get_pr_ut.py 11.3 KB
Newer Older
C
chalsliu 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# 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.
""" For the PR that only modified the unit test, get cases in pull request. """

import os
import json
18 19
import re
import sys
20 21
import time
import subprocess
22
import requests
23 24
import urllib.request
import ssl
Y
YUNSHEN XIE 已提交
25
import platform
C
chalsliu 已提交
26 27 28
from github import Github

PADDLE_ROOT = os.getenv('PADDLE_ROOT', '/paddle/')
C
chalsliu 已提交
29 30
PADDLE_ROOT += '/'
PADDLE_ROOT = PADDLE_ROOT.replace('//', '/')
31
ssl._create_default_https_context = ssl._create_unverified_context
C
chalsliu 已提交
32 33 34 35 36 37 38 39


class PRChecker(object):
    """ PR Checker. """

    def __init__(self):
        self.github = Github(os.getenv('GITHUB_API_TOKEN'), timeout=60)
        self.repo = self.github.get_repo('PaddlePaddle/Paddle')
40
        self.py_prog_oneline = re.compile('\d+\|\s*#.*')
C
chalsliu 已提交
41 42
        self.py_prog_multiline_a = re.compile('\d+\|\s*r?""".*?"""', re.DOTALL)
        self.py_prog_multiline_b = re.compile("\d+\|\s*r?'''.*?'''", re.DOTALL)
43 44 45
        self.cc_prog_online = re.compile('\d+\|\s*//.*')
        self.cc_prog_multiline = re.compile('\d+\|\s*/\*.*?\*/', re.DOTALL)
        self.lineno_prog = re.compile('@@ \-\d+,\d+ \+(\d+),(\d+) @@')
C
chalsliu 已提交
46
        self.pr = None
47
        self.suffix = ''
C
chalsliu 已提交
48
        self.full_case = False
C
chalsliu 已提交
49 50 51 52 53

    def init(self):
        """ Get pull request. """
        pr_id = os.getenv('GIT_PR_ID')
        if not pr_id:
54
            print('PREC No PR ID')
C
chalsliu 已提交
55
            exit(0)
56 57 58
        suffix = os.getenv('PREC_SUFFIX')
        if suffix:
            self.suffix = suffix
C
chalsliu 已提交
59
        self.pr = self.repo.get_pull(int(pr_id))
C
chalsliu 已提交
60 61 62 63 64 65 66 67 68
        last_commit = None
        ix = 0
        while True:
            commits = self.pr.get_commits().get_page(ix)
            for c in commits:
                last_commit = c.commit
            else:
                break
            ix = ix + 1
69 70
        if last_commit.message.find('test=allcase') != -1:
            print('PREC test=allcase is set')
C
chalsliu 已提交
71
            self.full_case = True
C
chalsliu 已提交
72

73 74 75 76 77 78 79 80
    #todo: exception
    def __wget_with_retry(self, url):
        ix = 1
        proxy = '--no-proxy'
        while ix < 6:
            if ix // 2 == 0:
                proxy = ''
            else:
81 82 83 84
                if platform.system() == 'Windows':
                    proxy = '-Y off'
                else:
                    proxy = '--no-proxy'
85 86 87 88 89 90 91 92 93 94 95 96
            code = subprocess.call(
                'wget -q {} --no-check-certificate {}'.format(proxy, url),
                shell=True)
            if code == 0:
                return True
            print(
                'PREC download {} error, retry {} time(s) after {} secs.[proxy_option={}]'.
                format(url, ix, ix * 10, proxy))
            time.sleep(ix * 10)
            ix += 1
        return False

