New gitignore implementation based on pathspec

Refs #40

Author: simonw

files_to_prompt/cli.py

@@ -1,7 +1,8 @@
 import os
 import sys
 from fnmatch import fnmatch
-
+from files_to_prompt.utils import allowed_by_gitignore
+import pathlib
 import click
 
 global_index = 1
@@ -24,25 +25,6 @@
 }
 
 
-def should_ignore(path, gitignore_rules):
-    for rule in gitignore_rules:
-        if fnmatch(os.path.basename(path), rule):
-            return True
-        if os.path.isdir(path) and fnmatch(os.path.basename(path) + "/", rule):
-            return True
-    return False
-
-
-def read_gitignore(path):
-    gitignore_path = os.path.join(path, ".gitignore")
-    if os.path.isfile(gitignore_path):
-        with open(gitignore_path, "r") as f:
-            return [
-                line.strip() for line in f if line.strip() and not line.startswith("#")
-            ]
-    return []
-
-
 def add_line_numbers(content):
     lines = content.splitlines()
 
@@ -104,7 +86,6 @@ def process_path(
     include_hidden,
     ignore_files_only,
     ignore_gitignore,
-    gitignore_rules,
     ignore_patterns,
     writer,
     claude_xml,
@@ -124,17 +105,13 @@ def process_path(
                 dirs[:] = [d for d in dirs if not d.startswith(".")]
                 files = [f for f in files if not f.startswith(".")]
 
+            root_path = pathlib.Path(root)
             if not ignore_gitignore:
-                gitignore_rules.extend(read_gitignore(root))
                 dirs[:] = [
-                    d
-                    for d in dirs
-                    if not should_ignore(os.path.join(root, d), gitignore_rules)
+                    d for d in dirs if allowed_by_gitignore(root_path, root_path / d)
                 ]
                 files = [
-                    f
-                    for f in files
-                    if not should_ignore(os.path.join(root, f), gitignore_rules)
+                    f for f in files if allowed_by_gitignore(root_path, root_path / f)
                 ]
 
             if ignore_patterns:
@@ -302,7 +279,6 @@ def cli(
     # Combine paths from arguments and stdin
     paths = [*paths, *stdin_paths]
 
-    gitignore_rules = []
     writer = click.echo
     fp = None
     if output_file:
@@ -311,8 +287,6 @@ def cli(
     for path in paths:
         if not os.path.exists(path):
             raise click.BadArgumentUsage(f"Path does not exist: {path}")
-        if not ignore_gitignore:
-            gitignore_rules.extend(read_gitignore(os.path.dirname(path)))
         if claude_xml and path == paths[0]:
             writer("<documents>")
         process_path(
@@ -321,7 +295,6 @@ def cli(
             include_hidden,
             ignore_files_only,
             ignore_gitignore,
-            gitignore_rules,
             ignore_patterns,
             writer,
             claude_xml,

files_to_prompt/utils.py

@@ -0,0 +1,75 @@
+from pathlib import Path
+from pathspec.gitignore import GitIgnoreSpec
+
+
+def allowed_by_gitignore(root: Path, file_path: Path) -> bool:
+    """
+    Check whether the file (file_path) should be included (i.e. not ignored)
+    based on all .gitignore files encountered from the root directory down to
+    the directory where the file resides.
+
+    Parameters:
+      root (Path): The root directory under which .gitignore files are searched.
+      file_path (Path): The file to be checked.
+
+    Returns:
+      bool: True if the file should be included (not ignored); False if it should be ignored.
+    """
+    # Resolve absolute paths.
+    abs_root = root.resolve()
+    abs_file = file_path.resolve()
+
+    # Ensure file is under the provided root.
+    try:
+        _ = abs_file.relative_to(abs_root)
+    except ValueError:
+        raise ValueError(f"File {abs_file!r} is not under the root {abs_root!r}.")
+
+    # Build a list of directories from the root to the file's directory.
+    directories = [abs_root]
+    file_dir = abs_file.parent
+    rel_dir = file_dir.relative_to(abs_root)
+    for part in rel_dir.parts:
+        directories.append(directories[-1] / part)
+
+    # The decision will be updated by any matching .gitignore rule encountered.
+    decision = None
+
+    # Process each directory (from root to file's directory)
+    for directory in directories:
+        gitignore_file = directory / ".gitignore"
+        if gitignore_file.is_file():
+            try:
+                # Read nonempty lines (ignoring blank lines).
+                lines = [
+                    line.rstrip("\n")
+                    for line in gitignore_file.read_text(encoding="utf-8").splitlines()
+                    if line.strip()
+                ]
+            except Exception as e:
+                print(f"Could not read {gitignore_file}: {e}")
+                continue
+
+            # Compile a GitIgnoreSpec for the rules in the current directory.
+            spec = GitIgnoreSpec.from_lines(lines)
+
+            # .gitignore patterns are relative to the directory they are in.
+            # Compute the file path relative to this directory in POSIX format.
+            rel_file = abs_file.relative_to(directory).as_posix()
+
+            # Check the file against these rules.
+            result = spec.check_file(rel_file)
+
+            # If a rule from this .gitignore file applied, update the decision.
+            if result.include is not None:
+                decision = result.include
+
+    # If no .gitignore rule matched, the file is included by default.
+    if decision is None:
+        return True
+
+    # Interpretation:
+    #   • decision == True  --> a normal ignore rule matched (file should be ignored)
+    #   • decision == False --> a negation rule matched (file re-included)
+    # So, we return not decision.
+    return not decision

pyproject.toml

@@ -10,7 +10,8 @@ classifiers = [
     "License :: OSI Approved :: Apache Software License"
 ]
 dependencies = [
-    "click"
+    "click",
+    "pathspec",
 ]
 
