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
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
|
# Copyright (c) 2021 Cisco and/or its affiliates.
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at:
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Convert output_info.xml files into JSON structures.
Version: 0.1.0
Date: 22nd June 2021
The json structure is defined in https://gerrit.fd.io/r/c/csit/+/28992
"""
import os
import re
import json
import logging
import gzip
from os.path import join
from shutil import rmtree
from copy import deepcopy
from json import loads
from pal_utils import get_files
class JSONData:
"""A Class storing and manipulating data from tests.
"""
def __init__(self, template=None):
"""Initialization.
:param template: JSON formatted template used to store data. It can
include default values.
:type template: dict
"""
self._template = deepcopy(template)
self._data = self._template if self._template else dict()
def __str__(self):
"""Return a string with human readable data.
:returns: Readable description.
:rtype: str
"""
return str(self._data)
def __repr__(self):
"""Return a string executable as Python constructor call.
:returns: Executable constructor call.
:rtype: str
"""
return f"JSONData(template={self._template!r})"
@property
def data(self):
"""Getter
:return: Data stored in the object.
:rtype: dict
"""
return self._data
def update(self, kwargs):
"""Update the data with new data from the dictionary.
:param kwargs: Key value pairs to be added to the data.
:type kwargs: dict
"""
self._data.update(kwargs)
def set_key(self, key, val):
"""Setter.
:param key: The key to be updated / added.
:param val: The key value.
:type key: str
:type val: object
"""
self._data[key] = deepcopy(val)
def add_to_list(self, key, val):
"""Add an item to the list identified by key.
:param key: The key identifying the list.
:param val: The val to be appended to the list. If val is a list,
extend is used.
"""
if self._data.get(key, None) is None:
self._data[key] = list()
if isinstance(val, list):
self._data[key].extend(val)
else:
self._data[key].append(val)
def dump(self, file_out, indent=None):
"""Write JSON data to a file.
:param file_out: Path to the output JSON file.
:param indent: Indentation of items in JSON string. It is directly
passed to json.dump method.
:type file_out: str
:type indent: str
"""
try:
with open(file_out, u"w") as file_handler:
json.dump(self._data, file_handler, indent=indent)
except OSError as err:
logging.warning(f"{repr(err)} Skipping")
def load(self, file_in):
"""Load JSON data from a file.
:param file_in: Path to the input JSON file.
:type file_in: str
:raises: ValueError if the data being deserialized is not a valid
JSON document.
:raises: IOError if the file is not found or corrupted.
"""
with open(file_in, u"r") as file_handler:
self._data = json.load(file_handler)
def _export_test_from_xml_to_json(tid, in_data, out, template, metadata):
"""Export data from a test to a json structure.
:param tid: Test ID.
:param in_data: Test data.
:param out: Path to output json file.
:param template: JSON template with optional default values.
:param metadata: Data which are not stored in XML structure.
