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
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
|
from qolab.file_utils import save_table_with_header, infer_compression
import datetime
import numpy as np
import yaml
import pandas
import os
def headerFromDictionary(d, prefix=''):
"""Converts dictionary to YAML format with optional prefix for every line."""
header = []
tail = yaml.dump(d, default_flow_style=False, sort_keys=False)
tail = tail.split('\n')
header.extend(tail)
prefixed_header = [prefix+l for l in header]
return prefixed_header
def from_timestamps_to_dates(timestamps):
"""Formats timestamps to datetime format"""
dates = [datetime.datetime.fromtimestamp(float(ts)) for ts in timestamps]
return(dates)
def loadTraceRawHeaderAndData(fname, tryCompressedIfMissing=True):
"""Load trace file and return header and data.
Attempts to load a compressed file if the main is missing
and `tryCompressedIfMissing` is set to `True`.
E.g. if we try to load 'data.dat' file and it is missing, attempt to load 'data.dat.gz'
this option is mainly for compatibility with old scripts which are not aware of compressed files
"""
headerstr=[]
data = None
if (not os.path.exists(fname)) and tryCompressedIfMissing:
# attempt to locate compressed file for the missing base file
for ext in ['gz', 'bz', 'bz2']:
if os.path.exists(fname+'.'+ext):
fname += '.'+ext
break
# we will try to guess if the file compressed
_open = open
compression = infer_compression(fname)
print(compression)
if compression == 'gzip':
# TODO improve detection: gzip files have first 2 bytes set to b'\x1f\x8b'
import gzip
_open = gzip.open
elif compression == 'bzip':
# TODO improve detection: bzip files have first 2 bytes set to b'BZ'
import bz2
_open = bz2.open
with _open(fname, mode='rb') as tracefile:
# Reading yaml header prefixed by '% '
# It sits at the top and below is just data in TSV format
while True:
ln = tracefile.readline()
if ln[0:2]==b'% ':
headerstr.append(ln[2:].decode('utf-8'))
else:
break
header=yaml.load(str.join('\n', headerstr), Loader=yaml.BaseLoader)
# now we load the data itself
# data=np.genfromtxt(fname, comments='%', delimiter='\t')
# Note: pandas reads csv faster by factor of 8 then numpy.genfromtxt
# data=pandas.read_csv('/home/evmik/hopping_trace_20220706_02141.dat', comment='%', delimiter='\t', header=None)
tracefile.seek(0) # rewind file to the beginning
df = pandas.read_csv(tracefile, comment='%', delimiter='\t', header=None)
data = df.to_numpy()
return(header, data)
def loadTrace(fname, tryCompressedIfMissing=True):
"""Load trace file."""
(header, data) = loadTraceRawHeaderAndData(fname, tryCompressedIfMissing=tryCompressedIfMissing)
return traceFromHeaderAndData(header, data)
def traceFromHeaderAndData(header, data=None):
"""Generate trace class from it description (header) and data."""
