Update type hints using pyupgrade --py39-plus and --py310-plus

Author: filipchristiansen

.pre-commit-config.yaml

@@ -48,7 +48,7 @@ repos:
     hooks:
       - id: pyupgrade
         description: 'Automatically upgrade syntax for newer versions.'
-        args: [--py3-plus, --py36-plus, --py38-plus]
+        args: [--py3-plus, --py36-plus, --py38-plus, --py39-plus, --py310-plus]
 
   - repo: https://github.com/pre-commit/pygrep-hooks
     rev: v1.10.0

src/gitingest/cli.py

@@ -1,9 +1,7 @@
 import os
-from typing import Optional, Tuple
 
 import click
 
-from gitingest.ignore_patterns import DEFAULT_IGNORE_PATTERNS
 from gitingest.ingest import ingest
 from gitingest.ingest_from_query import MAX_FILE_SIZE
 
@@ -24,10 +22,10 @@ def normalize_pattern(pattern: str) -> str:
 @click.option('--include-pattern', '-i', multiple=True, help='Patterns to include')
 def main(
     source: str,
-    output: Optional[str],
+    output: str | None,
     max_size: int,
-    exclude_pattern: Tuple[str, ...],
-    include_pattern: Tuple[str, ...],
+    exclude_pattern: tuple[str, ...],
+    include_pattern: tuple[str, ...],
 ) -> None:
     """Analyze a directory and create a text dump of its contents."""
     try:

src/gitingest/clone.py

@@ -1,6 +1,5 @@
 import asyncio
 from dataclasses import dataclass
-from typing import Optional, Tuple
 
 from gitingest.utils import AsyncTimeoutError, async_timeout
 
@@ -11,8 +10,8 @@
 class CloneConfig:
     url: str
     local_path: str
-    commit: Optional[str] = None
-    branch: Optional[str] = None
+    commit: str | None = None
+    branch: str | None = None
 
 
 async def check_repo_exists(url: str) -> bool:
@@ -44,7 +43,7 @@ async def check_repo_exists(url: str) -> bool:
     return "HTTP/1.1 404" not in stdout_str and "HTTP/2 404" not in stdout_str
 
 
-async def run_git_command(*args: str) -> Tuple[bytes, bytes]:
+async def run_git_command(*args: str) -> tuple[bytes, bytes]:
     """
     Executes a git command asynchronously and captures its output.
 
@@ -77,7 +76,7 @@ async def run_git_command(*args: str) -> Tuple[bytes, bytes]:
 
 
 @async_timeout(CLONE_TIMEOUT)
-async def clone_repo(config: CloneConfig) -> Tuple[bytes, bytes]:
+async def clone_repo(config: CloneConfig) -> tuple[bytes, bytes]:
     """
     Clones a repository to a local path based on the provided query parameters.
 
@@ -107,8 +106,8 @@ async def clone_repo(config: CloneConfig) -> Tuple[bytes, bytes]:
     # Extract and validate query parameters
     url: str = config.url
     local_path: str = config.local_path
-    commit: Optional[str] = config.commit
-    branch: Optional[str] = config.branch
+    commit: str | None = config.commit
+    branch: str | None = config.branch
 
     if not url:
         raise ValueError("The 'url' parameter is required.")

src/gitingest/ignore_patterns.py

@@ -1,6 +1,4 @@
-from typing import List
-
-DEFAULT_IGNORE_PATTERNS: List[str] = [
+DEFAULT_IGNORE_PATTERNS: list[str] = [
     # Python
     '*.pyc',
     '*.pyo',

src/gitingest/ingest.py

@@ -2,7 +2,6 @@
 import inspect
 import shutil
 from pathlib import Path
-from typing import List, Optional, Tuple, Union
 
 from gitingest.clone import CloneConfig, clone_repo
 from gitingest.ingest_from_query import ingest_from_query
@@ -12,10 +11,10 @@
 def ingest(
     source: str,
     max_file_size: int = 10 * 1024 * 1024,  # 10 MB
-    include_patterns: Union[List[str], str, None] = None,
-    exclude_patterns: Union[List[str], str, None] = None,
-    output: Optional[str] = None,
-) -> Tuple[str, str, str]:
+    include_patterns: list[str] | str | None = None,
+    exclude_patterns: list[str] | str | None = None,
+    output: str | None = None,
+) -> tuple[str, str, str]:
     try:
         query = parse_query(
             source=source,

src/gitingest/ingest_from_query.py

@@ -1,6 +1,6 @@
 import os
 from fnmatch import fnmatch
-from typing import Any, Dict, List, Optional, Set, Tuple
+from typing import Any
 
 import tiktoken
 
@@ -10,7 +10,7 @@
 MAX_TOTAL_SIZE_BYTES = 500 * 1024 * 1024  # 500 MB
 
 
-def should_include(path: str, base_path: str, include_patterns: List[str]) -> bool:
+def should_include(path: str, base_path: str, include_patterns: list[str]) -> bool:
     rel_path = path.replace(base_path, "").lstrip(os.sep)
     include = False
     for pattern in include_patterns:
@@ -19,7 +19,7 @@ def should_include(path: str, base_path: str, include_patterns: List[str]) -> bo
     return include
 
