Add support for ip git repo urls (#827)

* Add support for ip v4 git repo urls
* Add tests for git urls

.travis.yml

@@ -12,5 +12,5 @@ before_script:
   - docker-compose -f docker-compose.yml -f docker-compose.ci.yml up --build -d
 
 script:
-  - docker exec -it cvat /bin/bash -c 'python3 manage.py test cvat/apps/engine utils/cli'
+  - docker exec -it cvat /bin/bash -c 'python3 manage.py test cvat/apps utils/cli'
   - docker exec -it cvat /bin/bash -c 'cd cvat-core && npm install && npm run test && npm run coveralls'

.vscode/launch.json

@@ -134,7 +134,7 @@
                 "test",
                 "--settings",
                 "cvat.settings.testing",
-                "cvat/apps/engine",
+                "cvat/apps",
                 "utils/cli"
             ],
             "django": true,

CHANGELOG.md

@@ -25,6 +25,7 @@ https://github.com/opencv/cvat/issues/750).
 
 ### Fixed
 - [Mask problem on coco json style](https://github.com/opencv/cvat/issues/718)
+- [Exception in Git plugin](https://github.com/opencv/cvat/issues/826)
 
 ### Security
 -

cvat/apps/git/git.py

@@ -76,8 +76,13 @@ class Git:
     # HTTP/HTTPS: [http://]github.com/proj/repos[.git]
     def _parse_url(self):
         try:
-            http_pattern = "([https|http]+)*[://]*([a-zA-Z0-9._-]+.[a-zA-Z]+)/([a-zA-Z0-9._-]+)/([a-zA-Z0-9._-]+)"
-            ssh_pattern = "([a-zA-Z0-9._-]+)@([a-zA-Z0-9._-]+):([a-zA-Z0-9._-]+)/([a-zA-Z0-9._-]+)"
+            # Almost STD66 (RFC3986), but schema can include a leading digit
+            # Reference on URL formats accepted by Git:
+            # https://github.com/git/git/blob/77bd3ea9f54f1584147b594abc04c26ca516d987/url.c
+
+            host_pattern = r"((?:(?:(?:\d{1,3}\.){3}\d{1,3})|(?:[a-zA-Z0-9._-]+.[a-zA-Z]+))(?::\d+)?)"
+            http_pattern = r"(?:http[s]?://)?" + host_pattern + r"((?:/[a-zA-Z0-9._-]+){2})"
+            ssh_pattern = r"([a-zA-Z0-9._-]+)@" + host_pattern + r":([a-zA-Z0-9._-]+)/([a-zA-Z0-9._-]+)"
 
             http_match = re.match(http_pattern, self._url)
             ssh_match = re.match(ssh_pattern, self._url)
@@ -87,21 +92,21 @@ class Git:
             repos = None
 
             if http_match:
-                host = http_match.group(2)
-                repos = "{}/{}".format(http_match.group(3), http_match.group(4))
+                host = http_match.group(1)
+                repos = http_match.group(2)[1:]
             elif ssh_match:
                 user = ssh_match.group(1)
                 host = ssh_match.group(2)
                 repos = "{}/{}".format(ssh_match.group(3), ssh_match.group(4))
             else:
-                raise Exception("Got URL doesn't sutisfy for regular expression")
+                raise Exception("Git repository URL does not satisfy pattern")
 
             if not repos.endswith(".git"):
                 repos += ".git"
 
             return user, host, repos
         except Exception as ex:
-            slogger.glob.exception('URL parsing errors occured', exc_info = True)
+            slogger.glob.exception('URL parsing errors occurred', exc_info = True)
             raise ex
 
 

cvat/apps/git/tests.py

@@ -2,6 +2,58 @@
 #
 # SPDX-License-Identifier: MIT
 
+from itertools import product
+
 from django.test import TestCase
 
 # Create your tests here.
+
+from cvat.apps.git.git import Git
+
+
+class GitUrlTest(TestCase):
+    class FakeGit:
+        def __init__(self, url):
+            self._url = url
+
+    def _check_correct_urls(self, samples):
+        for i, (expected, url) in enumerate(samples):
+            git = GitUrlTest.FakeGit(url)
+            try:
+                actual = Git._parse_url(git)
+                self.assertEqual(expected, actual, "URL #%s: '%s'" % (i, url))
+            except Exception:
+                self.assertFalse(True, "URL #%s: '%s'" % (i, url))
+
+    def test_correct_urls_can_be_parsed(self):
+        hosts = ['host.zone', '1.2.3.4']
+        ports = ['', ':42']
+        repo_groups = ['repo', 'r4p0']
+        repo_repos = ['nkjl23', 'hewj']
+        git_suffixes = ['', '.git']
+
+        samples = []
+
+        # http samples
+        protocols = ['', 'http://', 'https://']
+        for protocol, host, port, repo_group, repo, git in product(
+                protocols, hosts, ports, repo_groups, repo_repos, git_suffixes):
+            url = '{protocol}{host}{port}/{repo_group}/{repo}{git}'.format(
+                protocol=protocol, host=host, port=port,
+                repo_group=repo_group, repo=repo, git=git
+            )
+            expected = ('git', host + port, '%s/%s.git' % (repo_group, repo))
+            samples.append((expected, url))
+
+        # git samples
+        users = ['user', 'u123_.']
+        for user, host, port, repo_group, repo, git in product(
+                users, hosts, ports, repo_groups, repo_repos, git_suffixes):
+            url = '{user}@{host}{port}:{repo_group}/{repo}{git}'.format(
+                user=user, host=host, port=port,
+                repo_group=repo_group, repo=repo, git=git
+            )
+            expected = (user, host + port, '%s/%s.git' % (repo_group, repo))
+            samples.append((expected, url))
+
+        self._check_correct_urls(samples)
\ No newline at end of file