label = None
model = None
tr = None
if 'config' not in header:
print('Error: trace has now config')
return None
else:
if 'label' in header['config']:
label = header['config']['label']
if 'model' not in header['config']:
print('Error: unknown trace model')
return None
else:
model = header['config']['model']
if model == 'Trace':
tr = Trace(label)
if data is not None:
tr.values = data
elif model == 'TraceXY':
tx = traceFromHeaderAndData(header['TraceX'])
ty = traceFromHeaderAndData(header['TraceY'])
if data is not None:
tx.values=data[:,0]
ty.values=data[:,1]
tr = TraceXY(label)
tr.x = tx
tr.y = ty
elif model == 'TraceSetSameX':
tx = traceFromHeaderAndData(header['TraceX'])
tx.values=data[:,0]
tr = TraceSetSameX(label)
tr.addTraceX(tx)
ytrs_header = header['TraceY']
cnt=0
for l, h in ytrs_header.items():
ty = traceFromHeaderAndData(h)
cnt += 1
ty.values=data[:,cnt]
trxy = TraceXY(l)
trxy.x = tx
trxy.y = ty
tr.addTrace( trxy )
else:
print(f'Error: unknown trace model: {model}')
return None
tr.config = header['config']
return tr
class Trace:
"""Base Trace class, which holds only one variable"""
def __init__(self, label):
self.config = {}
self.config['label'] = label
self.config['model'] = 'Trace'
self.config['version'] = '0.1'
# 'type' is useful to indicate way of representation, make sense on y_vs_x traces
# if set to none we have normal y vs x
# if set to 'timestamp' x will be converted to datetime dates
self.config['type'] = None
self.config['item_format']='.15e'
self.config['tags'] = {}
self.last_saved_pos = 0
self._trace_specific_init()
self.clear_data()
def _trace_specific_init(self):
self.config['unit'] = None
self.values = np.empty(0)
self.last_saved_pos = 0
def clear_last_saved_pos(self):
self.last_saved_pos = 0
def clear_data(self):
self.clear_last_saved_pos()
if self.values is not None:
self.values = np.empty(0, dtype=self.values.dtype)
def __repr__(self):
lbl = self.config['label']
cls_name = f"{self.__class__.__name__}('{lbl}'"
return "".join([cls_name, f', N={self.values.size}', ')'])
def plot(self):
import matplotlib.pyplot as plt
x=self.values
if self.config['type'] is not None:
if self.config['type'] == 'timestamp':
x = from_timestamps_to_dates(self.values)
plt.plot(x, label=self.config['label'])
plt.xlabel('index')
plt.ylabel(f"{self.config['unit']}")
plt.legend()
plt.grid(True)
def getConfig(self):
d ={}
d['config'] = self.config.copy()
return( d )
def getData(self):
return( self.values )
def getHeader(self, prefix=''):
d = self.getConfig()
return headerFromDictionary(d, prefix='')
def save(self, fname, last_saved_pos=None, skip_headers_if_file_exist=False, **kwargs):
if last_saved_pos is None:
last_saved_pos = self.last_saved_pos
data = self.getData()
if last_saved_pos > 0:
skip_headers_if_file_exist=True
fname = save_table_with_header(fname, data[last_saved_pos:,:], self.getHeader(), item_format=self.config['item_format'], skip_headers_if_file_exist=skip_headers_if_file_exist, **kwargs)
self.last_saved_pos = data.shape[0]
return(fname)
def addPoint(self, val):
self.values = np.append(self.values, val)
class TraceXY(Trace):
"""Data structure to handle two variables X and Y, handy for Y vs X data arrangements."""
def __init__(self, label):
super().__init__(label)
self.config['model'] = 'TraceXY'
def _trace_specific_init(self):
self.x = None
self.y = None
def clear_data(self):
self.clear_last_saved_pos()
if self.x is not None:
self.x.clear_data()
if self.y is not None:
self.y.clear_data()
def __repr__(self):
lbl = self.config['label']
cls_name = f"{self.__class__.__name__}('{lbl}'"
xlabel= f"{self.x.config['label']}"
xparam= f", {self.x}"
yparam= f", {self.y}"
return "".join([cls_name, xparam, yparam, ')'])
def plot(self):
import matplotlib.pyplot as plt
x=self.x.values
if self.x.config['type'] is not None:
if self.x.config['type'] == 'timestamp':
x = from_timestamps_to_dates(x)
plt.plot(x, self.y.values, label=self.config['label'])
plt.xlabel(f"{self.x.config['label']} ({self.x.config['unit']})")
plt.ylabel(f"{self.y.config['label']} ({self.y.config['unit']})")
plt.legend()
plt.grid(True)
def getConfig(self):
config = {}
config['config'] = self.config.copy()
config['TraceX'] = {}
config['TraceX'] = self.x.getConfig()
config['TraceY'] = {}
config['TraceY'] = self.y.getConfig()
return( config )
def getData(self):
data=self.x.values
if data.ndim == 1:
data = data[:, np.newaxis]
vals = self.y.values
if vals.ndim == 1:
vals = vals[:, np.newaxis]
data=np.concatenate((data, vals), 1)
return( data )
def addPoint(self, valX, valY):
self.x.values = np.append(self.x.values, valX)
self.y.values = np.append(self.y.values, valY)
class TraceSetSameX(Trace):
"""Data structure to handle multiple Ys vs X dependencies, handy for scope traces."""
