename(c, destination)
    print("Number: " + str(negative_contrast_rel))


def plot_visualizations(folder_path):

    for cell_data in icelldata_of_dir("invivo_data/"):

        name = os.path.split(cell_data.get_data_path())[-1]
        print(name)
        save_path = folder_path + name + "/"
        if not os.path.exists(save_path):
            os.mkdir(save_path)

        baseline = BaselineCellData(cell_data)
        baseline.plot_baseline(save_path)
        baseline.plot_serial_correlation(10, save_path)
        baseline.plot_polar_vector_strength(save_path)
        baseline.plot_interspike_interval_histogram(save_path)

        ficurve = FICurveCellData(cell_data, cell_data.get_fi_contrasts())
        ficurve.plot_fi_curve(save_path)


def full_overview(save_path_table, folder_path):
    with open(save_path_table, "w") as table:
        table.write("Name, Path, Baseline Frequency Hz,Vector Strength, serial correlation lag=1,"
                    " serial correlation lag=2, burstiness, coefficient of variation,"
                    " fi-curve inf slope, fi-curve zero slope at straight, contrast at fi-curve zero straight\n")

        # add contrasts, f-inf values, f_zero_values
        count = 0
        start = 0
        for cell_data in icelldata_of_dir("invivo_data/"):
            count += 1
            if count < start:
                continue
            save_dir = cell_data.get_data_path()
            name = os.path.split(cell_data.get_data_path())[-1]
            line = name + ","
            line += cell_data.get_data_path() + ","

            baseline = BaselineCellData(cell_data)
            if not baseline.load_values(save_dir):
                baseline.save_values(save_dir)

            line += "{:.1f},".format(baseline.get_baseline_frequency())
            line += "{:.2f},".format(baseline.get_vector_strength())
            sc = baseline.get_serial_correlation(2)
            line += "{:.2f},".format(sc[0])
            line += "{:.2f},".format(sc[1])
            line += "{:.2f},".format(baseline.get_burstiness())
            line += "{:.2f},".format(baseline.get_coefficient_of_variation())

            ficurve = FICurveCellData(cell_data, cell_data.get_fi_contrasts(), save_dir)

            line += "{:.2f},".format(ficurve.get_f_inf_slope())
            line += "{:.2f}\n".format(ficurve.get_f_zero_fit_slope_at_straight())
            line += "{:.2f}\n".format(ficurve.f_zero_fit[3])
            table.write(line)

            name = os.path.split(cell_data.get_data_path())[-1]
            print(name)
            save_path = folder_path + name + "/"
            if not os.path.exists(save_path):
                os.mkdir(save_path)

            baseline.plot_baseline(save_path)
            baseline.plot_serial_correlation(10, save_path)
            baseline.plot_polar_vector_strength(save_path)
            baseline.plot_interspike_interval_histogram(save_path)

            ficurve.plot_fi_curve(save_path)


if __name__ == '__main__':
    main()