 
-def should_exclude(path: str, base_path: str, ignore_patterns: List[str]) -> bool:
+def should_exclude(path: str, base_path: str, ignore_patterns: list[str]) -> bool:
     rel_path = path.replace(base_path, "").lstrip(os.sep)
     for pattern in ignore_patterns:
         if pattern == '':
@@ -60,11 +60,11 @@ def read_file_content(file_path: str) -> str:
 
 def scan_directory(
     path: str,
-    query: Dict[str, Any],
-    seen_paths: Optional[Set[str]] = None,
+    query: dict[str, Any],
+    seen_paths: set[str] | None = None,
     depth: int = 0,
-    stats: Optional[Dict[str, int]] = None,
-) -> Optional[Dict[str, Any]]:
+    stats: dict[str, int] | None = None,
+) -> dict[str, Any] | None:
     """Recursively analyzes a directory and its contents with safety limits."""
     if seen_paths is None:
         seen_paths = set()
@@ -220,11 +220,11 @@ def scan_directory(
 
 
 def extract_files_content(
-    query: Dict[str, Any],
-    node: Dict[str, Any],
+    query: dict[str, Any],
+    node: dict[str, Any],
     max_file_size: int,
-    files: Optional[List[Dict[str, Any]]] = None,
-) -> List[Dict[str, Any]]:
+    files: list[dict[str, Any]] | None = None,
+) -> list[dict[str, Any]]:
     """Recursively collects all text files with their contents."""
     if files is None:
         files = []
@@ -248,7 +248,7 @@ def extract_files_content(
     return files
 
 
-def create_file_content_string(files: List[Dict[str, Any]]) -> str:
+def create_file_content_string(files: list[dict[str, Any]]) -> str:
     """Creates a formatted string of file contents with separators."""
     output = ""
     separator = "=" * 48 + "\n"
@@ -278,7 +278,7 @@ def create_file_content_string(files: List[Dict[str, Any]]) -> str:
     return output
 
 
-def create_summary_string(query: Dict[str, Any], nodes: Dict[str, Any]) -> str:
+def create_summary_string(query: dict[str, Any], nodes: dict[str, Any]) -> str:
     """Creates a summary string with file counts and content size."""
     if "user_name" in query:
         summary = f"Repository: {query['user_name']}/{query['repo_name']}\n"
@@ -297,7 +297,7 @@ def create_summary_string(query: Dict[str, Any], nodes: Dict[str, Any]) -> str:
     return summary
 
 
-def create_tree_structure(query: Dict[str, Any], node: Dict[str, Any], prefix: str = "", is_last: bool = True) -> str:
+def create_tree_structure(query: dict[str, Any], node: dict[str, Any], prefix: str = "", is_last: bool = True) -> str:
     """Creates a tree-like string representation of the file structure."""
     tree = ""
 
@@ -319,7 +319,7 @@ def create_tree_structure(query: Dict[str, Any], node: Dict[str, Any], prefix: s
     return tree
 
 
-def generate_token_string(context_string: str) -> Optional[str]:
+def generate_token_string(context_string: str) -> str | None:
     """Returns the number of tokens in a text string."""
     formatted_tokens = ""
     try:
@@ -340,7 +340,7 @@ def generate_token_string(context_string: str) -> Optional[str]:
     return formatted_tokens
 
 
-def ingest_single_file(path: str, query: Dict[str, Any]) -> Tuple[str, str, str]:
+def ingest_single_file(path: str, query: dict[str, Any]) -> tuple[str, str, str]:
     if not os.path.isfile(path):
         raise ValueError(f"Path {path} is not a file")
 
@@ -376,7 +376,7 @@ def ingest_single_file(path: str, query: Dict[str, Any]) -> Tuple[str, str, str]
     return summary, tree, files_content
 
 
-def ingest_directory(path: str, query: Dict[str, Any]) -> Tuple[str, str, str]:
+def ingest_directory(path: str, query: dict[str, Any]) -> tuple[str, str, str]:
     nodes = scan_directory(path=path, query=query)
     if not nodes:
         raise ValueError(f"No files found in {path}")
@@ -392,7 +392,7 @@ def ingest_directory(path: str, query: Dict[str, Any]) -> Tuple[str, str, str]:
     return summary, tree, files_content
 
