mirror of
https://github.com/cyclotruc/gitingest.git
synced 2026-04-28 11:09:30 +00:00
e5fadce158
• Accept hosts starting with “git.” or “gitlab.” in _looks_like_git_host • Update doc-strings to document the heuristic • Adjust git-host-agnostic tests: expect ValueError for slug form with custom hosts; add real GitLab instance (git.rwth-aachen.de) to matrix
74 lines
2.4 KiB
Python
74 lines
2.4 KiB
Python
"""
|
|
Tests to verify that the query parser is Git host agnostic.
|
|
|
|
These tests confirm that `parse_query` correctly identifies user/repo pairs and canonical URLs for GitHub, GitLab,
|
|
Bitbucket, Gitea, and Codeberg, even if the host is omitted.
|
|
"""
|
|
|
|
from typing import List, Tuple
|
|
|
|
import pytest
|
|
|
|
from gitingest.query_parsing import parse_query
|
|
from gitingest.utils.query_parser_utils import KNOWN_GIT_HOSTS
|
|
|
|
# Repository matrix: (host, user, repo)
|
|
_REPOS: List[Tuple[str, str, str]] = [
|
|
("github.com", "tiangolo", "fastapi"),
|
|
("gitlab.com", "gitlab-org", "gitlab-runner"),
|
|
("bitbucket.org", "na-dna", "llm-knowledge-share"),
|
|
("gitea.com", "xorm", "xorm"),
|
|
("codeberg.org", "forgejo", "forgejo"),
|
|
("git.rwth-aachen.de", "medialab", "19squared"),
|
|
("gitlab.alpinelinux.org", "alpine", "apk-tools"),
|
|
]
|
|
|
|
|
|
# Generate cartesian product of repository tuples with URL variants.
|
|
@pytest.mark.parametrize("host, user, repo", _REPOS, ids=[f"{h}:{u}/{r}" for h, u, r in _REPOS])
|
|
@pytest.mark.parametrize("variant", ["full", "noscheme", "slug"])
|
|
@pytest.mark.asyncio
|
|
async def test_parse_query_without_host(
|
|
host: str,
|
|
user: str,
|
|
repo: str,
|
|
variant: str,
|
|
) -> None:
|
|
"""Verify that `parse_query` handles URLs, host-omitted URLs and raw slugs."""
|
|
|
|
# Build the input URL based on the selected variant
|
|
if variant == "full":
|
|
url = f"https://{host}/{user}/{repo}"
|
|
elif variant == "noscheme":
|
|
url = f"{host}/{user}/{repo}"
|
|
else: # "slug"
|
|
url = f"{user}/{repo}"
|
|
|
|
expected_url = f"https://{host}/{user}/{repo}"
|
|
|
|
# For slug form with a custom host (not in KNOWN_GIT_HOSTS) we expect a failure,
|
|
# because the parser cannot guess which domain to use.
|
|
if variant == "slug" and host not in KNOWN_GIT_HOSTS:
|
|
with pytest.raises(ValueError):
|
|
await parse_query(url, max_file_size=50, from_web=True)
|
|
return
|
|
|
|
query = await parse_query(url, max_file_size=50, from_web=True)
|
|
|
|
# Compare against the canonical dict while ignoring unpredictable fields.
|
|
actual = query.model_dump(exclude={"id", "local_path", "ignore_patterns"})
|
|
|
|
expected = {
|
|
"user_name": user,
|
|
"repo_name": repo,
|
|
"url": expected_url,
|
|
"slug": f"{user}-{repo}",
|
|
"subpath": "/",
|
|
"type": None,
|
|
"branch": None,
|
|
"commit": None,
|
|
"max_file_size": 50,
|
|
"include_patterns": None,
|
|
}
|
|
|
|
assert actual == expected
|