|
1
|
|
|
import logging |
|
|
|
|
|
|
2
|
|
|
import math |
|
3
|
|
|
from datetime import date, datetime, timedelta |
|
4
|
|
|
from typing import SupportsFloat |
|
5
|
|
|
|
|
6
|
|
|
import regex |
|
|
|
|
|
|
7
|
|
|
from pint import Quantity, UnitRegistry |
|
|
|
|
|
|
8
|
|
|
from pint.errors import PintTypeError |
|
|
|
|
|
|
9
|
|
|
|
|
10
|
|
|
from pocketutils.core.exceptions import OutOfRangeError, StringPatternError |
|
|
|
|
|
|
11
|
|
|
from pocketutils.tools.string_tools import StringTools |
|
12
|
|
|
|
|
13
|
|
|
logger = logging.getLogger("pocketutils") |
|
14
|
|
|
_UNIT_REG = UnitRegistry() |
|
15
|
|
|
|
|
16
|
|
|
|
|
17
|
|
|
class UnitTools: |
|
|
|
|
|
|
18
|
|
|
@classmethod |
|
19
|
|
|
def format_approx_big_number(cls, n: int) -> str: |
|
|
|
|
|
|
20
|
|
|
for k, v in {1e15: "", 1e12: "T", 1e9: "B", 1e6: "M", 1e3: "k"}.items(): |
|
|
|
|
|
|
21
|
|
|
if n >= k: |
|
22
|
|
|
return str(n // k) + v |
|
23
|
|
|
return str(n) |
|
24
|
|
|
|
|
25
|
|
|
@classmethod |
|
26
|
|
|
def approx_time_wrt( |
|
|
|
|
|
|
27
|
|
|
cls, |
|
|
|
|
|
|
28
|
|
|
now: date | datetime, |
|
|
|
|
|
|
29
|
|
|
then: date | datetime, |
|
|
|
|
|
|
30
|
|
|
*, |
|
|
|
|
|
|
31
|
|
|
skip_today: bool = False, |
|
|
|
|
|
|
32
|
|
|
sig: int = 3, |
|
|
|
|
|
|
33
|
|
|
) -> str: |
|
34
|
|
|
""" |
|
35
|
|
|
Describes ``then`` with higher resolution for smaller differences to ``now``. |
|
36
|
|
|
|
|
37
|
|
|
Examples: |
|
38
|
|
|
- ``approx_time_wrt(date(2021, 1, 12), date(1996, 10, 1)) # "1996"`` |
|
39
|
|
|
- ``approx_time_wrt(date(2021, 1, 12), date(2021, 10, 1)) # "2021-01-12"`` |
|
40
|
|
|
- ``approx_time_wrt(date(2021, 10, 1), datetime(2021, 10, 1, 11, 55)) # "2021-01-12 11:55"`` |
|
|
|
|
|
|
41
|
|
|
- ``approx_time_wrt(date(2021, 10, 1), datetime(2021, 10, 1, 11, 0, 0, 30, 222222)) # "2021-01-12 00:00:30"`` |
|
|
|
|
|
|
42
|
|
|
- ``approx_time_wrt(date(2021, 10, 1), datetime(2021, 10, 1, 11, 0, 0, 2, 222222)) # "2021-01-12 00:00:02.222"`` |
|
|
|
|
|
|
43
|
|
|
- ``approx_time_wrt(date(2021, 10, 1), datetime(2021, 10, 1, 11, 0, 0, 2, 22)) # "2021-01-12 00:00:02.000022"`` |
|
|
|
|
|
|
44
|
|
|
""" |
|
45
|
|
|
delta = now - then if now > then else then - now |
|
46
|
|
|
tot_days = (delta.days) + (delta.seconds / 86400) + (delta.microseconds / 86400 / 10**6) |
|
47
|
|
|
tot_secs = tot_days * 86400 |
|
48
|
|
|
_today = "" if skip_today and then.date() == now.date() else "%Y-%m-%d " |
|
49
|
|
|
if tot_days > sig * 365.24219: |
|
|
|
|
|
|
50
|
|
|
return str(then.year) |
|
51
|
|
|
elif tot_days > sig * 30.437: |
|
52
|
|
|
return then.strftime("%Y-%m") |
|
53
|
|
|
elif tot_days > sig: |
|
54
|
|
|
return then.strftime("%Y-%m-%d") |
|
55
|
|
|
elif tot_secs > sig * 60: |
|
56
|
|
|
