pr fixes

Author: filipchristiansen

src/gitingest/cli.py

@@ -131,6 +131,7 @@ async def _async_main(
         If ``True``, also ingest files matched by ``.gitignore`` or ``.gitingestignore`` (default: ``False``).
     token : str | None
         GitHub personal access token (PAT) for accessing private repositories.
+        Can also be set via the ``GITHUB_TOKEN`` environment variable.
     output : str | None
         The path where the output file will be written (default: ``digest.txt`` in current directory).
         Use ``"-"`` to write to ``stdout``.

src/gitingest/utils/exceptions.py

@@ -41,6 +41,9 @@ def __init__(self, message: str) -> None:
 class InvalidGitHubTokenError(ValueError):
     """Exception raised when a GitHub Personal Access Token is malformed."""
 
-    def __init__(self, token: str) -> None:
-        msg = f"Invalid GitHub token format: {token!r}. To generate a token, see https://github.com/settings/tokens."
+    def __init__(self) -> None:
+        msg = (
+            "Invalid GitHub token format. To generate a token, go to "
+            "https://github.com/settings/tokens/new?description=gitingest&scopes=repo."
+        )
         super().__init__(msg)

src/gitingest/utils/git_utils.py

@@ -9,17 +9,22 @@
 from typing import Final
 from urllib.parse import urlparse
 
+from starlette.status import (
+    HTTP_200_OK,
+    HTTP_301_MOVED_PERMANENTLY,
+    HTTP_302_FOUND,
+    HTTP_401_UNAUTHORIZED,
+    HTTP_403_FORBIDDEN,
+    HTTP_404_NOT_FOUND,
+)
+
 from gitingest.utils.exceptions import InvalidGitHubTokenError
 
 # GitHub Personal-Access tokens (classic + fine-grained).
 #   - ghp_ / gho_ / ghu_ / ghs_ / ghr_  → 36 alphanumerics
 #   - github_pat_                       → 22 alphanumerics + "_" + 59 alphanumerics
 _GITHUB_PAT_PATTERN: Final[str] = r"^(?:gh[pousr]_[A-Za-z0-9]{36}|github_pat_[A-Za-z0-9]{22}_[A-Za-z0-9]{59})$"
 
-_OK: Final[set[str]] = {"200", "301"}  # reachable / canonical redirect
-_MISSING: Final[set[str]] = {"404", "302"}  # not found / redirect
-_NEEDS_AUTH: Final[set[str]] = {"401", "403"}  # login or forbidden
-
 
 def is_github_host(url: str) -> bool:
     """Check if a URL is from a GitHub host (github.com or GitHub Enterprise).
@@ -109,6 +114,7 @@ async def check_repo_exists(url: str, token: str | None = None) -> bool:
         If the host returns an unrecognised status code.
 
     """
+    # TODO: use `requests` instead of `curl`
     cmd: list[str] = [
         "curl",
         "--silent",
@@ -139,14 +145,15 @@ async def check_repo_exists(url: str, token: str | None = None) -> bool:
     if proc.returncode != 0:
         return False
 
-    status = stdout.decode().strip()
-    if status in _OK:
+    status = int(stdout.decode().strip())
+    if status in {HTTP_200_OK, HTTP_301_MOVED_PERMANENTLY}:
         return True
-    if status in _MISSING:
+    # TODO: handle 302 redirects
+    if status in {HTTP_404_NOT_FOUND, HTTP_302_FOUND}:
         return False
-    if status in _NEEDS_AUTH:
+    if status in {HTTP_401_UNAUTHORIZED, HTTP_403_FORBIDDEN}:
         return False
-    msg = f"Unexpected HTTP status {status!r} for {url}"
+    msg = f"Unexpected HTTP status {status} for {url}"
     raise RuntimeError(msg)
 
 
@@ -294,4 +301,4 @@ def validate_github_token(token: str) -> None:
 
     """
     if not re.fullmatch(_GITHUB_PAT_PATTERN, token):
-        raise InvalidGitHubTokenError(token)
+        raise InvalidGitHubTokenError

src/server/routers/download.py

@@ -2,6 +2,7 @@
 
 from fastapi import APIRouter, HTTPException
 from fastapi.responses import FileResponse
+from starlette.status import HTTP_403_FORBIDDEN, HTTP_404_NOT_FOUND
 
 from gitingest.config import TMP_BASE_PATH
 
@@ -32,14 +33,17 @@ async def download_ingest(digest_id: str) -> FileResponse:
     directory = TMP_BASE_PATH / digest_id
 
     if not directory.is_dir():
-        raise HTTPException(status_code=404, detail=f"Digest {digest_id!r} not found")
+        raise HTTPException(status_code=HTTP_404_NOT_FOUND, detail=f"Digest {digest_id!r} not found")
 
     try:
         first_txt_file = next(directory.glob("*.txt"))
     except StopIteration as exc:
-        raise HTTPException(status_code=404, detail=f"No .txt file found for digest {digest_id!r}") from exc
+        raise HTTPException(
+            status_code=HTTP_404_NOT_FOUND,
+            detail=f"No .txt file found for digest {digest_id!r}",
+        ) from exc
 
     try:
         return FileResponse(path=first_txt_file, media_type="text/plain", filename=first_txt_file.name)
     except PermissionError as exc:
-        raise HTTPException(status_code=403, detail=f"Permission denied for {first_txt_file}") from exc
+        raise HTTPException(status_code=HTTP_403_FORBIDDEN, detail=f"Permission denied for {first_txt_file}") from exc