1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
|
# -*- coding: utf-8 -*-
"""Compare module."""
from __future__ import division
from .errors import pycmCompareError
from .output import *
from .utils import *
from .params import *
from .cm import ConfusionMatrix
import os
import numpy
from warnings import warn
class Compare():
"""
Compare class.
>>> cm1 = ConfusionMatrix(matrix={0:{0:2,1:50,2:6}, 1:{0:5,1:50,2:3}, 2:{0:1,1:7,2:50}})
>>> cm2 = ConfusionMatrix(matrix={0:{0:50,1:2,2:6}, 1:{0:50,1:5,2:3}, 2:{0:1,1:55,2:2}})
>>> cp = Compare({"cm1": cm1, "cm2": cm2})
>>> print(cp)
Best : cm1
<BLANKLINE>
Rank Name Class-Score Overall-Score
1 cm1 0.50278 0.58095
2 cm2 0.33611 0.52857
<BLANKLINE>
>>> cp.best
pycm.ConfusionMatrix(classes: [0, 1, 2])
>>> cp.sorted
['cm1', 'cm2']
>>> cp.best_name
'cm1'
"""
def __init__(
self,
cm_dict,
by_class=False,
class_weight=None,
class_benchmark_weight=None,
overall_benchmark_weight=None,
digit=5):
"""
Init method.
:param cm_dict: dictionary of confusion matrices
:type cm_dict: dict
:param by_class: compare by class flag
:type by_class: bool
:param class_weight: class weights
:type class_weight: dict
:param class_benchmark_weight: class benchmark weights
:type class_benchmark_weight: dict
:param overall_benchmark_weight: overall benchmark weights
:type overall_benchmark_weight: dict
:param digit: scale (number of fraction digits)(default value: 5)
:type digit: int
"""
self.scores = None
self.sorted = None
self.classes = None
__compare_assign_handler__(
self,
cm_dict,
class_weight,
class_benchmark_weight,
overall_benchmark_weight,
digit)
__compare_class_handler__(self, cm_dict)
__compare_overall_handler__(self, cm_dict)
__compare_rounder__(self, cm_dict)
scores_list = list(self.scores.values())
(max_overall_name, max_class_name) = __compare_sort_handler__(self)
if scores_list.count(self.scores[max_class_name]) == 1:
if by_class:
self.best = cm_dict[max_class_name]
self.best_name = max_class_name
else:
if max_overall_name == max_class_name:
self.best = cm_dict[max_class_name]
self.best_name = max_overall_name
else:
warn(COMPARE_RESULT_WARNING, RuntimeWarning)
else:
warn(COMPARE_RESULT_WARNING, RuntimeWarning)
def print_report(self):
"""
Print Compare report.
:return: None
"""
report = compare_report_print(
self.sorted, self.scores, self.best_name)
print(report)
def save_report(
self,
name,
address=True):
"""
Save Compare report in .comp (flat file format).
:param name: filename
:type name: str
:param address: flag for address return
:type address: bool
:return: saving address as dict {"Status": bool, "Message": str}
"""
try:
message = None
file = open(name + ".comp", "w")
report = compare_report_print(
self.sorted, self.scores, self.best_name)
file.write(report)
file.close()
if address:
message = os.path.join(
os.getcwd(), name + ".comp") # pragma: no cover
return {"Status": True, "Message": message}
except Exception as e:
return {"Status": False, "Message": str(e)}
def __repr__(self):
"""
Compare object representation method.
:return: representation as str
"""
return "pycm.Compare(classes: " + str(self.classes) + ")"
def __str__(self):
"""
Compare object string representation method.
:return: representation as str
"""
report = compare_report_print(
self.sorted, self.scores, self.best_name)
return report
def __compare_class_handler__(compare, cm_dict):
"""
Handle class score of Compare class.
:param compare: Compare
:type compare: pycm.Compare object
:param cm_dict: dictionary of confusion matrices
:type cm_dict: dict
:return: None
"""
class_weight_sum = sum(compare.class_weight.values())
class_benchmark_weight_sum = sum(compare.class_benchmark_weight.values())
for c in compare.classes:
for item in CLASS_BENCHMARK_SCORE_DICT:
max_item_score = len(CLASS_BENCHMARK_SCORE_DICT[item]) - 1
all_class_score = [CLASS_BENCHMARK_SCORE_DICT[item][
cm.class_stat[item][c]] for cm in cm_dict.values()]
if all([isinstance(x, int) for x in all_class_score]):
for cm_name in cm_dict:
score = (compare.class_weight[c] / class_weight_sum) * (
CLASS_BENCHMARK_SCORE_DICT[item][cm_dict[cm_name].class_stat[item][c]] / max_item_score)
score = score * \
(compare.class_benchmark_weight[item] / class_benchmark_weight_sum)
compare.scores[cm_name]["class"] += score
def __compare_overall_handler__(compare, cm_dict):
"""
Handle overall score of Compare class.