return then.strftime(_today + "%H:%M") |
|
57
|
|
|
elif tot_secs > sig: |
|
58
|
|
|
return then.strftime(_today + "%H:%M:%S") |
|
59
|
|
|
elif tot_secs > sig / 1000: |
|
60
|
|
|
return then.strftime(_today + "%H:%M:%S") + "." + str(round(then.microsecond / 1000)) |
|
61
|
|
|
else: |
|
62
|
|
|
return then.strftime(_today + "%H:%M:%S.%f") |
|
63
|
|
|
|
|
64
|
|
|
@classmethod |
|
65
|
|
|
def delta_time_to_str(cls, delta_sec: float | timedelta, *, space: str = "") -> str: |
|
66
|
|
|
""" |
|
67
|
|
|
Returns a pretty string from a difference in time in seconds. |
|
68
|
|
|
Rounds hours and minutes to 2 decimal places, and seconds to 1. |
|
69
|
|
|
Ex: delta_time_to_str(313) == 5.22min |
|
70
|
|
|
delta_sec: The time in seconds |
|
71
|
|
|
space: Space char between digits and units; |
|
72
|
|
|
good choices are empty, ASCII space, Chars.narrownbsp, Chars.thinspace, |
|
73
|
|
|
and Chars.nbsp. |
|
74
|
|
|
|
|
75
|
|
|
Returns: |
|
76
|
|
|
A string with units 'hr', 'min', or 's' |
|
77
|
|
|
""" |
|
78
|
|
|
if isinstance(delta_sec, timedelta): |
|
79
|
|
|
delta_sec = delta_sec.total_seconds() |
|
80
|
|
|
if abs(delta_sec) > 60 * 60 * 3: |
|
|
|
|
|
|
81
|
|
|
return ( |
|
82
|
|
|
StringTools.strip_empty_decimal(str(round(delta_sec / 60 / 60, 2))) + space + "hr" |
|
83
|
|
|
) |
|
84
|
|
|
elif abs(delta_sec) > 180: |
|
85
|
|
|
return StringTools.strip_empty_decimal(str(round(delta_sec / 60, 2))) + space + "min" |
|
86
|
|
|
else: |
|
87
|
|
|
return StringTools.strip_empty_decimal(str(round(delta_sec, 1))) + space + "s" |
|
88
|
|
|
|
|
89
|
|
|
@classmethod |
|
90
|
|
|
def ms_to_minsec(cls, ms: int, space: str = "") -> str: |
|
|
|
|
|
|
91
|
|
|
""" |
|
92
|
|
|
Converts a number of milliseconds to one of the following formats. |
|
93
|
|
|
Will be one of these: |
|
94
|
|
|
- 10ms if < 1 sec |
|
95
|
|
|
- 10:15 if < 1 hour |
|
96
|
|
|
- 10:15:33 if < 1 day |
|
97
|
|
|
- 5d:10:15:33 if > 1 day |
|
98
|
|
|
Prepends a minus sign (−) if negative. |
|
99
|
|
|
|
|
100
|
|
|
Args: |
|
101
|
|
|
ms: The milliseconds |
|
102
|
|
|
space: Space char between digits and 'ms' (if used); |
|
103
|
|
|
good choices are empty, ASCII space, Chars.narrownbsp, |
|
104
|
|
|
Chars.thinspace, and Chars.nbsp. |
|
105
|
|
|
|
|
106
|
|
|
Returns: |
|
107
|
|
|
A string of one of the formats above |
|
108
|
|
|
""" |
|
109
|
|
|
ms = abs(int(ms)) |
|
110
|
|
|
seconds = int((ms / 1000) % 60) |
|
111
|
|
|
minutes = int((ms / (1000 * 60)) % 60) |
|
112
|
|
|
hours = int((ms / (1000 * 60 * 60)) % 24) |
|
113
|
|
|
days = int(ms / (1000 * 60 * 60 * 24)) |
|
114
|
|
|
z_hr = str(hours).zfill(2) |
|
115
|
|
|
z_min = str(minutes).zfill(2) |
|
116
|
|
|
z_sec = str(seconds).zfill(2) |
|
117
|
|
|
sgn = "−" if ms < 0 else "" |
