feat(#34): support self-managed GitLab (#145)

* feat(#34): support self-hosted GitLab

* fix: apply suggestions from lgtm code review

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

* fix: syntax and add test

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Author: scastlara

README.md

@@ -143,7 +143,7 @@ There are three available severities for comments:
 
 lgtm aims to work with as many services as possible, and that includes remote repository providers. At the moment, lgtm supports:
 
-- [GitLab](https://gitlab.com) (only gitlab.com, not self-hosted)
+- [GitLab](https://gitlab.com) (both gitlab.com and [self-managed](https://about.gitlab.com/install/)).
 - [GitHub](https://github.com)
 
 lgtm will autodetect the url of the pull request passed as an argument.

src/lgtm_ai/__main__.py

@@ -3,6 +3,7 @@
 from collections.abc import Callable
 from importlib.metadata import version
 from typing import Any, assert_never, get_args
+from urllib.parse import urlparse
 
 import click
 import httpx
@@ -179,7 +180,12 @@ def review(target: PRUrl | LocalRepository, config: str | None, verbose: int, **
     formatter: Formatter[Any] = MarkDownFormatter(
         add_ranges_to_suggestions=git_source_supports_multiline_suggestions(target.source)
     )
-    git_client = get_git_client(source=target.source, token=resolved_config.git_api_key, formatter=formatter)
+    git_client = get_git_client(
+        source=target.source,
+        token=resolved_config.git_api_key,
+        formatter=formatter,
+        url=target.base_url if isinstance(target, PRUrl) else None,
+    )
     issues_client = _get_issues_client(resolved_config, git_client, formatter)
 
     code_reviewer = CodeReviewer(
@@ -235,7 +241,9 @@ def guide(
         config_file=config,
     ).resolve_config(target)
     agent_extra_settings = AgentSettings(retries=resolved_config.ai_retries)
-    git_client = get_git_client(source=target.source, token=resolved_config.git_api_key, formatter=MarkDownFormatter())
+    git_client = get_git_client(
+        source=target.source, token=resolved_config.git_api_key, formatter=MarkDownFormatter(), url=target.base_url
+    )
     review_guide = ReviewGuideGenerator(
         guide_agent=get_guide_agent_with_settings(agent_extra_settings),
         model=get_ai_model(
@@ -283,20 +291,23 @@ def _get_formatter_and_printer(output_format: OutputFormat) -> tuple[Formatter[A
 def _get_issues_client(
     resolved_config: ResolvedConfig, git_client: GitClient | None, formatter: Formatter[Any]
 ) -> IssuesClient | None:
-    """Select a different issues client for retrieving issues.
+    """Get the issues client based on the resolved configuration.
 
-    Will only return a different client if all of the following are true:
-        1) Be used at all
-        2) Be retrieved from a git platform and not elsewhere (e.g., Jira, Asana, etc.)
-        3) Have a specific API key configured
+    It can be a GitClient for GitHub/GitLab issues, or a Jira client.
+    If issues are not configured with a specific platform, it will fall back
+    to using the main `git_client`.
     """
     issues_client: IssuesClient | None = git_client
     if not resolved_config.issues_url or not resolved_config.issues_platform or not resolved_config.issues_regex:
         return issues_client
     if resolved_config.issues_platform.is_git_platform:
         if resolved_config.issues_api_key:
+            parsed_issues_url = urlparse(str(resolved_config.issues_url))
             issues_client = get_git_client(
-                source=resolved_config.issues_platform, token=resolved_config.issues_api_key, formatter=formatter
+                source=resolved_config.issues_platform,
+                token=resolved_config.issues_api_key,
+                formatter=formatter,
+                url=f"{parsed_issues_url.scheme}://{parsed_issues_url.netloc}",
             )
     elif resolved_config.issues_platform == IssuesPlatform.jira:
         if not resolved_config.issues_api_key or not resolved_config.issues_user:

src/lgtm_ai/base/schemas.py

@@ -26,6 +26,7 @@ def is_git_platform(self) -> bool:
 @dataclass(frozen=True, slots=True)
 class PRUrl:
     full_url: str
+    base_url: str
     repo_path: str
     pr_number: int
     source: PRSource

src/lgtm_ai/git_client/utils.py

@@ -7,13 +7,16 @@
 from lgtm_ai.git_client.gitlab import GitlabClient
 
 
-def get_git_client(source: PRSource | IssuesPlatform, token: str, formatter: Formatter[str]) -> GitClient | None:
+def get_git_client(
+    source: PRSource | IssuesPlatform, token: str, formatter: Formatter[str], url: str | None = None
+) -> GitClient | None:
     """Return a GitClient instance based on the provided PR URL."""
     git_client: GitClient
 