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
    def __urlretrieve(self, url, filename):
        ix = 1
        with_proxy = urllib.request.getproxies()
        without_proxy = {'http': '', 'http': ''}
        while ix < 6:
            if ix // 2 == 0:
                cur_proxy = urllib.request.ProxyHandler(without_proxy)
            else:
                cur_proxy = urllib.request.ProxyHandler(with_proxy)
            opener = urllib.request.build_opener(cur_proxy,
                                                 urllib.request.HTTPHandler)
            urllib.request.install_opener(opener)
            try:
                urllib.request.urlretrieve(url, filename)
            except Exception as e:
                print(e)
                print(
                    'PREC download {} error, retry {} time(s) after {} secs.[proxy_option={}]'.
                    format(url, ix, ix * 10, proxy))
                continue
            else:
                return True
            time.sleep(ix * 10)
            ix += 1

        return False

C
chalsliu 已提交
124 125 126 127 128 129 130 131 132 133 134 135 136
    def get_pr_files(self):
        """ Get files in pull request. """
        page = 0
        file_list = []
        while True:
            files = self.pr.get_files().get_page(page)
            if not files:
                break
            for f in files:
                file_list.append(PADDLE_ROOT + f.filename)
            page += 1
        return file_list

137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
    def __get_comment_by_filetype(self, content, filetype):
        result = []
        if filetype == 'py':
            result = self.__get_comment_by_prog(content, self.py_prog_oneline)
            result.extend(
                self.__get_comment_by_prog(content, self.py_prog_multiline_a))
            result.extend(
                self.__get_comment_by_prog(content, self.py_prog_multiline_b))
        if filetype == 'cc':
            result = self.__get_comment_by_prog(content, self.cc_prog_oneline)
            result.extend(
                self.__get_comment_by_prog(content, self.cc_prog_multiline))
        return result

    def __get_comment_by_prog(self, content, prog):
        result_list = prog.findall(content)
        if not result_list:
C
chalsliu 已提交
154 155
            return []
        result = []
156 157 158 159 160 161
        for u in result_list:
            result.extend(u.split('\n'))
        return result

    def get_comment_of_file(self, f):
        #content = self.repo.get_contents(f.replace(PADDLE_ROOT, ''), 'pull/').decoded_content
162
        #todo: get file from github
163 164 165 166 167 168 169 170 171 172 173 174 175 176 177
        with open(f) as fd:
            lines = fd.readlines()
        lineno = 1
        inputs = ''
        for line in lines:
            #for line in content.split('\n'):
            #input += str(lineno) + '|' + line + '\n'
            inputs += str(lineno) + '|' + line
            lineno += 1
        fietype = ''
        if f.endswith('.h') or f.endswith('.cc') or f.endswith('.cu'):
            filetype = 'cc'
        if f.endswith('.py'):
            filetype = 'py'
        else:
C
chalsliu 已提交
178
            return []
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
        return self.__get_comment_by_filetype(inputs, filetype)

    def get_pr_diff_lines(self):
        file_to_diff_lines = {}
        r = requests.get(self.pr.diff_url)
        data = r.text
        data = data.split('\n')
        ix = 0
        while ix < len(data):
            if data[ix].startswith('+++'):
                if data[ix].rstrip('\r\n') == '+++ /dev/null':
                    ix += 1
                    continue
                filename = data[ix][6:]
                ix += 1
                while ix < len(data):
                    result = self.lineno_prog.match(data[ix])
                    if not result:
                        break
                    lineno = int(result.group(1))
                    length = int(result.group(2))
                    ix += 1
                    end = ix + length
                    while ix < end:
                        if data[ix][0] == '-':
                            end += 1
                        if data[ix][0] == '+':
                            line_list = file_to_diff_lines.get(filename)
C
chalsliu 已提交
207 208
                            line = '{}{}'.format(lineno,
                                                 data[ix].replace('+', '|', 1))
209 210 211 212 213 214 215 216 217 218 219 220 221
                            if line_list:
                                line_list.append(line)
                            else:
                                file_to_diff_lines[filename] = [line, ]
                        if data[ix][0] != '-':
                            lineno += 1
                        ix += 1
            ix += 1
        return file_to_diff_lines

    def is_only_comment(self, f):
        file_to_diff_lines = self.get_pr_diff_lines()
        comment_lines = self.get_comment_of_file(f)
C
chalsliu 已提交
222 223 224
        diff_lines = file_to_diff_lines.get(f.replace(PADDLE_ROOT, '', 1))
        if not diff_lines:
            return False
225 226 227
        for l in diff_lines:
            if l not in comment_lines:
                return False
228
        print('PREC {} is only comment'.format(f))
229 230
        return True

