1
|
|
|
""" |
2
|
|
|
KnowYourData |
3
|
|
|
============ |
4
|
|
|
|
5
|
|
|
A rapid and lightweight module to describe the statistics and structure of |
6
|
|
|
data arrays for interactive use. |
7
|
|
|
|
8
|
|
|
The most simple use case to display data is if you have a numpy array 'x': |
9
|
|
|
|
10
|
|
|
>>> from knowyourdata import kyd |
11
|
|
|
>>> kyd(x) |
12
|
|
|
|
13
|
|
|
""" |
14
|
|
|
|
15
|
|
|
import sys |
16
|
|
|
import numpy as np |
17
|
|
|
from IPython.display import display |
18
|
|
|
|
19
|
|
|
# Getting HTML Template |
20
|
|
|
from . import kyd_html_display_template |
21
|
|
|
kyd_htmltemplate = kyd_html_display_template.kyd_htmltemplate |
|
|
|
|
22
|
|
|
|
23
|
|
|
|
24
|
|
|
class KYD_datasummary(object): |
|
|
|
|
25
|
|
|
"""A class to store and display the summary information""" |
26
|
|
|
|
27
|
|
|
text_repr = "" |
28
|
|
|
html_repr = "" |
29
|
|
|
|
30
|
|
|
# Display Settings |
31
|
|
|
col_width = 10 |
32
|
|
|
precision = 4 |
33
|
|
|
|
34
|
|
|
def __repr__(self): |
35
|
|
|
""" |
36
|
|
|
The Plain String Representation of the Data Summary |
37
|
|
|
""" |
38
|
|
|
return self.text_repr |
39
|
|
|
|
40
|
|
|
def _repr_html_(self): |
41
|
|
|
""" |
42
|
|
|
The HTML Representation of the Data Summary |
43
|
|
|
""" |
44
|
|
|
return self.html_repr |
45
|
|
|
|
46
|
|
|
def make_html_repr(self): |
47
|
|
|
"""Make HTML Representation of Data Summary""" |
48
|
|
|
self.html_repr = kyd_htmltemplate.format(kyd_class=self.kyd_class) |
49
|
|
|
|
50
|
|
|
def make_txt_basic_stats(self): |
51
|
|
|
"""Make Text Representation of Basic Statistics""" |
52
|
|
|
pstr_list = [] |
53
|
|
|
|
54
|
|
|
pstr_struct_header1 = "Basic Statistics " |
55
|
|
|
pstr_struct_header2 = '' |
56
|
|
|
|
57
|
|
|
pstr_list.append(pstr_struct_header1) |
58
|
|
|
pstr_list.append(pstr_struct_header2) |
59
|
|
|
|
60
|
|
|
template_str = ( |
61
|
|
|
" {0:^10} " |
62
|
|
|
" {1:>8} " |
63
|
|
|
" {2:<10} " |
64
|
|
|
" {3:>8} " |
65
|
|
|
" {4:<10} " |
66
|
|
|
) |
67
|
|
|
|
68
|
|
|
tmp_data = [ |
69
|
|
|
[ |
70
|
|
|
"Mean:", "{kyd_class.mean:.{kyd_class.precision}}".format( |
71
|
|
|
kyd_class=self.kyd_class), |
72
|
|
|
"", |
73
|
|
|
"Std Dev:", "{kyd_class.std:.{kyd_class.precision}}".format( |
74
|
|
|
kyd_class=self.kyd_class) |
75
|
|
|
], |
76
|
|
|
["Min:", "1Q:", "Median:", "3Q:", "Max:"], |
77
|
|
|
[ |
78
|
|
|
"{kyd_class.min: .{kyd_class.precision}}".format( |
79
|
|
|
kyd_class=self.kyd_class), |
80
|
|
|
"{kyd_class.firstquartile: .{kyd_class.precision}}".format( |
81
|
|
|
kyd_class=self.kyd_class), |
82
|
|
|
"{kyd_class.median: .{kyd_class.precision}}".format( |
83
|