def __init__(self, label):
super().__init__(label)
self.config['model'] = 'TraceSetSameX'
def _trace_specific_init(self):
self.x = None
self.traces={}
def clear_data(self):
self.clear_last_saved_pos()
if self.x is not None:
self.x.clear_data()
for k, tr in self.traces.items():
tr.clear_data()
def __repr__(self):
lbl = self.config['label']
cls_name = f"{self.__class__.__name__}('{lbl}'"
xparam= f", x: {self.x}"
yparam = f", traces: {list(self.traces.keys())}"
return "".join([cls_name, xparam, yparam, ')'])
def addTraceX(self, tr):
self.x = tr
def addTrace(self, tr):
if tr.config['model'] == 'TraceXY':
if self.x == None:
self.x = tr.x
trY = tr.y
self.traces[tr.config['label']]=trY
elif tr.config['model'] == 'Trace':
if self.x == None:
self.x = tr
else:
self.traces[tr.config['label']]=tr
def plot(self):
import matplotlib.pyplot as plt
nplots = len(self.traces.keys())
fig = plt.gcf()
fig, axs = plt.subplots(nplots, 1, sharex=True, num=fig.number)
cnt=0
x=self.x.values
if self.x.config['type'] is not None:
if self.x.config['type'] == 'timestamp':
x = from_timestamps_to_dates(x)
for k, tr in self.traces.items():
p=axs[cnt].plot(x, tr.values, label=k)
axs[cnt].set_ylabel(f"{tr.config['label']} ({tr.config['unit']})")
axs[cnt].legend()
axs[cnt].grid(True)
cnt+=1
plt.xlabel(f"{self.x.config['label']} ({self.x.config['unit']})")
def items(self):
return (self.traces.items())
def keys(self):
return (self.traces.keys())
def getTrace(self, label):
tr = TraceXY(label)
tr.x = self.x
tr.y = self.traces[label]
return (tr)
def getConfig(self):
config = {}
config['config'] = self.config.copy()
config['TraceX'] = {}
config['TraceX'] = self.x.getConfig()
config['TraceY'] = {}
for k,v in self.traces.items():
config['TraceY'][k] = v.getConfig()
return( config )
def getData(self):
data=self.x.values
if data.ndim == 1:
data = data[:, np.newaxis]
for k,v in self.traces.items():
vals = v.values
if vals.ndim == 1:
vals = vals[:, np.newaxis]
data=np.concatenate((data, vals), 1)
return( data )
def addPointToTrace(self, val, name=None):
if name is None:
self.x.values = np.append(self.x.values, val)
else:
a = self.traces[name].values
a = np.append(a, val)
self.traces[name].values = a
if __name__ == '__main__':
print("Testing trace")
x=Trace('x trace')
x.values = np.random.normal(2,2,(4,1))
x.values = np.array(x.values, int)
x.config['unit']='s'
x.config['tags']['tag1'] = 'xxxx'
x.config['tags']['tag2'] = 'xxxx'
x.save('xtrace.dat', skip_headers_if_file_exist=True)
# print(x.getHeader())
y=Trace('y trace')
y.values = np.random.normal(2,2,(4,1))
y.config['unit']='V'
y.config['tags']['ytag2'] = 'yyyy'
xy=TraceXY('xy trace')
xy.config['tags']['xy tag']= 'I am xy tag'
xy.x = x
xy.y = y
xy.save('xytrace.dat')
# print(xy.getHeader())
xyn = TraceSetSameX('many ys trace')
xyn.config['tags']['descr'] = 'I am many ys trace'
xy.config['label']='y1'
xyn.addTrace(xy)
xy.config['label']='y2'
xyn.addTrace(xy)
xy.config['label']='y3'
xyn.addTrace(xy)
xyn.save('xyntrace.dat')
# print(xyn.getHeader())
|