:type tid: str
:type in_data: dict
:type out: str
:type template: dict
:type metadata: dict
"""
data = JSONData(template=template)
data.update(metadata)
data.set_key(u"test_id", tid)
t_type = in_data.get(u"type", u"")
t_type = u"NDRPDR" if t_type == u"CPS" else t_type # It is NDRPDR
data.set_key(u"test_type", t_type)
tags = in_data.get(u"tags", list())
data.set_key(u"tags", tags)
data.set_key(u"documentation", in_data.get(u"documentation", u""))
data.set_key(u"message", in_data.get(u"msg", u""))
data.set_key(u"start_time", in_data.get(u"starttime", u""))
data.set_key(u"end_time", in_data.get(u"endtime", u""))
data.set_key(u"status", in_data.get(u"status", u"FAILED"))
sut_type = u""
if u"vpp" in tid:
sut_type = u"vpp"
elif u"dpdk" in tid:
sut_type = u"dpdk"
data.set_key(u"sut_type", sut_type)
# Process configuration history:
in_papi = deepcopy(in_data.get(u"conf_history", None))
if in_papi:
regex_dut = re.compile(r'\*\*DUT(\d):\*\*')
node_id = u"dut1"
for line in in_papi.split(u"\n"):
if not line:
continue
groups = re.search(regex_dut, line)
if groups:
node_id = f"dut{groups.group(1)}"
else:
data.add_to_list(
u"log",
{
u"source_type": u"node",
u"source_id": node_id,
u"msg_type": u"papi",
u"log_level": u"INFO",
u"timestamp": in_data.get(u"starttime", u""),
u"msg": line,
u"data": list()
}
)
# Process show runtime:
if in_data.get(u"telemetry-show-run", None):
for item in in_data[u"telemetry-show-run"].values():
data.add_to_list(u"log", item.get(u"runtime", dict()))
else:
in_sh_run = deepcopy(in_data.get(u"show-run", None))
if in_sh_run:
# Transform to openMetrics format
for key, val in in_sh_run.items():
log_item = {
u"source_type": u"node",
u"source_id": key,
u"msg_type": u"metric",
u"log_level": u"INFO",
u"timestamp": in_data.get(u"starttime", u""),
u"msg": u"show_runtime",
u"data": list()
}
runtime = loads(val.get(u"runtime", list()))
for item in runtime:
for metric, m_data in item.items():
if metric == u"name":
continue
for idx, m_item in enumerate(m_data):
log_item[u"data"].append(
{
u"name": metric,
u"value": m_item,
u"labels": {
u"host": val.get(u"host", u""),
u"socket": val.get(u"socket", u""),
u"graph_node": item.get(u"name", u""),
u"thread_id": str(idx)
}
}
)
data.add_to_list(u"log", log_item)
# Process results:
results = dict()
if t_type == u"DEVICETEST":
pass # Nothing to add.
elif t_type == u"NDRPDR":
results = {
u"throughput": {
u"unit":
u"cps" if u"TCP_CPS" in tags or u"UDP_CPS" in tags
else u"pps",
u"ndr": {
u"value": {
u"lower": in_data.get(u"throughput", dict()).
get(u"NDR", dict()).get(u"LOWER", u"NaN"),
u"upper": in_data.get(u"throughput", dict()).
get(u"NDR", dict()).get(u"UPPER", u"NaN")
},
u"value_gbps": {
u"lower": in_data.get(u"gbps", dict()).
get(u"NDR", dict()).get(u"LOWER", u"NaN"),
u"upper": in_data.get(u"gbps", dict()).
get(u"NDR", dict()).get(u"UPPER", u"NaN")
}
},
u"pdr": {
u"value": {
u"lower": in_data.get(u"throughput", dict()).
get(u"PDR", dict()).get(u"LOWER", u"NaN"),
u"upper": in_data.get(u"throughput", dict()).
get(u"PDR", dict()).get(u"UPPER", u"NaN")
},
u"value_gbps": {
u"lower": in_data.get(u"gbps", dict()).
get(u"PDR", dict()).get(u"LOWER", u"NaN"),
u"upper": in_data.get(u"gbps", dict()).
get(u"PDR", dict()).get(u"UPPER", u"NaN")
}
}
},
u"latency": {
u"forward": {
u"pdr_90": in_data.get(u"latency", dict()).
get(u"PDR90", dict()).get(u"direction1", u"NaN"),
u"pdr_50": in_data.get(u"latency", dict()).
get(u"PDR50", dict()).get(u"direction1", u"NaN"),
u"pdr_10": in_data.get(u"latency", dict()).
get(u"PDR10", dict()).get(u"direction1", u"NaN"),
u"pdr_0": in_data.get(u"latency", dict()).
get(u"LAT0", dict()).get(u"direction1", u"NaN")
},
u"reverse": {
u"pdr_90": in_data.get(u"latency", dict()).
get(u"PDR90", dict()).get(u"direction2", u"NaN"),
u"pdr_50": in_data.get(u"latency", dict()).
get(u"PDR50", dict()).get(u"direction2", u"NaN"),
u"pdr_10": in_data.get(u"latency", dict()).