 
-def ingest_from_query(query: Dict[str, Any]) -> Tuple[str, str, str]:
+def ingest_from_query(query: dict[str, Any]) -> tuple[str, str, str]:
     """Main entry point for analyzing a codebase directory or single file."""
     path = f"{query['local_path']}{query['subpath']}"
     if not os.path.exists(path):

src/gitingest/parse_query.py

@@ -1,7 +1,7 @@
 import os
 import string
 import uuid
-from typing import Any, Dict, List, Optional, Union
+from typing import Any
 from urllib.parse import unquote
 
 from gitingest.ignore_patterns import DEFAULT_IGNORE_PATTERNS
@@ -10,7 +10,7 @@
 HEX_DIGITS = set(string.hexdigits)
 
 
-def parse_url(url: str) -> Dict[str, Any]:
+def parse_url(url: str) -> dict[str, Any]:
     url = url.split(" ")[0]
     url = unquote(url)  # Decode URL-encoded characters
 
@@ -88,7 +88,7 @@ def normalize_pattern(pattern: str) -> str:
     return pattern
 
 
-def parse_patterns(pattern: Union[List[str], str]) -> List[str]:
+def parse_patterns(pattern: list[str] | str) -> list[str]:
     patterns = pattern if isinstance(pattern, list) else [pattern]
     patterns = [p.strip() for p in patterns]
 
@@ -102,7 +102,7 @@ def parse_patterns(pattern: Union[List[str], str]) -> List[str]:
     return [normalize_pattern(p) for p in patterns]
 
 
-def override_ignore_patterns(ignore_patterns: List[str], include_patterns: List[str]) -> List[str]:
+def override_ignore_patterns(ignore_patterns: list[str], include_patterns: list[str]) -> list[str]:
     """
     Removes patterns from ignore_patterns that are present in include_patterns using set difference.
 
@@ -121,7 +121,7 @@ def override_ignore_patterns(ignore_patterns: List[str], include_patterns: List[
     return list(set(ignore_patterns) - set(include_patterns))
 
 
-def parse_path(path: str) -> Dict[str, Any]:
+def parse_path(path: str) -> dict[str, Any]:
     query = {
         "url": None,
         "local_path": os.path.abspath(path),
@@ -136,9 +136,9 @@ def parse_query(
     source: str,
     max_file_size: int,
     from_web: bool,
-    include_patterns: Optional[Union[List[str], str]] = None,
-    ignore_patterns: Optional[Union[List[str], str]] = None,
-) -> Dict[str, Any]:
+    include_patterns: list[str] | str | None = None,
+    ignore_patterns: list[str] | str | None = None,
+) -> dict[str, Any]:
     """
     Parses the input source to construct a query dictionary with specified parameters.
 

src/gitingest/tests/test_ingest.py

@@ -1,5 +1,5 @@
 from pathlib import Path
-from typing import Any, Dict
+from typing import Any
 
 import pytest
 
@@ -8,7 +8,7 @@
 
 # Test fixtures
 @pytest.fixture
-def sample_query() -> Dict[str, Any]:
+def sample_query() -> dict[str, Any]:
     return {
         'user_name': 'test_user',
         'repo_name': 'test_repo',
@@ -73,7 +73,7 @@ def temp_directory(tmp_path: Path) -> Path:
     return test_dir
 
 
-def test_scan_directory(temp_directory: Path, sample_query: Dict[str, Any]) -> None:
+def test_scan_directory(temp_directory: Path, sample_query: dict[str, Any]) -> None:
     result = scan_directory(str(temp_directory), query=sample_query)
     if result is None:
         assert False, "Result is None"
@@ -84,7 +84,7 @@ def test_scan_directory(temp_directory: Path, sample_query: Dict[str, Any]) -> N
     assert len(result['children']) == 5  # file1.txt, file2.py, src, dir1, dir2
 
 
-def test_extract_files_content(temp_directory: Path, sample_query: Dict[str, Any]) -> None:
+def test_extract_files_content(temp_directory: Path, sample_query: dict[str, Any]) -> None:
     nodes = scan_directory(str(temp_directory), query=sample_query)
     if nodes is None:
         assert False, "Nodes is None"

src/gitingest/utils.py

@@ -1,7 +1,8 @@
 ## Async Timeout decorator
 import asyncio
 import functools
-from typing import Awaitable, Callable, ParamSpec, TypeVar
+from collections.abc import Awaitable, Callable
+from typing import ParamSpec, TypeVar
 
 T = TypeVar("T")
 P = ParamSpec("P")

src/main.py

@@ -1,5 +1,4 @@
 import os
-from typing import Dict
 
 from api_analytics.fastapi import Analytics
 from dotenv import load_dotenv
@@ -52,7 +51,7 @@ async def rate_limit_exception_handler(request: Request, exc: Exception) -> Resp
 
 
 @app.get("/health")
-async def health_check() -> Dict[str, str]:
+async def health_check() -> dict[str, str]:
     return {"status": "healthy"}