|
|
kyd_class=self.kyd_class), |
84
|
|
|
"{kyd_class.thirdquartile: .{kyd_class.precision}}".format( |
85
|
|
|
kyd_class=self.kyd_class), |
86
|
|
|
"{kyd_class.max: .{kyd_class.precision}}".format( |
87
|
|
|
kyd_class=self.kyd_class), |
88
|
|
|
], |
89
|
|
|
['-99 CI:', '-95 CI:', '-68 CI:', '+68 CI:', '+95 CI:', '+99 CI:'], |
90
|
|
|
[ |
91
|
|
|
"{kyd_class.ci_99[0]: .{kyd_class.precision}}".format( |
92
|
|
|
kyd_class=self.kyd_class), |
93
|
|
|
"{kyd_class.ci_95[0]: .{kyd_class.precision}}".format( |
94
|
|
|
kyd_class=self.kyd_class), |
95
|
|
|
"{kyd_class.ci_68[0]: .{kyd_class.precision}}".format( |
96
|
|
|
kyd_class=self.kyd_class), |
97
|
|
|
"{kyd_class.ci_68[1]: .{kyd_class.precision}}".format( |
98
|
|
|
kyd_class=self.kyd_class), |
99
|
|
|
"{kyd_class.ci_95[1]: .{kyd_class.precision}}".format( |
100
|
|
|
kyd_class=self.kyd_class), |
101
|
|
|
"{kyd_class.ci_99[1]: .{kyd_class.precision}}".format( |
102
|
|
|
kyd_class=self.kyd_class), |
103
|
|
|
], |
104
|
|
|
] |
105
|
|
|
|
106
|
|
|
n_tmp_data = len(tmp_data) |
107
|
|
|
|
108
|
|
|
num_rows_in_cols = [len(i) for i in tmp_data] |
109
|
|
|
num_rows = np.max(num_rows_in_cols) |
110
|
|
|
|
111
|
|
|
for i in range(n_tmp_data): |
112
|
|
|
tmp_col = tmp_data[i] |
113
|
|
|
for j in range(num_rows_in_cols[i], num_rows): |
|
|
|
|
114
|
|
|
tmp_col.append("") |
115
|
|
|
|
116
|
|
|
for i in range(num_rows): |
117
|
|
|
pstr_list.append( |
118
|
|
|
template_str.format( |
119
|
|
|
tmp_data[0][i], |
120
|
|
|
tmp_data[1][i], |
121
|
|
|
tmp_data[2][i], |
122
|
|
|
tmp_data[3][i], |
123
|
|
|
tmp_data[4][i], |
124
|
|
|
) |
125
|
|
|
) |
126
|
|
|
|
127
|
|
|
return pstr_list |
128
|
|
|
|
129
|
|
|
def make_txt_struct(self): |
130
|
|
|
"""Make Text Representation of Array""" |
131
|
|
|
|
132
|
|
|
pstr_list = [] |
133
|
|
|
|
134
|
|
|
# pstr_struct_header0 = "................." |
135
|
|
|
# Commenting out Ansi Coloured Version |
136
|
|
|
# pstr_struct_header1 = '\033[1m' + "Array Structure " + '\033[0m' |
137
|
|
|
pstr_struct_header1 = "Array Structure " |
138
|
|
|
pstr_struct_header2 = " " |
139
|
|
|
|
140
|
|
|
# pstr_list.append(pstr_struct_header0) |
141
|
|
|
pstr_list.append(pstr_struct_header1) |
142
|
|
|
pstr_list.append(pstr_struct_header2) |
143
|
|
|
|
144
|
|
|
pstr_n_dim = ( |
145
|
|
|
"Number of Dimensions:\t" |
146
|
|
|
"{kyd_class.ndim}").format( |
147
|
|
|
kyd_class=self.kyd_class) |
148
|
|
|
pstr_list.append(pstr_n_dim) |
149
|
|
|
|
150
|
|
|
pstr_shape = ( |
151
|
|
|
"Shape of Dimensions:\t" |
152
|
|
|
"{kyd_class.shape}").format( |
153
|
|
|
kyd_class=self.kyd_class) |
154
|
|
|
pstr_list.append(pstr_shape) |
