Fixing lint suggestions

Author: gowthamkishore3799

src/gitingest/query_parser.py

@@ -4,6 +4,7 @@
 import re
 import string
 import uuid
+import warnings
 from pathlib import Path
 from typing import Any
 from urllib.parse import unquote, urlparse
@@ -172,36 +173,35 @@ async def _parse_repo_source(source: str) -> dict[str, Any]:
     if _is_valid_git_commit_hash(commit_or_branch):
         parsed["commit"] = commit_or_branch
         parsed["subpath"] += "/".join(remaining_parts[1:])
-
     else:
         parsed["branch"] = await _configure_branch_and_subpath(remaining_parts, url)
         parsed["subpath"] += "/".join(remaining_parts)
     return parsed
 
-async def _configure_branch_and_subpath(remaining_parts: list[str],url: str) -> str | None:
+
+async def _configure_branch_and_subpath(remaining_parts: list[str], url: str) -> str | None:
     """
-    Find the branch name from the remaining parts of the URL path.
+    Configure the branch and subpath based on the remaining parts of the URL.
     Parameters
     ----------
     remaining_parts : list[str]
-        List of path parts extracted from the URL.
+        The remaining parts of the URL path.
     url : str
-        The repository URL to determine branches.
-
+        The URL of the repository.
     Returns
     -------
-    str (branch name) or None
+    str | None
+        The branch name if found, otherwise None.
 
     """
     try:
         # Fetch the list of branches from the remote repository
         branches: list[str] = await fetch_remote_branch_list(url)
     except Exception as e:
-        print(f"Warning: Failed to fetch branch list: {str(e)}")
-        return remaining_parts.pop(0) if remaining_parts else None
+        warnings.warn(f"Warning: Failed to fetch branch list: {str(e)}")
+        return remaining_parts.pop(0)
 
     branch = []
-
     while remaining_parts:
         branch.append(remaining_parts.pop(0))
         branch_name = "/".join(branch)
@@ -210,6 +210,7 @@ async def _configure_branch_and_subpath(remaining_parts: list[str],url: str) ->
 
     return None
 
+
 def _is_valid_git_commit_hash(commit: str) -> bool:
     """
     Validate if the provided string is a valid Git commit hash.

src/gitingest/repository_clone.py

@@ -144,27 +144,25 @@ async def _check_repo_exists(url: str) -> bool:
 @async_timeout(TIMEOUT)
 async def fetch_remote_branch_list(url: str) -> list[str]:
     """
-    Get the list of branches from the remote repo.
-
+    Fetch the list of branches from a remote Git repository.
     Parameters
     ----------
     url : str
-       The URL of the repository.
-
+        The URL of the Git repository to fetch branches from.
     Returns
     -------
     list[str]
-      list of the branches in the remote repository
+        A list of branch names available in the remote repository.
     """
     fetch_branches_command = ["git", "ls-remote", "--heads", url]
-    stdout, stderr = await _run_git_command(*fetch_branches_command)
+    stdout, _ = await _run_git_command(*fetch_branches_command)
     stdout_decoded = stdout.decode()
 
     return [
-            line.split('refs/heads/', 1)[1]
-            for line in stdout_decoded.splitlines()
-            if line.strip() and 'refs/heads/' in line
-        ]
+        line.split("refs/heads/", 1)[1]
+        for line in stdout_decoded.splitlines()
+        if line.strip() and "refs/heads/" in line
+    ]
 
 
 async def _run_git_command(*args: str) -> tuple[bytes, bytes]:

tests/query_parser/test_query_parser.py

@@ -1,10 +1,9 @@
 """ Tests for the query_parser module. """
 
 from pathlib import Path
+from unittest.mock import AsyncMock, patch
 
 import pytest
-from unittest.mock import patch, AsyncMock
-from gitingest.repository_clone import _check_repo_exists, fetch_remote_branch_list
 
 from gitingest.ignore_patterns import DEFAULT_IGNORE_PATTERNS
 from gitingest.query_parser import _parse_patterns, _parse_repo_source, parse_query
@@ -98,15 +97,18 @@ async def test_parse_query_invalid_pattern() -> None:
     with pytest.raises(ValueError, match="Pattern.*contains invalid characters"):
         await parse_query(url, max_file_size=50, from_web=True, include_patterns="*.py;rm -rf")
 
+
 async def test_parse_url_with_subpaths() -> None:
     """
     Test `_parse_repo_source` with a URL containing a branch and subpath.
     Verifies that user name, repository name, branch, and subpath are correctly extracted.
     """
     url = "https://github.com/user/repo/tree/main/subdir/file"
-    with patch('gitingest.repository_clone._run_git_command', new_callable=AsyncMock) as mock_run_git_command:
+    with patch("gitingest.repository_clone._run_git_command", new_callable=AsyncMock) as mock_run_git_command:
         mock_run_git_command.return_value = (b"refs/heads/main\nrefs/heads/dev\nrefs/heads/feature-branch\n", b"")
-        with patch('gitingest.repository_clone.fetch_remote_branch_list', new_callable=AsyncMock) as mock_fetch_branches:
+        with patch(
+            "gitingest.repository_clone.fetch_remote_branch_list", new_callable=AsyncMock
+        ) as mock_fetch_branches:
             mock_fetch_branches.return_value = ["main", "dev", "feature-branch"]
             result = await _parse_repo_source(url)
             assert result["user_name"] == "user"
@@ -227,9 +229,11 @@ async def test_parse_url_branch_and_commit_distinction() -> None:
     url_branch = "https://github.com/user/repo/tree/main"
     url_commit = "https://github.com/user/repo/tree/abcd1234abcd1234abcd1234abcd1234abcd1234"
 