 [project.urls]

tests/test_allowed_by_gitignore.py

@@ -0,0 +1,54 @@
+from files_to_prompt.utils import allowed_by_gitignore
+from pathlib import Path
+
+
+def test_allowed_by_gitignore(tmpdir):
+    # Create a temporary directory structure.
+    base = Path(tmpdir)
+    repo = base / "repo"
+    repo.mkdir()
+
+    # Create a top-level .gitignore in repo that ignores the "build/" directory.
+    (repo / ".gitignore").write_text("build/\n", encoding="utf-8")
+
+    # Create a "build" subdirectory and add an output file which should be ignored.
+    build_dir = repo / "build"
+    build_dir.mkdir()
+    output_file = build_dir / "output.txt"
+    output_file.write_text("dummy build output", encoding="utf-8")
+
+    # Create a "src" subdirectory with its own .gitignore.
+    src_dir = repo / "src"
+    src_dir.mkdir()
+    # In src, ignore "temp.txt"
+    (src_dir / ".gitignore").write_text("temp.txt\n", encoding="utf-8")
+
+    # Create files in "src"
+    main_file = src_dir / "main.py"
+    main_file.write_text("print('Hello')", encoding="utf-8")
+    temp_file = src_dir / "temp.txt"
+    temp_file.write_text("should be ignored", encoding="utf-8")
+    keep_file = src_dir / "keep.txt"
+    keep_file.write_text("keep this file", encoding="utf-8")
+
+    # Create a file at repo root that is not ignored.
+    root_file = repo / "README.md"
+    root_file.write_text("# Repo README", encoding="utf-8")
+
+    # Test cases:
+    # 1. File in "build" should be ignored.
+    assert (
+        allowed_by_gitignore(repo, output_file) is False
+    ), "build/output.txt should be ignored"
+
+    # 2. File in "src" that is ignored per src/.gitignore.
+    assert allowed_by_gitignore(repo, temp_file) is False, "src/temp.txt should be ignored"
+
+    # 3. Files in "src" not mentioned in .gitignore should be included.
+    assert allowed_by_gitignore(repo, main_file) is True, "src/main.py should be included"
+    assert allowed_by_gitignore(repo, keep_file) is True, "src/keep.txt should be included"
+
+    # 4. File at the repo root not mentioned in .gitignore.
+    assert (
+        allowed_by_gitignore(repo, root_file) is True
+    ), "repo/README.md should be included"

tests/test_files_to_prompt.py

@@ -21,7 +21,7 @@ def test_basic_functionality(tmpdir):
         with open("test_dir/file2.txt", "w") as f:
             f.write("Contents of file2")
 
-        result = runner.invoke(cli, ["test_dir"])
+        result = runner.invoke(cli, ["test_dir"], catch_exceptions=False)
         assert result.exit_code == 0
         assert "test_dir/file1.txt" in result.output
         assert "Contents of file1" in result.output
@@ -36,7 +36,7 @@ def test_include_hidden(tmpdir):
         with open("test_dir/.hidden.txt", "w") as f:
             f.write("Contents of hidden file")
 
-        result = runner.invoke(cli, ["test_dir"])
+        result = runner.invoke(cli, ["test_dir"], catch_exceptions=False)
         assert result.exit_code == 0
         assert "test_dir/.hidden.txt" not in result.output
 
@@ -61,11 +61,9 @@ def test_ignore_gitignore(tmpdir):
         with open("test_dir/nested_include/included2.txt", "w") as f:
             f.write("This nested file should be included")
         with open("test_dir/nested_ignore/.gitignore", "w") as f:
-            f.write("nested_ignore.txt")
+            f.write("*")
         with open("test_dir/nested_ignore/nested_ignore.txt", "w") as f:
             f.write("This nested file should not be included")
-        with open("test_dir/nested_ignore/actually_include.txt", "w") as f:
-            f.write("This nested file should actually be included")
 
         result = runner.invoke(cli, ["test_dir", "-c"])
         assert result.exit_code == 0
@@ -74,7 +72,6 @@ def test_ignore_gitignore(tmpdir):
         assert filenames == {
             "test_dir/included.txt",
             "test_dir/nested_include/included2.txt",
-            "test_dir/nested_ignore/actually_include.txt",
         }
 
         result2 = runner.invoke(cli, ["test_dir", "-c", "--ignore-gitignore"])
@@ -86,7 +83,6 @@ def test_ignore_gitignore(tmpdir):
             "test_dir/ignored.txt",
             "test_dir/nested_include/included2.txt",
             "test_dir/nested_ignore/nested_ignore.txt",
-            "test_dir/nested_ignore/actually_include.txt",
         }
 
 
@@ -243,7 +239,7 @@ def test_binary_file_warning(tmpdir):
         with open("test_dir/text_file.txt", "w") as f:
             f.write("This is a text file")
 
-        result = runner.invoke(cli, ["test_dir"])
+        result = runner.invoke(cli, ["test_dir"], catch_exceptions=False)
         assert result.exit_code == 0
 
         stdout = result.stdout
@@ -331,7 +327,7 @@ def test_line_numbers(tmpdir):
         with open("test_dir/multiline.txt", "w") as f:
             f.write(test_content)
 
-        result = runner.invoke(cli, ["test_dir"])
+        result = runner.invoke(cli, ["test_dir"], catch_exceptions=False)
         assert result.exit_code == 0
         assert "1  First line" not in result.output
         assert test_content in result.output