155
|
|
|
|
156
|
|
|
pstr_dtype = ( |
157
|
|
|
"Array Data Type:\t" |
158
|
|
|
"{kyd_class.dtype}").format( |
159
|
|
|
kyd_class=self.kyd_class) |
160
|
|
|
pstr_list.append(pstr_dtype) |
161
|
|
|
|
162
|
|
|
pstr_memsize = ( |
163
|
|
|
"Memory Size:\t\t" |
164
|
|
|
"{kyd_class.human_memsize}").format( |
165
|
|
|
kyd_class=self.kyd_class) |
166
|
|
|
pstr_list.append(pstr_memsize) |
167
|
|
|
|
168
|
|
|
pstr_spacer = ("") |
169
|
|
|
pstr_list.append(pstr_spacer) |
170
|
|
|
|
171
|
|
|
pstr_numnan = ( |
172
|
|
|
"Number of NaN:\t" |
173
|
|
|
"{kyd_class.num_nan}").format( |
174
|
|
|
kyd_class=self.kyd_class) |
175
|
|
|
pstr_list.append(pstr_numnan) |
176
|
|
|
|
177
|
|
|
pstr_numinf = ( |
178
|
|
|
"Number of Inf:\t" |
179
|
|
|
"{kyd_class.num_inf}").format( |
180
|
|
|
kyd_class=self.kyd_class) |
181
|
|
|
pstr_list.append(pstr_numinf) |
182
|
|
|
|
183
|
|
|
return pstr_list |
184
|
|
|
|
185
|
|
|
def make_text_repr(self): |
186
|
|
|
"""Making final text string for plain text representation""" |
187
|
|
|
|
188
|
|
|
tmp_text_repr = "" |
189
|
|
|
|
190
|
|
|
tmp_text_repr += "\n" |
191
|
|
|
|
192
|
|
|
pstr_basic = self.make_txt_basic_stats() |
193
|
|
|
pstr_struct = self.make_txt_struct() |
194
|
|
|
|
195
|
|
|
n_basic = len(pstr_basic) |
196
|
|
|
n_struct = len(pstr_struct) |
197
|
|
|
|
198
|
|
|
l_colwidth = max([len(x) for x in pstr_basic]) + 1 |
199
|
|
|
|
200
|
|
|
r_colwidth = max([len(x) for x in pstr_struct]) + 2 |
201
|
|
|
|
202
|
|
|
# new_colwidth = self.col_width + 20 |
203
|
|
|
|
204
|
|
|
# Finding the longest string |
205
|
|
|
len_list = max([n_basic, n_struct]) |
206
|
|
|
|
207
|
|
|
for i in range(len_list): |
208
|
|
|
tmp_str = '| ' |
209
|
|
|
if i < n_basic: |
210
|
|
|
tmp_str += (pstr_basic[i].ljust(l_colwidth)) |
211
|
|
|
else: |
212
|
|
|
tmp_str += ''.ljust(l_colwidth) |
213
|
|
|
tmp_str += ' | ' |
214
|
|
|
|
215
|
|
|
if i < n_struct: |
216
|
|
|
tmp_str += (pstr_struct[i].expandtabs().ljust(r_colwidth)) |
217
|
|
|
else: |
218
|
|
|
tmp_str += ''.ljust(r_colwidth) |
219
|
|
|
tmp_str += '\t|' |
220
|
|
|
|
221
|
|
|
tmp_text_repr += tmp_str + "\n" |
222
|
|
|
|
223
|
|
|
tmp_text_repr += "\n" |
224
|
|
|
self.text_repr = tmp_text_repr |
225
|
|
|
|
226
|
|
|
def __init__(self, kyd_class): |
227
|
|
|
super(KYD_datasummary, self).__init__() |
228
|
|
|
self.kyd_class = kyd_class |
229
|
|
|
self.make_text_repr() |
230
|
|
|
self.make_html_repr() |
231
|
|
|
|
232
|
|
|
|
233
|
|
|
class KYD(object): |
|
|
|
|
234
|
|
|
"""The Central Class for KYD""" |
235
|
|
|
|
236
|
|
|
# Variable for Data Vector |
237
|
|
|
data = None |