C
chalsliu 已提交
231 232
    def get_pr_ut(self):
        """ Get unit tests in pull request. """
C
chalsliu 已提交
233 234
        if self.full_case:
            return ''
C
chalsliu 已提交
235
        check_added_ut = False
C
chalsliu 已提交
236 237
        ut_list = []
        file_ut_map = None
238
        ret = self.__urlretrieve(
239
            'https://sys-p0.bj.bcebos.com/prec/file_ut.json{}'.format(
240
                self.suffix), 'file_ut.json{}'.format(self.suffix))
241 242 243
        if not ret:
            print('PREC download file_ut.json failed')
            exit(1)
244
        with open('file_ut.json' + self.suffix) as jsonfile:
C
chalsliu 已提交
245 246
            file_ut_map = json.load(jsonfile)
        for f in self.get_pr_files():
Y
YUNSHEN XIE 已提交
247 248
            current_system = platform.system()
            if current_system == "Darwin" or current_system == "Windows":
249 250 251 252 253
                f_judge = f.replace(PADDLE_ROOT, '/paddle/', 1)
                f_judge = f_judge.replace('//', '/')
            else:
                f_judge = f
            if f_judge not in file_ut_map:
254 255 256 257 258 259
                if f.endswith('.md'):
                    ut_list.append('md_placeholder')
                elif f.endswith('.h') or f.endswith('.cu'):
                    if self.is_only_comment(f):
                        ut_list.append('h_cu_comment_placeholder')
                    else:
260 261 262
                        print(
                            'PREC dismatch: {} not in file ut map and not md or comment'.
                            format(f))
263
                        return ''
C
chalsliu 已提交
264 265
                elif f.endswith('.cc') or f.endswith('.py') or f.endswith(
                        '.cu'):
266
                    if f.find('test_') != -1 or f.find('_test') != -1:
267
                        print('PREC {} need check new ut'.format(f))
268
                        check_added_ut = True
269
                    elif self.is_only_comment(f):
C
chalsliu 已提交
270
                        ut_list.append('nomap_comment_placeholder')
271
                    else:
272 273 274
                        print(
                            'PREC dismatch: {} not in file ut map and not new ut or comment'.
                            format(f))
275
                        return ''
C
chalsliu 已提交
276
                else:
277
                    print('PREC dismatch: {} not in file ut map'.format(f))
C
chalsliu 已提交
278
                    return ''
C
chalsliu 已提交
279
            else:
280
                if self.is_only_comment(f):
C
chalsliu 已提交
281
                    ut_list.append('map_comment_placeholder')
282
                else:
283
                    ut_list.extend(file_ut_map.get(f_judge))
C
chalsliu 已提交
284
        ut_list = list(set(ut_list))
C
chalsliu 已提交
285

C
chalsliu 已提交
286
        if check_added_ut:
C
chalsliu 已提交
287 288
            with open('{}/added_ut'.format(PADDLE_ROOT)) as utfile:
                for ut in utfile:
289
                    print('PREC NEW UT: {}'.format(ut.rstrip('\r\n')))
C
chalsliu 已提交
290
                    ut_list.append(ut.rstrip('\r\n'))
C
chalsliu 已提交
291

C
chalsliu 已提交
292
        if ut_list:
293
            ret = self.__urlretrieve(
C
chalsliu 已提交
294
                'https://sys-p0.bj.bcebos.com/prec/prec_delta{}'.format(
295
                    self.suffix), 'prec_delta{}'.format(self.suffix))
C
chalsliu 已提交
296 297 298 299 300 301 302 303
            if ret:
                with open('prec_delta' + self.suffix) as delta:
                    for ut in delta:
                        ut_list.append(ut.rstrip('\r\n'))
            else:
                print('PREC download prec_delta failed')
                exit(1)

304
        return '\n'.join(ut_list)
C
chalsliu 已提交
305 306 307 308 309


if __name__ == '__main__':
    pr_checker = PRChecker()
    pr_checker.init()
310 311 312
    #print(pr_checker.get_pr_ut())
    with open('ut_list', 'w') as f:
        f.write(pr_checker.get_pr_ut())