|
118
|
|
|
if ms < 1000: |
|
|
|
|
|
|
119
|
|
|
return f"{sgn}{ms}{space}ms" |
|
120
|
|
|
elif days > 1: |
|
121
|
|
|
return f"{days}d:{z_hr}:{z_min}:{z_sec}" |
|
122
|
|
|
elif hours > 1: |
|
123
|
|
|
return f"{sgn}{z_hr}:{z_min}:{z_sec}" |
|
124
|
|
|
else: |
|
125
|
|
|
return f"{sgn}{z_min}:{z_sec}" |
|
126
|
|
|
|
|
127
|
|
|
@classmethod |
|
128
|
|
|
def round_to_sigfigs(cls, num: SupportsFloat, sig_figs: int | None) -> float: |
|
129
|
|
|
""" |
|
130
|
|
|
Round to specified number of sigfigs. |
|
131
|
|
|
|
|
132
|
|
|
Args: |
|
133
|
|
|
num: A Python or Numpy float or something that supports __float__ |
|
134
|
|
|
sig_figs: The number of significant figures, non-negative |
|
135
|
|
|
|
|
136
|
|
|
Returns: |
|
137
|
|
|
A Python integer |
|
138
|
|
|
""" |
|
139
|
|
|
if sig_figs is None: |
|
140
|
|
|
return float(num) |
|
141
|
|
|
if sig_figs < 0: |
|
142
|
|
|
raise OutOfRangeError(f"sig_figs {sig_figs} is negative", minimum=0) |
|
143
|
|
|
num = float(num) |
|
144
|
|
|
if num != 0: |
|
|
|
|
|
|
145
|
|
|
digits = -int(math.floor(math.log10(abs(num))) - (sig_figs - 1)) |
|
146
|
|
|
return round(num, digits) |
|
147
|
|
|
else: |
|
148
|
|
|
return 0 # can't take the log of 0 |
|
149
|
|
|
|
|
150
|
|
|
@classmethod |
|
151
|
|
|
def format_micromolar( |
|
152
|
|
|
cls, |
|
|
|
|
|
|
153
|
|
|
micromolar: float, |
|
|
|
|
|
|
154
|
|
|
n_sigfigs: int | None = 5, |
|
|
|
|
|
|
155
|
|
|
*, |
|
|
|
|
|
|
156
|
|
|
adjust_units: bool = True, |
|
|
|
|
|
|
157
|
|
|
use_sigfigs: bool = True, |
|
|
|
|
|
|
158
|
|
|
space: str = "", |
|
|
|
|
|
|
159
|
|
|
) -> str: |
|
160
|
|
|
""" |
|
161
|
|
|
Returns a concentration with units, with the units scaled as needed. |
|
162
|
|
|
Can handle millimolar, micromolar, nanomolar, and picomolar. |
|
163
|
|
|
|
|
164
|
|
|
Args: |
|
165
|
|
|
micromolar: Value |
|
166
|
|
|
n_sigfigs: For rounding; no rounding if None |
|
167
|
|
|
adjust_units: If False, will always use micromolar |
|
168
|
|
|
use_sigfigs: If True, rounds to a number of significant figures; otherwise round to decimal places |
|
|
|
|
|
|
169
|
|
|
space: Space char between digits and units; |
|
170
|
|
|
good choices are empty, ASCII space, |
|
171
|
|
|
:attr:`pocketutils.core.chars.Chars.narrownbsp`, |
|
172
|
|
|
:attr:`pocketutils.core.chars.Chars.thinspace`, |
|
173
|
|
|
and :attr:`pocketutils.core.chars.Chars.nbsp`. |
|
174
|
|
|
|
|
175
|
|
|
Returns: |
|
176
|
|
|
The concentration with a suffix of µM, mM, nM, or mM |
|
177
|
|
|
""" |
|
178
|
|
|
d = micromolar |
|
|
|
|
|
|
179
|
|
|
m = abs(d) |
|
|
|
|
|
|
180
|
|
|
unit = "µM" |
|
181
|
|
|
if adjust_units: |
|
182
|
|
|
if m < 1e-6: |
|
183
|
|
|
d *= 1e9 |
|
|
|
|
|
|
184
|
|
|
unit = "fM" |
|
185
|
|
|
elif m < 1e-3: |
