Fix : coverage

Author: Sabrina-Hassaim

CONTRIBUTING.md

@@ -1 +1 @@
-mkdocs/devguide.md
+mkdocs/devguide.md

README.md

@@ -1 +1 @@
-mkdocs/index.md
+mkdocs/index.md

janitor/io.py

@@ -866,6 +866,7 @@ def _infer_file_type(file_path: str) -> str:
     Raises:
         ValueError if the file extension is unsupported.
     """
+    file_path = file_path.lower()
     if file_path.endswith(".zip"):
         return "zip"
     elif file_path.endswith((".tar", ".tar.gz")):

mkdocs/AUTHORS.md

@@ -1 +1 @@
-../AUTHORS.md
+../AUTHORS.md

mkdocs/CHANGELOG.md

@@ -1 +1 @@
-../CHANGELOG.md
+../CHANGELOG.md

tests/io/test_read_archive.py

@@ -1,27 +1,32 @@
 import io
 import tarfile
 import zipfile
+from unittest.mock import patch
 
 import pandas as pd
 import pytest
 
-from janitor.io import read_archive
+from janitor.io import (
+    _infer_file_type,
+    read_archive,
+)
 
 
+# Fixtures for creating test archives
 @pytest.fixture
-def zip_test_file(tmp_path):
-    """Fixture pour créer un fichier ZIP de test."""
-    zip_path = tmp_path / "test.zip"
+def dummy_zip_file(tmp_path):
+    """Create a dummy ZIP file containing two CSV files."""
+    zip_path = tmp_path / "dummy.zip"
     with zipfile.ZipFile(zip_path, mode="w") as zf:
         zf.writestr("file1.csv", "col1,col2\n1,2\n3,4")
         zf.writestr("file2.csv", "col3,col4\n5,6\n7,8")
     return zip_path
 
 
 @pytest.fixture
-def tar_test_file(tmp_path):
-    """Fixture pour créer un fichier TAR de test."""
-    tar_path = tmp_path / "test.tar.gz"
+def dummy_tar_file(tmp_path):
+    """Create a dummy TAR file containing two CSV files."""
+    tar_path = tmp_path / "dummy.tar.gz"
     with tarfile.open(tar_path, mode="w:gz") as tf:
         info1 = tarfile.TarInfo(name="file1.csv")
         data1 = io.BytesIO(b"col1,col2\n1,2\n3,4")
@@ -35,23 +40,27 @@ def tar_test_file(tmp_path):
     return tar_path
 
 
-def test_read_zip_archive(zip_test_file):
+# Tests for reading archives via `read_archive`
+def test_read_zip_archive(dummy_zip_file):
+    """Test reading a specific file from a ZIP archive."""
     result = read_archive(
-        str(zip_test_file), extract_to_df=True, selected_files=["file1.csv"]
+        str(dummy_zip_file), extract_to_df=True, selected_files=["file1.csv"]
     )
     assert isinstance(result, pd.DataFrame)
     assert list(result.columns) == ["col1", "col2"]
     assert result.shape == (2, 2)
 
 
-def test_list_files_in_zip(zip_test_file):
-    result = read_archive(str(zip_test_file), extract_to_df=False)
+def test_list_files_in_zip(dummy_zip_file):
+    """Test listing files in a ZIP archive."""
+    result = read_archive(str(dummy_zip_file), extract_to_df=False)
     assert isinstance(result, list)
     assert "file1.csv" in result
     assert "file2.csv" in result
 
 
-def test_no_compatible_files(tmp_path):
+def test_no_compatible_files_in_zip(tmp_path):
+    """Test handling a ZIP archive with no compatible files."""
     zip_path = tmp_path / "empty.zip"
     with zipfile.ZipFile(zip_path, mode="w") as zf:
         zf.writestr("file1.txt", "Just some text")
@@ -61,17 +70,82 @@ def test_no_compatible_files(tmp_path):
         read_archive(str(zip_path))
 
 
-def test_read_tar_archive(tar_test_file):
+def test_read_tar_archive(dummy_tar_file):
+    """Test reading a specific file from a TAR archive."""
     result = read_archive(
-        str(tar_test_file), extract_to_df=True, selected_files=["file1.csv"]
+        str(dummy_tar_file), extract_to_df=True, selected_files=["file1.csv"]
     )
     assert isinstance(result, pd.DataFrame)
     assert list(result.columns) == ["col1", "col2"]
     assert result.shape == (2, 2)
 
 
-def test_list_files_in_tar(tar_test_file):
-    result = read_archive(str(tar_test_file), extract_to_df=False)
+def test_list_files_in_tar(dummy_tar_file):
+    """Test listing files in a TAR archive."""
+    result = read_archive(str(dummy_tar_file), extract_to_df=False)
     assert isinstance(result, list)
     assert "file1.csv" in result
     assert "file2.csv" in result
+
+
+def test_no_compatible_files_in_tar(tmp_path):
+    """Test handling a TAR archive with no compatible files."""
+    tar_path = tmp_path / "invalid.tar.gz"
+    with tarfile.open(tar_path, mode="w:gz") as tf:
+        info = tarfile.TarInfo(name="file1.txt")
+        data = io.BytesIO(b"Just some text")
+        info.size = data.getbuffer().nbytes
+        tf.addfile(info, data)
+    with pytest.raises(
+        ValueError, match="No compatible files found in the archive"
+    ):
+        read_archive(str(tar_path))
+
+
+# Tests for unsupported file types
+def test_read_archive_unsupported_file():
+    """Test handling unsupported file types."""
+    with pytest.raises(
+        ValueError,
+        match="Cannot infer file type from the file extension. "
+        "Please specify the 'file_type' parameter.",
+    ):
+        read_archive("test.unsupported")
+
+
+def test_read_archive_no_extension():
+    """Test handling files with no extension."""
+    with pytest.raises(
+        ValueError,
+        match="Cannot infer file type from the file extension. "
+        "Please specify the 'file_type' parameter.",
+    ):
+        read_archive("testfile")
+
+
+# Tests for interactive file selection
+def test_interactive_file_selection_valid(dummy_zip_file):
+    """Test valid input for interactive file selection."""
+    user_input = "1,2"
+    with patch("builtins.input", return_value=user_input):
+        result = read_archive(str(dummy_zip_file), extract_to_df=False)
+        assert "file1.csv" in result
+        assert "file2.csv" in result
+
+
+# Tests for file type inference
+def test_infer_file_type_valid():
+    """Test valid file type inference."""
+    assert _infer_file_type("test.zip") == "zip"
+    assert _infer_file_type("test.tar") == "tar"
+    assert _infer_file_type("test.tar.gz") == "tar.gz"
+
+
+def test_infer_file_type_invalid():
+    """Test invalid file type inference."""
+    with pytest.raises(
+        ValueError,
+        match="Cannot infer file type from the file extension. "
+        "Please specify the 'file_type' parameter.",
+    ):
+        _infer_file_type("testfile")