:param compare: Compare
:type compare: pycm.Compare object
:param cm_dict: dictionary of confusion matrices
:type cm_dict: dict
:return: None
"""
overall_benchmark_weight_sum = sum(
compare.overall_benchmark_weight.values())
for item in OVERALL_BENCHMARK_SCORE_DICT:
max_item_score = len(OVERALL_BENCHMARK_SCORE_DICT[item]) - 1
all_overall_score = [OVERALL_BENCHMARK_SCORE_DICT[item][
cm.overall_stat[OVERALL_BENCHMARK_MAP[item]]] for cm in cm_dict.values()]
if all([isinstance(x, int) for x in all_overall_score]):
for cm_name in cm_dict:
score = OVERALL_BENCHMARK_SCORE_DICT[item][cm_dict[cm_name]
.overall_stat[OVERALL_BENCHMARK_MAP[item]]] / max_item_score
score = score * \
(compare.overall_benchmark_weight[item] / overall_benchmark_weight_sum)
compare.scores[cm_name]["overall"] += score
def __compare_rounder__(compare, cm_dict):
"""
Round Compare.scores .
:param compare: Compare
:type compare: pycm.Compare object
:param cm_dict: dictionary of confusion matrices
:type cm_dict: dict
:return: None
"""
for cm_name in cm_dict:
compare.scores[cm_name]["overall"] = numpy.around(
compare.scores[cm_name]["overall"], compare.digit)
compare.scores[cm_name]["class"] = numpy.around(
compare.scores[cm_name]["class"], compare.digit)
def __compare_sort_handler__(compare):
"""
Handle sorting of scores.
:param compare: Compare
:type compare: pycm.Compare object
:return: (max_overall_name,max_class_name) as tuple
"""
sorted_by_class = sorted(
compare.scores,
key=lambda x: (
compare.scores[x]['class'],
compare.scores[x]['overall']))
sorted_by_overall = sorted(
compare.scores,
key=lambda x: (
compare.scores[x]['overall'],
compare.scores[x]['class']))
sorted_by_class.reverse()
sorted_by_overall.reverse()
compare.sorted = sorted_by_class
max_overall_name = sorted_by_overall[0]
max_class_name = sorted_by_class[0]
return (max_overall_name, max_class_name)
def __compare_weight_handler__(compare, weight, weight_type):
"""
Handle different weights validation.
:param compare: Compare
:type compare: pycm.Compare object
:param weight: input weight
:type weight: dict
:param weight_type: input weight type
:type weight_type: str
:return: None
"""
valid_dict = {
"class_weight": compare.classes,
"class_benchmark_weight": CLASS_BENCHMARK_SCORE_DICT.keys(),
"overall_benchmark_weight": OVERALL_BENCHMARK_SCORE_DICT.keys()}
error_dict = {
"class_weight": COMPARE_CLASS_WEIGHT_ERROR,
"class_benchmark_weight": COMPARE_CLASS_BENCHMARK_WEIGHT_ERROR,
"overall_benchmark_weight": COMPARE_OVERALL_BENCHMARK_WEIGHT_ERROR}
warning_dict = {
"class_weight": COMPARE_CLASS_WEIGHT_WARNING,
"class_benchmark_weight": COMPARE_CLASS_BENCHMARK_WEIGHT_WARNING,
"overall_benchmark_weight": COMPARE_OVERALL_BENCHMARK_WEIGHT_WARNING}
if weight is None:
return None
if not isinstance(weight, dict):
raise pycmCompareError(error_dict[weight_type])
if set(weight) == set(valid_dict[weight_type]):
if all([isfloat(x) for x in weight.values()]
) and sum(weight.values()) != 0:
setattr(compare, weight_type, weight)
else:
warn(warning_dict[weight_type], RuntimeWarning)
else:
raise pycmCompareError(error_dict[weight_type])
def __compare_assign_handler__(
compare,
cm_dict,
class_weight,
class_benchmark_weight,
overall_benchmark_weight,
digit):
"""
Assign basic parameters to Compare.
:param compare: Compare
:type compare: pycm.Compare object
:param cm_dict: dictionary of confusion matrices
:type cm_dict: dict
:param class_weight: class weights
:type class_weight: dict
:param class_benchmark_weight: class benchmark weights
:type class_benchmark_weight: dict
:param overall_benchmark_weight: overall benchmark weights
:type overall_benchmark_weight: dict
:param digit: scale (number of fraction digits)(default value: 5)
:type digit: int
:return: None
"""
if not isinstance(cm_dict, dict):
raise pycmCompareError(COMPARE_FORMAT_ERROR)
if not all(isinstance(item, ConfusionMatrix)
for item in cm_dict.values()):
raise pycmCompareError(COMPARE_TYPE_ERROR)
if any(item.metrics_off for item in cm_dict.values()):
raise pycmCompareError(COMPARE_METRICS_OFF_ERROR)
if not list_check_equal([getattr(item, "POP")
for item in cm_dict.values()]):
raise pycmCompareError(COMPARE_DOMAIN_ERROR)
if len(cm_dict) < 2:
raise pycmCompareError(COMPARE_NUMBER_ERROR)
compare.classes = list(cm_dict.values())[0].classes
compare.class_weight = {k: 1 for k in compare.classes}
compare.class_benchmark_weight = {k: 1 for k in CLASS_BENCHMARK_LIST}
compare.overall_benchmark_weight = {
k: 0 if k in KAPPA_BENCHMARK_LIST[1:] else 1 for k in OVERALL_BENCHMARK_LIST}
compare.digit = digit
compare.best = None
compare.best_name = None
compare.sorted = None
compare.scores = {k: {"overall": 0, "class": 0}.copy()
for k in cm_dict}
__compare_weight_handler__(compare, class_weight, "class_weight")
__compare_weight_handler__(
compare,
class_benchmark_weight,
"class_benchmark_weight")
__compare_weight_handler__(
compare,
overall_benchmark_weight,
"overall_benchmark_weight")
|