|
186
|
|
|
d *= 1e6 |
|
|
|
|
|
|
187
|
|
|
unit = "pM" |
|
188
|
|
|
elif m < 1: |
|
189
|
|
|
d *= 1e3 |
|
|
|
|
|
|
190
|
|
|
unit = "nM" |
|
191
|
|
|
elif m >= 1e6: |
|
192
|
|
|
d /= 1e6 |
|
|
|
|
|
|
193
|
|
|
unit = "M" |
|
194
|
|
|
elif m >= 1e3: |
|
195
|
|
|
d /= 1e3 |
|
|
|
|
|
|
196
|
|
|
unit = "mM" |
|
197
|
|
|
if n_sigfigs is None: |
|
198
|
|
|
pass |
|
199
|
|
|
elif use_sigfigs: |
|
200
|
|
|
d = cls.round_to_sigfigs(d, n_sigfigs) |
|
|
|
|
|
|
201
|
|
|
else: |
|
202
|
|
|
d = round(d, n_sigfigs) |
|
|
|
|
|
|
203
|
|
|
if round(d) == d and str(d).endswith(".0"): |
|
|
|
|
|
|
204
|
|
|
return str(d)[:-2] + space + unit |
|
205
|
|
|
else: |
|
206
|
|
|
return str(d) + space + unit |
|
207
|
|
|
|
|
208
|
|
|
@classmethod |
|
209
|
|
|
def split_species_micromolar(cls, text: str) -> tuple[str, float | None]: |
|
|
|
|
|
|
210
|
|
|
""" |
|
211
|
|
|
Splits a name into a chemical/concentration pair, falling back with the full name. |
|
212
|
|
|
Ex: "abc 3.5uM" → (abc, 3.5) |
|
213
|
|
|
Ex: "abc 3.5 µM" → (abc, 3.5) |
|
214
|
|
|
Ex: "abc (3.5mM)" → (abc, 3500.0) |
|
215
|
|
|
Ex: "abc 3.5mM" → (abc, None) |
|
216
|
|
|
Ex: "3.5mM" → (3.5mM, None) # an edge case: don't pass in only units |
|
217
|
|
|
Uses a moderately strict pattern for the drug and dose: |
|
218
|
|
|
- The dose must terminate the string, except for end parenthesis or whitespace. |
|
219
|
|
|
- The drug and dose must be separated by at least one non-alphanumeric, non-dot, non-hyphen character. |
|
|
|
|
|
|
220
|
|
|
- Units must follow the digits, separated by at most whitespace, and are case-sensitive. |
|
221
|
|
|
""" |
|
222
|
|
|
# lazy ops in the first group and in the non-(alphanumeric/dot/dash) separator between the drug and dose |
|
|
|
|
|
|
223
|
|
|
pat = regex.compile( |
|
224
|
|
|
r"^\s*(.*?)(?:[^A-Za-z0-9.\-]+?[\s(\[{]*(\d+(?:.\d*)?)\s*([mµunpf]M)\s*[)\]}]*)?\s*$", |
|
225
|
|
|
flags=regex.V1, |
|
226
|
|
|
) |
|
227
|
|
|
match = pat.fullmatch(text) |
|
228
|
|
|
if match is None: |
|
229
|
|
|
raise StringPatternError(f"Text {text} couldn't be parsed", value=text, pattern=pat) |
|
230
|
|
|
if match.group(2) is None: |
|
|
|
|
|
|
231
|
|
|
return text.strip(), None |
|
232
|
|
|
else: |
|
233
|
|
|
drug = match.group(1).strip("([{)]}") |
|
234
|
|
|
dose = UnitTools.concentration_to_micromolar(float(match.group(2)), match.group(3)) |
|
235
|
|
|
return drug, dose |
|
236
|
|
|
|
|
237
|
|
|
@classmethod |
|
238
|
|
|
def extract_micromolar(cls, text: str) -> float | None: |
|
239
|
|
|
""" |
|
240
|
|
|
Returns what looks like a concentration with units. Accepts one of: mM, µM, uM, nM, pM. |
|
241
|
|
|
Searches pretty flexibly. |
|
242
|
|
|
If no matches are found, returns None. |
|
243
|
|