-    with patch('gitingest.repository_clone._run_git_command', new_callable=AsyncMock) as mock_run_git_command:
+    with patch("gitingest.repository_clone._run_git_command", new_callable=AsyncMock) as mock_run_git_command:
         mock_run_git_command.return_value = (b"refs/heads/main\nrefs/heads/dev\nrefs/heads/feature-branch\n", b"")
-        with patch('gitingest.repository_clone.fetch_remote_branch_list', new_callable=AsyncMock) as mock_fetch_branches:
+        with patch(
+            "gitingest.repository_clone.fetch_remote_branch_list", new_callable=AsyncMock
+        ) as mock_fetch_branches:
             mock_fetch_branches.return_value = ["main", "dev", "feature-branch"]
 
             result_branch = await _parse_repo_source(url_branch)
@@ -240,6 +244,7 @@ async def test_parse_url_branch_and_commit_distinction() -> None:
             assert result_commit["branch"] is None
             assert result_commit["commit"] == "abcd1234abcd1234abcd1234abcd1234abcd1234"
 
+
 async def test_parse_query_uuid_uniqueness() -> None:
     """
     Test `parse_query` to ensure that each call generates a unique UUID for the query result.
@@ -283,38 +288,52 @@ async def test_parse_query_with_branch() -> None:
     assert result["commit"] is None
     assert result["type"] == "blob"
 
+
 @pytest.mark.asyncio
-@pytest.mark.parametrize("url, expected_branch, expected_subpath", [
-    ("https://github.com/user/repo/tree/main/src", "main", "/src"),
-    ("https://github.com/user/repo/tree/fix1", "fix1", "/"),
-    ("https://github.com/user/repo/tree/nonexistent-branch/src", "nonexistent-branch", "/src"),
-])
+@pytest.mark.parametrize(
+    "url, expected_branch, expected_subpath",
+    [
+        ("https://github.com/user/repo/tree/main/src", "main", "/src"),
+        ("https://github.com/user/repo/tree/fix1", "fix1", "/"),
+        ("https://github.com/user/repo/tree/nonexistent-branch/src", "nonexistent-branch", "/src"),
+    ],
+)
 async def test_parse_repo_source_with_failed_git_command(url, expected_branch, expected_subpath):
     """
     Test `_parse_repo_source` when git command fails.
     Verifies that the function returns the first path component as the branch.
     """
-    with patch('gitingest.repository_clone.fetch_remote_branch_list', new_callable=AsyncMock) as mock_fetch_branches:
+    with patch("gitingest.repository_clone.fetch_remote_branch_list", new_callable=AsyncMock) as mock_fetch_branches:
         mock_fetch_branches.side_effect = Exception("Failed to fetch branch list")
-        
+
         result = await _parse_repo_source(url)
-        
+
         assert result["branch"] == expected_branch
         assert result["subpath"] == expected_subpath
 
+
 @pytest.mark.asyncio
-@pytest.mark.parametrize("url, expected_branch, expected_subpath", [
-    ("https://github.com/user/repo/tree/feature/fix1/src", "feature/fix1", "/src"),
-    ("https://github.com/user/repo/tree/main/src", "main", "/src"),
-    ("https://github.com/user/repo", None, "/"),  # No 
-    ("https://github.com/user/repo/tree/nonexistent-branch/src", None, "/"),  # Non-existent branch
-    ("https://github.com/user/repo/tree/fix", "fix", "/"),
-])
+@pytest.mark.parametrize(
+    "url, expected_branch, expected_subpath",
+    [
+        ("https://github.com/user/repo/tree/feature/fix1/src", "feature/fix1", "/src"),
+        ("https://github.com/user/repo/tree/main/src", "main", "/src"),
+        ("https://github.com/user/repo", None, "/"),  # No
+        ("https://github.com/user/repo/tree/nonexistent-branch/src", None, "/"),  # Non-existent branch
+        ("https://github.com/user/repo/tree/fix", "fix", "/"),
+        ("https://github.com/user/repo/blob/fix/page.html", "fix", "/page.html"),
+    ],
+)
 async def test_parse_repo_source_with_various_url_patterns(url, expected_branch, expected_subpath):
-    with patch('gitingest.repository_clone._run_git_command', new_callable=AsyncMock) as mock_run_git_command, \
-         patch('gitingest.repository_clone.fetch_remote_branch_list', new_callable=AsyncMock) as mock_fetch_branches:
-        
-        mock_run_git_command.return_value = (b"refs/heads/feature/fix1\nrefs/heads/main\nrefs/heads/feature-branch\nrefs/heads/fix\n", b"")
+    with (
+        patch("gitingest.repository_clone._run_git_command", new_callable=AsyncMock) as mock_run_git_command,
+        patch("gitingest.repository_clone.fetch_remote_branch_list", new_callable=AsyncMock) as mock_fetch_branches,
+    ):
+
+        mock_run_git_command.return_value = (
+            b"refs/heads/feature/fix1\nrefs/heads/main\nrefs/heads/feature-branch\nrefs/heads/fix\n",
+            b"",
+        )
         mock_fetch_branches.return_value = ["feature/fix1", "main", "feature-branch"]
 
         result = await _parse_repo_source(url)