Sleds/scorewalker-utils/RunTest/AccuracyGraphSetup.py

import json

FILES = 'files'
COLORS = 'colors'
LABELS = 'labels'

TITLE = 'Title'
X_LBL = 'xLabel'
Y_LBL = 'yLabel'

NUM_VAL = 'number'
STR_VAL = 'string'

package_compare_colors = {'Correct & High Confidence': 'g', 'Incorrect & Low Confidence': 'y',
                          'Incorrect & High Confidence': 'r', 'Correct & Low Confidence': 'b',
                          'Correct Pagination': 'm', 'Classification Score': 'teal'}

doctype_compare_colors = {'Correct & High Confidence': 'g', 'Incorrect & Low Confidence': 'y',
                          'Incorrect & High Confidence': 'r', 'Correct & Low Confidence': 'b'}


def load_files(files):
    result = {'Incorrect & High Confidence': 0.0,
              'Correct & Low Confidence': 0.0,
              'Correct & High Confidence': 0.0,
              'Incorrect & Low Confidence': 0.0,
              'Correct Pagination': 0.0}

    for f in files:
        cfg_data = load_data_from_file(f)
        for key, _ in result.items():
            result[key] += cfg_data[key]

    for key, _ in result.items():
        result[key] /= len(files)

    return result


def load_data_from_file(file_path):
    result = {}
    with open(file_path) as reader:
        lines = reader.readlines()
        # Get the data from the lines.
        for line in lines:
            key, val = parse_line(line)
            result[key] = val

    return result


def parse_line(line_txt):
    if line_txt.endswith('\n'):
        line_txt = line_txt[:-1]

    tag, value = line_txt.split('=', 1)
    value, val_type = value.split(':', 1)
    return tag, cast_to_type(val_type, value)


def cast_to_type(t_val, val):
    if t_val == NUM_VAL:
        return float(val)

    elif t_val == STR_VAL:
        return str(val)

    else:
        return val


def make_avg_cfg(files, title, config_file_loc):
    data_file = config_file_loc.replace('graph-config.json', 'graph-data.txt')
    make_global_graph_data_file(files, data_file)
    # to_json = {FILES: make_config(files, tags), COLORS: package_compare_colors, LABELS: get_package_labels(title)}
    to_json = {FILES: {'Averages': data_file}, COLORS: package_compare_colors, LABELS: get_package_labels(title)}

    with open(config_file_loc, 'w+') as writer:
        json.dump(to_json, writer, indent=3)


def make_package_graph_config(files, tags, title, config_file_loc):

    to_json = {FILES: make_config(files, tags), COLORS: package_compare_colors, LABELS: get_package_labels(title)}

    with open(config_file_loc, 'w+') as writer:
        json.dump(to_json, writer, indent=3)


def make_global_graph_data_file(files, out_path):
    data = load_files(files)

    with open(out_path, 'w+', newline='') as writer:
        for key, value in data.items():
            writer.write('%s=%s:%s\n' % (key, value, NUM_VAL))


def change_title(old_file, new_file, new_title):
    # Load and change the title.
    new_config = load_config(old_file)
    new_config[LABELS][TITLE] = new_title
    # Save the file.
    with open(new_file, 'w+') as writer:
        json.dump(new_config, writer)


def load_config(config_file):
    with open(config_file) as reader:
        result = json.load(reader)
    return result


def load_files_from_config(config_file):
    with open(config_file) as cfg_reader:
        config = json.load(cfg_reader)

    return get_files(config[FILES])


def get_files(config_files):
    result = []

    for tag, file in config_files.items():
        result.append(file)

    return result


def make_config(files, tags):
    result = {}
    for i in range(len(files)):
        result[tags[i]] = files[i]
    return result


def get_package_labels(title):
    return {X_LBL: 'Document.', Y_LBL: 'Pages Classified (%).', TITLE: title}


def get_doctype_labels(title):
    return {X_LBL: 'Doctype', Y_LBL: 'Pages Classified (%)', TITLE: title}