|
If multiple matches are found, warns and returns None. |
|
244
|
|
|
""" |
|
245
|
|
|
# we need to make sure mM ex isn't part of a larger name |
|
246
|
|
|
pat1 = regex.compile(r"(\d+(?:.\d*)?)\s*([mµunpf]M)\s*[)\]}]*", flags=regex.V1) |
|
247
|
|
|
|
|
248
|
|
|
def find(pat): |
|
249
|
|
|
return { |
|
250
|
|
|
UnitTools.concentration_to_micromolar(float(match.group(1)), match.group(2)) |
|
251
|
|
|
for match in pat.finditer(text) |
|
252
|
|
|
if match is not None |
|
253
|
|
|
} |
|
254
|
|
|
|
|
255
|
|
|
matches = find(pat1) |
|
256
|
|
|
if len(matches) == 1: |
|
|
|
|
|
|
257
|
|
|
return next(iter(matches)) |
|
258
|
|
|
elif len(matches) > 1: |
|
259
|
|
|
logger.warning(f"Found {len(matches)} potential doses: {matches} . Returning None.") |
|
|
|
|
|
|
260
|
|
|
return None |
|
261
|
|
|
|
|
262
|
|
|
@classmethod |
|
263
|
|
|
def concentration_to_micromolar(cls, digits: SupportsFloat, units: str) -> float: |
|
264
|
|
|
""" |
|
265
|
|
|
Converts a concentration with units to micromolar. |
|
266
|
|
|
|
|
267
|
|
|
Args: |
|
268
|
|
|
digits: Float or float-compatible value |
|
269
|
|
|
units: Units that ``digits`` are in |
|
270
|
|
|
|
|
271
|
|
|
Example: |
|
272
|
|
|
.. code-block:: |
|
273
|
|
|
|
|
274
|
|
|
concentration_to_micromolar(53, 'nM') # returns 0.053 |
|
275
|
|
|
|
|
276
|
|
|
See Also: |
|
277
|
|
|
:meth:`extract_micromolar` |
|
278
|
|
|
""" |
|
279
|
|
|
return ( |
|
280
|
|
|
float(digits) |
|
281
|
|
|
* { |
|
282
|
|
|
"M": 1e6, |
|
283
|
|
|
"mM": 1e3, |
|
284
|
|
|
"µM": 1, |
|
285
|
|
|
"uM": 1, |
|
286
|
|
|
"nM": 1e-3, |
|
287
|
|
|
"pM": 1e-6, |
|
288
|
|
|
"fM": 1e-9, |
|
289
|
|
|
}[units] |
|
290
|
|
|
) |
|
291
|
|
|
|
|
292
|
|
|
@classmethod |
|
293
|
|
|
def canonicalize_quantity(cls, s: str, dimensionality: str) -> Quantity: |
|
|
|
|
|
|
294
|
|
|
""" |
|
295
|
|
|
Returns a quantity in reduced units from a magnitude with units. |
|
296
|
|
|
|
|
297
|
|
|
Args: |
|
298
|
|
|
s: The string to parse; e.g. ``"1 m/s^2"``. |
|
299
|
|
|
Unit names and symbols permitted, and spaces may be omitted. |
|
300
|
|
|
dimensionality: The resulting Quantity is checked against this; |
|
301
|
|
|
e.g. ``"[length]/[meter]^2"`` |
|
302
|
|
|
|
|
303
|
|
|
Returns: |
|
304
|
|
|
a pint ``Quantity`` |
|
305
|
|
|
|
|
306
|
|
|
Raise: |
|
307
|
|
|
PintTypeError: If the dimensionality is inconsistent |
|
308
|
|
|
""" |
|
309
|
|
|
q = _UNIT_REG.Quantity(s).to_reduced_units() |
|
|
|
|
|
|
310
|
|
|
if not q.is_compatible_with(dimensionality): |
|
311
|
|
|
raise PintTypeError(f"{s} not of dimensionality {dimensionality}") |
|
312
|
|
|
return q |
|
313
|
|
|
|
|
314
|
|
|
|
|
315
|
|
|
__all__ = ["UnitTools"] |
|
316
|
|
|
|
dmyersturnbull /
pocketutils
Loading history...