From d052dd121f92dad56b3c7123b596fb56c42673fc Mon Sep 17 00:00:00 2001
From: ravi_kumar_pilla <ravi_kumar_pilla@mckinsey.com>
Date: Tue, 11 Feb 2025 22:00:37 -0600
Subject: [PATCH] ignore files which are hidden and in gitignore

Signed-off-by: ravi_kumar_pilla <ravi_kumar_pilla@mckinsey.com>
---
 package/kedro_viz/autoreload_file_filter.py   | 13 +++-------
 .../integrations/kedro/lite_parser.py         | 10 ++++++-
 package/kedro_viz/utils.py                    | 26 ++++++++++++++++++-
 3 files changed, 37 insertions(+), 12 deletions(-)

diff --git a/package/kedro_viz/autoreload_file_filter.py b/package/kedro_viz/autoreload_file_filter.py
index f8b13c6237..628fb931b7 100644
--- a/package/kedro_viz/autoreload_file_filter.py
+++ b/package/kedro_viz/autoreload_file_filter.py
@@ -7,6 +7,7 @@
 from pathlib import Path
 from typing import Optional, Set
 
+from kedro_viz.utils import load_gitignore_patterns
 from pathspec import GitIgnoreSpec
 from watchfiles import Change, DefaultFilter
 
@@ -35,16 +36,8 @@ def __init__(self, base_path: Optional[Path] = None):
         super().__init__()
 
         # Load .gitignore patterns
-        gitignore_path = self.cwd / ".gitignore"
-        try:
-            with open(gitignore_path, "r", encoding="utf-8") as gitignore_file:
-                ignore_patterns = gitignore_file.read().splitlines()
-            self.gitignore_spec: Optional[GitIgnoreSpec] = GitIgnoreSpec.from_lines(
-                "gitwildmatch", ignore_patterns
-            )
-        except FileNotFoundError:
-            self.gitignore_spec = None
-
+        self.gitignore_spec = load_gitignore_patterns(self.cwd)
+    
     def __call__(self, change: Change, path: str) -> bool:
         """
         Determine whether a file change should be processed.
diff --git a/package/kedro_viz/integrations/kedro/lite_parser.py b/package/kedro_viz/integrations/kedro/lite_parser.py
index f6a7dbeb8d..a1299dda68 100755
--- a/package/kedro_viz/integrations/kedro/lite_parser.py
+++ b/package/kedro_viz/integrations/kedro/lite_parser.py
@@ -8,6 +8,7 @@
 from unittest.mock import MagicMock
 
 from kedro_viz.integrations.utils import Spinner
+from kedro_viz.utils import is_file_ignored, load_gitignore_patterns
 
 logger = logging.getLogger(__name__)
 
@@ -243,7 +244,14 @@ def parse(self, target_path: Path) -> Union[Dict[str, Set[str]], None]:
 
         unresolved_imports: Dict[str, Set[str]] = {}
 
+        # Load .gitignore patterns
+        gitignore_spec = load_gitignore_patterns(target_path)
+
         if target_path.is_file():
+
+            if is_file_ignored(target_path):
+                return unresolved_imports
+            
             try:
                 missing_dependencies = self._get_unresolved_imports(target_path)
                 if len(missing_dependencies) > 0:
@@ -258,7 +266,7 @@ def parse(self, target_path: Path) -> Union[Dict[str, Set[str]], None]:
             return unresolved_imports
 
         # handling directories
-        _project_file_paths = set(target_path.rglob("*.py"))
+        _project_file_paths = set(file_path for file_path in target_path.rglob("*.py") if not is_file_ignored(file_path, target_path, gitignore_spec))
 
         for file_path in _project_file_paths:
             try:
diff --git a/package/kedro_viz/utils.py b/package/kedro_viz/utils.py
index a0a4a5abce..4033886031 100644
--- a/package/kedro_viz/utils.py
+++ b/package/kedro_viz/utils.py
@@ -1,7 +1,10 @@
 """Transcoding related utility functions."""
 
 import hashlib
-from typing import Tuple
+from pathlib import Path
+from typing import Optional, Tuple
+
+from pathspec import GitIgnoreSpec
 
 TRANSCODING_SEPARATOR = "@"
 
@@ -57,3 +60,24 @@ def _strip_transcoding(element: str) -> str:
 def is_dataset_param(dataset_name: str) -> bool:
     """Return whether a dataset is a parameter"""
     return dataset_name.lower().startswith("params:") or dataset_name == "parameters"
+
+def load_gitignore_patterns(project_path: Path) -> Optional[GitIgnoreSpec]:
+    gitignore_path = project_path / ".gitignore"
+
+    if not gitignore_path.exists():
+        return
+    
+    with open(gitignore_path, "r", encoding="utf-8") as gitignore_file:
+        ignore_patterns = gitignore_file.read().splitlines()
+        gitignore_spec = GitIgnoreSpec.from_lines(
+            "gitwildmatch", ignore_patterns
+        )
+        return gitignore_spec
+
+def is_file_ignored(file_path: Path, project_path: Optional[Path], gitignore_spec: Optional[GitIgnoreSpec]) -> bool:
+    """Returns True if the file should be ignored."""
+    if file_path.name.startswith("."):  # Ignore hidden files/folders
+        return True
+    if gitignore_spec and project_path and gitignore_spec.match_file(str(file_path.relative_to(project_path))):
+        return True
+    return False