get(u"PDR10", dict()).get(u"direction2", u"NaN"),
u"pdr_0": in_data.get(u"latency", dict()).
get(u"LAT0", dict()).get(u"direction2", u"NaN")
}
}
}
elif t_type == "MRR":
results = {
u"unit": u"pps", # Old data use only pps
u"samples": in_data.get(u"result", dict()).get(u"samples", list()),
u"avg": in_data.get(u"result", dict()).get(u"receive-rate", u"NaN"),
u"stdev": in_data.get(u"result", dict()).
get(u"receive-stdev", u"NaN")
}
elif t_type == "SOAK":
results = {
u"critical_rate": {
u"lower": in_data.get(u"throughput", dict()).
get(u"LOWER", u"NaN"),
u"upper": in_data.get(u"throughput", dict()).
get(u"UPPER", u"NaN"),
}
}
elif t_type == "HOSTSTACK":
results = in_data.get(u"result", dict())
# elif t_type == "TCP": # Not used ???
# results = in_data.get(u"result", u"NaN")
elif t_type == "RECONF":
results = {
u"loss": in_data.get(u"result", dict()).get(u"loss", u"NaN"),
u"time": in_data.get(u"result", dict()).get(u"time", u"NaN")
}
else:
pass
data.set_key(u"results", results)
data.dump(out, indent=u" ")
def convert_xml_to_json(spec, data):
"""Convert downloaded XML files into JSON.
Procedure:
- create one json file for each test,
- gzip all json files one by one,
- delete json files.
:param spec: Specification read from the specification files.
:param data: Input data parsed from output.xml files.
:type spec: Specification
:type data: InputData
"""
logging.info(u"Converting downloaded XML files to JSON ...")
template_name = spec.output.get(u"use-template", None)
structure = spec.output.get(u"structure", u"tree")
if template_name:
with open(template_name, u"r") as file_handler:
template = json.load(file_handler)
else:
template = None
build_dir = spec.environment[u"paths"][u"DIR[BUILD,JSON]"]
try:
rmtree(build_dir)
except FileNotFoundError:
pass # It does not exist
os.mkdir(build_dir)
for job, builds in data.data.items():
logging.info(f" Processing job {job}")
if structure == "tree":
os.makedirs(join(build_dir, job), exist_ok=True)
for build_nr, build in builds.items():
logging.info(f" Processing build {build_nr}")
if structure == "tree":
os.makedirs(join(build_dir, job, build_nr), exist_ok=True)
for test_id, test_data in build[u"tests"].items():
groups = re.search(re.compile(r'-(\d+[tT](\d+[cC]))-'), test_id)
if groups:
test_id = test_id.replace(groups.group(1), groups.group(2))
logging.info(f" Processing test {test_id}")
if structure == "tree":
dirs = test_id.split(u".")[:-1]
name = test_id.split(u".")[-1]
os.makedirs(
join(build_dir, job, build_nr, *dirs), exist_ok=True
)
file_name = \
f"{join(build_dir, job, build_nr, *dirs, name)}.json"
else:
file_name = join(
build_dir,
u'.'.join((job, build_nr, test_id, u'json'))
)
suite_id = test_id.rsplit(u".", 1)[0].replace(u" ", u"_")
_export_test_from_xml_to_json(
test_id, test_data, file_name, template,
{
u"ci": u"jenkins.fd.io",
u"job": job,
u"build_number": build_nr,
u"suite_id": suite_id,
u"suite_doc": build[u"suites"].get(suite_id, dict()).
get(u"doc", u""),
u"testbed": build[u"metadata"].get(u"testbed", u""),
u"sut_version": build[u"metadata"].get(u"version", u"")
}
)
# gzip the json files:
for file in get_files(build_dir, u"json"):
with open(file, u"rb") as src:
with gzip.open(f"{file}.gz", u"wb") as dst:
dst.writelines(src)
os.remove(file)
logging.info(u"Done.")
|