     if source == "gitlab":
-        git_client = GitlabClient(gitlab.Gitlab(private_token=token), formatter=formatter)
+        git_client = GitlabClient(gitlab.Gitlab(url=url, private_token=token), formatter=formatter)
     elif source == "github":
+        # TODO: Handle GitHub Enterprise with a custom URL
         git_client = GitHubClient(github.Github(login_or_token=token), formatter=formatter)
     elif source == "local":
         return None

src/lgtm_ai/validators.py

@@ -42,15 +42,6 @@ def parse_target(ctx: click.Context, param: str, value: object) -> PRUrl | Local
         )
 
     match parsed.netloc:
-        case AllowedLocations.Gitlab:
-            return _parse_pr_url(
-                parsed,
-                split_str="/-/merge_requests/",
-                source=PRSource.gitlab,
-                error_url_msg="The PR URL must be a merge request URL.",
-                error_num_msg="The PR URL must contain a valid MR number.",
-            )
-
         case AllowedLocations.Github:
             return _parse_pr_url(
                 parsed,
@@ -59,10 +50,22 @@ def parse_target(ctx: click.Context, param: str, value: object) -> PRUrl | Local
                 error_url_msg="The PR URL must be a pull request URL.",
                 error_num_msg="The PR URL must contain a valid PR number.",
             )
-
         case _:
-            raise click.BadParameter(
-                f"The PR URL host must be one of: {', '.join([s.value for s in AllowedLocations.__members__.values()])}"
+            # We support for GitLab cloud (.com) and self-hosted (custom domain)
+            # TODO: When we support more git providers with custom urls, we need to revisit this and
+            # probably add a `--git-platform` option to the CLI.
+            if "/-/merge_requests/" not in parsed.path:
+                raise click.BadParameter(
+                    f"The PR URL host '{parsed.netloc}' is not supported. "
+                    "lgtm-ai currently supports github.com and GitLab (cloud or self-hosted). "
+                    "GitLab merge request URLs must contain '/-/merge_requests/'."
+                )
+            return _parse_pr_url(
+                parsed,
+                split_str="/-/merge_requests/",
+                source=PRSource.gitlab,
+                error_url_msg="The PR URL must be a merge request URL.",
+                error_num_msg="The PR URL must contain a valid MR number.",
             )
 
 
@@ -135,6 +138,7 @@ def _parse_pr_url(
 
     return PRUrl(
         full_url=parsed.geturl(),
+        base_url=f"{parsed.scheme}://{parsed.netloc}",
         repo_path=project_path.strip("/"),
         pr_number=pr_num,
         source=source,

tests/config/test_handler.py

@@ -14,6 +14,7 @@
 
 target = PRUrl(
     full_url="https://gitlab.com/user/repo/-/merge_requests/1",
+    base_url="https://gitlab.com",
     repo_path="user/repo",
     pr_number=1,
     source=PRSource.gitlab,

tests/git_client/test_github.py

@@ -28,6 +28,7 @@
 MockGithubUrl = PRUrl(
     full_url="https://github.com/foo/bar/pull/1",
     repo_path="foo/bar",
+    base_url="https://github.com",
     pr_number=1,
     source=PRSource.github,
 )
@@ -234,7 +235,7 @@ def test_get_file_contents_single_file() -> None:
         mock.Mock(decoded_content=b"lorem ipsum dolor sit amet"),
     ]
     content = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
         file_path="important.py",
         branch_name="source",
     )
@@ -263,7 +264,7 @@ def test_get_file_contents_single_file_multiple_objects() -> None:
         ]
     ]
     content = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
         file_path="important.py",
         branch_name="source",
     )
@@ -303,12 +304,12 @@ def test_get_file_contents_one_file_missing() -> None:
         github.GithubException(status=404, message="Not Found"),  # target branch
     ]
     content_1 = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
         file_path="important.py",
         branch_name="source",
     )
     content_2 = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
         file_path="logic.py",
         branch_name="source",
     )
@@ -356,7 +357,7 @@ def test_post_review_successful() -> None:
     )
 
     client.publish_review(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
         fake_review,
     )
 
@@ -430,7 +431,7 @@ def test_post_review_fallback_to_single_line() -> None:
     ]
 
     client.publish_review(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
         fake_review,
     )
 