238
|
|
|
|
239
|
|
|
# Initial Flags |
240
|
|
|
f_allfinite = False |
241
|
|
|
f_allnonfinite = False |
242
|
|
|
f_hasnan = False |
243
|
|
|
f_hasinf = False |
244
|
|
|
|
245
|
|
|
# Initialized Numbers |
246
|
|
|
num_nan = 0 |
247
|
|
|
num_inf = 0 |
248
|
|
|
|
249
|
|
|
# Display Settings |
250
|
|
|
col_width = 10 |
251
|
|
|
precision = 4 |
252
|
|
|
|
253
|
|
|
def check_finite(self): |
254
|
|
|
"""Checking to see if all elements are finite and setting flags""" |
255
|
|
|
if np.all(np.isfinite(self.data)): |
256
|
|
|
self.filt_data = self.data |
257
|
|
|
self.f_allfinite = True |
258
|
|
|
else: |
259
|
|
|
finite_inds = np.where(np.isfinite(self.data)) |
260
|
|
|
|
261
|
|
|
self.filt_data = self.data[finite_inds] |
262
|
|
|
|
263
|
|
|
if self.filt_data.size == 0: |
264
|
|
|
self.f_allnonfinite = True |
265
|
|
|
|
266
|
|
|
if np.any(np.isnan(self.data)): |
267
|
|
|
self.f_hasnan = True |
268
|
|
|
self.num_nan = np.sum(np.isnan(self.data)) |
269
|
|
|
|
270
|
|
|
if np.any(np.isinf(self.data)): |
271
|
|
|
self.f_hasinf = True |
272
|
|
|
self.num_inf = np.sum(np.isinf(self.data)) |
273
|
|
|
|
274
|
|
|
def check_struct(self): |
275
|
|
|
"""Determining the Structure of the Numpy Array""" |
276
|
|
|
self.dtype = self.data.dtype |
277
|
|
|
self.ndim = self.data.ndim |
278
|
|
|
self.shape = self.data.shape |
279
|
|
|
self.size = self.data.size |
280
|
|
|
self.memsize = sys.getsizeof(self.data) |
281
|
|
|
self.human_memsize = sizeof_fmt(self.memsize) |
282
|
|
|
|
283
|
|
|
def get_basic_stats(self): |
284
|
|
|
"""Get basic statistics about array""" |
285
|
|
|
|
286
|
|
|
if self.f_allnonfinite: |
287
|
|
|
self.min = self.max = self.range = np.nan |
288
|
|
|
self.mean = self.std = self.median = np.nan |
289
|
|
|
self.firstquartile = self.thirdquartile = np.nan |
290
|
|
|
self.ci_68 = self.ci_95 = self.ci_99 = np.array([np.nan, np.nan]) |
291
|
|
|
|
292
|
|
|
return |
293
|
|
|
|
294
|
|
|
self.min = np.float_(np.min(self.filt_data)) |
295
|
|
|
self.max = np.float_(np.max(self.filt_data)) |
296
|
|
|
self.range = self.max - self.min |
297
|
|
|
self.mean = np.mean(self.filt_data) |
298
|
|
|
self.std = np.std(self.filt_data) |
299
|
|
|
self.median = np.float_(np.median(self.filt_data)) |
300
|
|
|
self.firstquartile = np.float_(np.percentile(self.filt_data, 25)) |
301
|
|
|
self.thirdquartile = np.float_(np.percentile(self.filt_data, 75)) |
302
|
|
|
self.ci_99 = np.float_( |
303
|
|
|
np.percentile(self.filt_data, np.array([0.5, 99.5]))) |
304
|
|
|
self.ci_95 = np.float_( |
305
|
|
|
np.percentile(self.filt_data, np.array([2.5, 97.5]))) |
306
|
|
|
self.ci_68 = np.float_( |
