xref: /openbmc/openbmc/poky/scripts/lib/resulttool/resultutils.py (revision 8460358c3d24c71d9d38fd126c745854a6301564)
1# resulttool - common library/utility functions
2#
3# Copyright (c) 2019, Intel Corporation.
4# Copyright (c) 2019, Linux Foundation
5#
6# SPDX-License-Identifier: GPL-2.0-only
7#
8
9import os
10import base64
11import zlib
12import json
13import scriptpath
14import copy
15import urllib.request
16import posixpath
17import logging
18scriptpath.add_oe_lib_path()
19
20logger = logging.getLogger('resulttool')
21
22flatten_map = {
23    "oeselftest": [],
24    "runtime": [],
25    "sdk": [],
26    "sdkext": [],
27    "manual": []
28}
29regression_map = {
30    "oeselftest": ['TEST_TYPE', 'MACHINE'],
31    "runtime": ['TESTSERIES', 'TEST_TYPE', 'IMAGE_BASENAME', 'MACHINE', 'IMAGE_PKGTYPE', 'DISTRO'],
32    "sdk": ['TESTSERIES', 'TEST_TYPE', 'IMAGE_BASENAME', 'MACHINE', 'SDKMACHINE'],
33    "sdkext": ['TESTSERIES', 'TEST_TYPE', 'IMAGE_BASENAME', 'MACHINE', 'SDKMACHINE'],
34    "manual": ['TEST_TYPE', 'TEST_MODULE', 'IMAGE_BASENAME', 'MACHINE']
35}
36store_map = {
37    "oeselftest": ['TEST_TYPE', 'TESTSERIES', 'MACHINE'],
38    "runtime": ['TEST_TYPE', 'DISTRO', 'MACHINE', 'IMAGE_BASENAME'],
39    "sdk": ['TEST_TYPE', 'MACHINE', 'SDKMACHINE', 'IMAGE_BASENAME'],
40    "sdkext": ['TEST_TYPE', 'MACHINE', 'SDKMACHINE', 'IMAGE_BASENAME'],
41    "manual": ['TEST_TYPE', 'TEST_MODULE', 'MACHINE', 'IMAGE_BASENAME']
42}
43
44rawlog_sections = {
45    "ptestresult.rawlogs": "ptest",
46    "ltpresult.rawlogs": "ltp",
47    "ltpposixresult.rawlogs": "ltpposix"
48}
49
50def is_url(p):
51    """
52    Helper for determining if the given path is a URL
53    """
54    return p.startswith('http://') or p.startswith('https://')
55
56extra_configvars = {'TESTSERIES': ''}
57
58#
59# Load the json file and append the results data into the provided results dict
60#
61def append_resultsdata(results, f, configmap=store_map, configvars=extra_configvars):
62    if type(f) is str:
63        if is_url(f):
64            with urllib.request.urlopen(f) as response:
65                data = json.loads(response.read().decode('utf-8'))
66            url = urllib.parse.urlparse(f)
67            testseries = posixpath.basename(posixpath.dirname(url.path))
68        else:
69            with open(f, "r") as filedata:
70                try:
71                    data = json.load(filedata)
72                except json.decoder.JSONDecodeError:
73                    print("Cannot decode {}. Possible corruption. Skipping.".format(f))
74                    data = ""
75            testseries = os.path.basename(os.path.dirname(f))
76    else:
77        data = f
78    for res in data:
79        if "configuration" not in data[res] or "result" not in data[res]:
80            raise ValueError("Test results data without configuration or result section?")
81        for config in configvars:
82            if config == "TESTSERIES" and "TESTSERIES" not in data[res]["configuration"]:
83                data[res]["configuration"]["TESTSERIES"] = testseries
84                continue
85            if config not in data[res]["configuration"]:
86                data[res]["configuration"][config] = configvars[config]
87        testtype = data[res]["configuration"].get("TEST_TYPE")
88        if testtype not in configmap:
89            raise ValueError("Unknown test type %s" % testtype)
90        testpath = "/".join(data[res]["configuration"].get(i) for i in configmap[testtype])
91        if testpath not in results:
92            results[testpath] = {}
93        results[testpath][res] = data[res]
94
95#
96# Walk a directory and find/load results data
97# or load directly from a file
98#
99def load_resultsdata(source, configmap=store_map, configvars=extra_configvars):
100    results = {}
101    if is_url(source) or os.path.isfile(source):
102        append_resultsdata(results, source, configmap, configvars)
103        return results
104    for root, dirs, files in os.walk(source):
105        for name in files:
106            f = os.path.join(root, name)
107            if name == "testresults.json":
108                append_resultsdata(results, f, configmap, configvars)
109    return results
110
111def filter_resultsdata(results, resultid):
112    newresults = {}
113    for r in results:
114        for i in results[r]:
115            if i == resultsid:
116                 newresults[r] = {}
117                 newresults[r][i] = results[r][i]
118    return newresults
119
120def strip_logs(results):
121    newresults = copy.deepcopy(results)
122    for res in newresults:
123        if 'result' not in newresults[res]:
124            continue
125        for logtype in rawlog_sections:
126            if logtype in newresults[res]['result']:
127                del newresults[res]['result'][logtype]
128        if 'ptestresult.sections' in newresults[res]['result']:
129            for i in newresults[res]['result']['ptestresult.sections']:
130                if 'log' in newresults[res]['result']['ptestresult.sections'][i]:
131                    del newresults[res]['result']['ptestresult.sections'][i]['log']
132    return newresults
133
134# For timing numbers, crazy amounts of precision don't make sense and just confuse
135# the logs. For numbers over 1, trim to 3 decimal places, for numbers less than 1,
136# trim to 4 significant digits
137def trim_durations(results):
138    for res in results:
139        if 'result' not in results[res]:
140            continue
141        for entry in results[res]['result']:
142            if 'duration' in results[res]['result'][entry]:
143                duration = results[res]['result'][entry]['duration']
144                if duration > 1:
145                    results[res]['result'][entry]['duration'] = float("%.3f" % duration)
146                elif duration < 1:
147                    results[res]['result'][entry]['duration'] = float("%.4g" % duration)
148    return results
149
150def handle_cleanups(results):
151    # Remove pointless path duplication from old format reproducibility results
152    for res2 in results:
153        try:
154            section = results[res2]['result']['reproducible']['files']
155            for pkgtype in section:
156                for filelist in section[pkgtype].copy():
157                    if section[pkgtype][filelist] and type(section[pkgtype][filelist][0]) == dict:
158                        newlist = []
159                        for entry in section[pkgtype][filelist]:
160                            newlist.append(entry["reference"].split("/./")[1])
161                        section[pkgtype][filelist] = newlist
162
163        except KeyError:
164            pass
165    # Remove pointless duplicate rawlogs data
166    try:
167        del results[res2]['result']['reproducible.rawlogs']
168    except KeyError:
169        pass
170
171def decode_log(logdata):
172    if isinstance(logdata, str):
173        return logdata
174    elif isinstance(logdata, dict):
175        if "compressed" in logdata:
176            data = logdata.get("compressed")
177            data = base64.b64decode(data.encode("utf-8"))
178            data = zlib.decompress(data)
179            return data.decode("utf-8", errors='ignore')
180    return None
181
182def generic_get_log(sectionname, results, section):
183    if sectionname not in results:
184        return None
185    if section not in results[sectionname]:
186        return None
187
188    ptest = results[sectionname][section]
189    if 'log' not in ptest:
190        return None
191    return decode_log(ptest['log'])
192
193def ptestresult_get_log(results, section):
194    return generic_get_log('ptestresult.sections', results, section)
195
196def generic_get_rawlogs(sectname, results):
197    if sectname not in results:
198        return None
199    if 'log' not in results[sectname]:
200        return None
201    return decode_log(results[sectname]['log'])
202
203def save_resultsdata(results, destdir, fn="testresults.json", ptestjson=False, ptestlogs=False):
204    for res in results:
205        if res:
206            dst = destdir + "/" + res + "/" + fn
207        else:
208            dst = destdir + "/" + fn
209        os.makedirs(os.path.dirname(dst), exist_ok=True)
210        resultsout = results[res]
211        if not ptestjson:
212            resultsout = strip_logs(results[res])
213        trim_durations(resultsout)
214        handle_cleanups(resultsout)
215        with open(dst, 'w') as f:
216            f.write(json.dumps(resultsout, sort_keys=True, indent=1))
217        for res2 in results[res]:
218            if ptestlogs and 'result' in results[res][res2]:
219                seriesresults = results[res][res2]['result']
220                for logtype in rawlog_sections:
221                    logdata = generic_get_rawlogs(logtype, seriesresults)
222                    if logdata is not None:
223                        logger.info("Extracting " + rawlog_sections[logtype] + "-raw.log")
224                        with open(dst.replace(fn, rawlog_sections[logtype] + "-raw.log"), "w+") as f:
225                            f.write(logdata)
226                if 'ptestresult.sections' in seriesresults:
227                    for i in seriesresults['ptestresult.sections']:
228                        sectionlog = ptestresult_get_log(seriesresults, i)
229                        if sectionlog is not None:
230                            with open(dst.replace(fn, "ptest-%s.log" % i), "w+") as f:
231                                f.write(sectionlog)
232
233def git_get_result(repo, tags, configmap=store_map):
234    git_objs = []
235    for tag in tags:
236        files = repo.run_cmd(['ls-tree', "--name-only", "-r", tag]).splitlines()
237        git_objs.extend([tag + ':' + f for f in files if f.endswith("testresults.json")])
238
239    def parse_json_stream(data):
240        """Parse multiple concatenated JSON objects"""
241        objs = []
242        json_d = ""
243        for line in data.splitlines():
244            if line == '}{':
245                json_d += '}'
246                objs.append(json.loads(json_d))
247                json_d = '{'
248            else:
249                json_d += line
250        objs.append(json.loads(json_d))
251        return objs
252
253    # Optimize by reading all data with one git command
254    results = {}
255    for obj in parse_json_stream(repo.run_cmd(['show'] + git_objs + ['--'])):
256        append_resultsdata(results, obj, configmap=configmap)
257
258    return results
259
260def test_run_results(results):
261    """
262    Convenient generator function that iterates over all test runs that have a
263    result section.
264
265    Generates a tuple of:
266        (result json file path, test run name, test run (dict), test run "results" (dict))
267    for each test run that has a "result" section
268    """
269    for path in results:
270        for run_name, test_run in results[path].items():
271            if not 'result' in test_run:
272                continue
273            yield path, run_name, test_run, test_run['result']
274
275