@@ -487,7 +488,8 @@ def test_publish_guide_successful() -> None:
     client = mock_github_client(m_repo)
 
     client.publish_guide(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github), FAKE_GUIDE
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
+        FAKE_GUIDE,
     )
 
     assert m_pr.create_review.call_args_list == [

tests/git_client/test_gitlab.py

@@ -24,6 +24,7 @@
 
 MockGitlabUrl = PRUrl(
     full_url="https://gitlab.com/foo/-/merge_requests/1",
+    base_url="https://gitlab.com",
     repo_path="foo",
     pr_number=1,
     source=PRSource.gitlab,
@@ -328,12 +329,12 @@ def test_get_file_contents_multiple_files() -> None:
     client = mock_gitlab_client(m_project)
 
     contents_1 = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
         file_path="important.py",
         branch_name="source",
     )
     contents_2 = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
         file_path="logic.py",
         branch_name="source",
     )
@@ -354,19 +355,19 @@ def test_get_file_contents_one_file_missing() -> None:
     client = mock_gitlab_client(m_project)
 
     missing_1 = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
         file_path="whatever",
         branch_name="source",
     )
 
     missing_2 = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
         file_path="whatever",
         branch_name="target",
     )
 
     contents = client.get_file_contents(
-        PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
+        PRUrl(full_url="https://foo", base_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
         file_path="logic.py",
         branch_name="source",
     )

tests/review/test_context.py

@@ -21,7 +21,11 @@ def test_retrieve_additional_context_from_git(self) -> None:
         )
 
         pr_url = PRUrl(
-            full_url="https://example.com/repo/pull/1", repo_path="repo", pr_number=1, source=PRSource.github
+            full_url="https://example.com/repo/pull/1",
+            base_url="https://example.com",
+            repo_path="repo",
+            pr_number=1,
+            source=PRSource.github,
         )
 
         additional_context = context_retriever.get_additional_context(
@@ -50,7 +54,11 @@ def test_retrieve_additional_context_from_url(self) -> None:
         )
 
         pr_url = PRUrl(
-            full_url="https://example.com/repo/pull/1", repo_path="repo", pr_number=1, source=PRSource.github
+            full_url="https://example.com/repo/pull/1",
+            base_url="https://example.com",
+            repo_path="repo",
+            pr_number=1,
+            source=PRSource.github,
         )
 
         additional_context = context_retriever.get_additional_context(
@@ -73,7 +81,11 @@ def test_retrieve_additional_context_from_config(self) -> None:
             git_client=MockGitClient(), issues_client=MockGitClient(), httpx_client=mock.Mock(spec=httpx.Client)
         )
         pr_url = PRUrl(
-            full_url="https://example.com/repo/pull/1", repo_path="repo", pr_number=1, source=PRSource.github
+            full_url="https://example.com/repo/pull/1",
+            base_url="https://example.com",
+            repo_path="repo",
+            pr_number=1,
+            source=PRSource.github,
         )
 
         additional_context = context_retriever.get_additional_context(
@@ -108,7 +120,13 @@ def test_get_context_multiple_files(self, client: GitClient) -> None:
         )
 
         context = context_retriever.get_code_context(
-            PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.github),
+            PRUrl(
+                full_url="https://foo",
+                base_url="https://example.com",
+                repo_path="path",
+                pr_number=1,
+                source=PRSource.github,
+            ),
             pr_diff=pr_diff,
         )
 
@@ -139,7 +157,13 @@ def test_get_context_one_file_missing(self, client: GitClient) -> None:
         )
 
         context = context_retriever.get_code_context(
-            PRUrl(full_url="https://foo", repo_path="path", pr_number=1, source=PRSource.gitlab),
+            PRUrl(
+                full_url="https://foo",
+                base_url="https://example.com",
+                repo_path="path",
+                pr_number=1,
+                source=PRSource.gitlab,
+            ),
             pr_diff=pr_diff,
         )
 

tests/review/test_guide.py

@@ -32,7 +32,7 @@ def test_get_guide_from_url_valid() -> None:
             config=ResolvedConfig(ai_api_key="", git_api_key=""),
         )
         guide = guide_generator.generate_review_guide(
-            pr_url=PRUrl(full_url="foo", repo_path="foo", pr_number=1, source=PRSource.gitlab)
+            pr_url=PRUrl(full_url="foo", base_url="foo", repo_path="foo", pr_number=1, source=PRSource.gitlab)
         )
 
     assert guide == ReviewGuide(