307
|
|
|
np.percentile(self.filt_data, np.array([16.0, 84.0]))) |
308
|
|
|
|
309
|
|
|
def make_summary(self): |
310
|
|
|
"""Making Data Summary""" |
311
|
|
|
self.data_summary = KYD_datasummary(self) |
312
|
|
|
|
313
|
|
|
def clear_memory(self): |
314
|
|
|
"""Ensuring the Numpy Array does not exist in memory""" |
315
|
|
|
del self.data |
316
|
|
|
del self.filt_data |
317
|
|
|
|
318
|
|
|
def display(self, short=False): |
319
|
|
|
"""Displaying all relevant statistics""" |
320
|
|
|
|
321
|
|
|
if short: |
322
|
|
|
pass |
323
|
|
|
try: |
324
|
|
|
get_ipython |
325
|
|
|
display(self.data_summary) |
326
|
|
|
except NameError: |
327
|
|
|
print(self.data_summary) |
328
|
|
|
|
329
|
|
|
def __init__(self, data): |
330
|
|
|
super(KYD, self).__init__() |
331
|
|
|
|
332
|
|
|
# Ensuring that the array is a numpy array |
333
|
|
|
if not isinstance(data, np.ndarray): |
334
|
|
|
data = np.array(data) |
335
|
|
|
|
336
|
|
|
self.data = data |
337
|
|
|
|
338
|
|
|
self.check_finite() |
339
|
|
|
self.check_struct() |
340
|
|
|
self.get_basic_stats() |
341
|
|
|
self.clear_memory() |
342
|
|
|
self.make_summary() |
343
|
|
|
|
344
|
|
|
|
345
|
|
|
def sizeof_fmt(num, suffix='B'): |
346
|
|
|
"""Return human readable version of in-memory size. |
347
|
|
|
Code from Fred Cirera from Stack Overflow: |
348
|
|
|
https://stackoverflow.com/questions/1094841/reusable-library-to-get-human-readable-version-of-file-size |
349
|
|
|
""" |
350
|
|
|
for unit in ['', 'Ki', 'Mi', 'Gi', 'Ti', 'Pi', 'Ei', 'Zi']: |
351
|
|
|
if abs(num) < 1024.0: |
352
|
|
|
return "%3.1f%s%s" % (num, unit, suffix) |
353
|
|
|
num /= 1024.0 |
354
|
|
|
return "%.1f%s%s" % (num, 'Yi', suffix) |
355
|
|
|
|
356
|
|
|
|
357
|
|
|
def kyd(data, full_statistics=False): |
358
|
|
|
"""Print statistics of any numpy array |
359
|
|
|
|
360
|
|
|
data -- Numpy Array of Data |
361
|
|
|
|
362
|
|
|
Keyword arguments: |
363
|
|
|
full_statistics -- printing all detailed statistics of the sources |
364
|
|
|
(Currently Not Implemented) |
365
|
|
|
|
366
|
|
|
""" |
367
|
|
|
|
368
|
|
|
data_kyd = KYD(data) |
369
|
|
|
if full_statistics: |
370
|
|
|
data_kyd.display() |
371
|
|
|
else: |
372
|
|
|
data_kyd.display(short=True) |
373
|
|
|
|
374
|
|
|
return data_kyd |
375
|
|
|
|
This check looks for invalid names for a range of different identifiers.
You can set regular expressions to which the identifiers must conform if the defaults do not match your requirements.
If your project includes a Pylint configuration file, the settings contained in that file take precedence.
To find out more about Pylint, please refer to their site.