BaseModel to validate RSS feed url input

src/models.py

@@ -0,0 +1,92 @@
+from dataclasses import dataclass
+from enum import StrEnum
+from typing import Self, assert_never
+from urllib.parse import urlsplit
+
+from pydantic import BaseModel, Field, field_validator, model_validator
+
+from feed_errors import InvalidFeedUrlError, UnsupportedPathError, WrongHostError
+from feed_handler import grab_gitea_release_information, grab_royal_road_chapter_information
+
+REPO_CHARS = frozenset("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_.-")
+
+@dataclass(frozen=True)
+class FeedUpdate():
+	entry_id: str
+	feed_title: str
+	body_title: str
+	link: str | None = None
+
+class FeedEntryType(StrEnum):
+	ROYALROAD = "royalroad"
+	GITEA = "gitea"
+
+def fetch_latest(feed_type: FeedEntryType, url: str) -> FeedUpdate:
+	match feed_type:
+		case FeedEntryType.ROYALROAD:
+			return grab_royal_road_chapter_information(url)
+		case FeedEntryType.GITEA:
+			return grab_gitea_release_information(url)
+		case _:
+			assert_never(feed_type)
+		
+
+def _path_segments(raw: str, feed_type: FeedEntryType, host: str) -> list[str]:
+	"""Check scheme and host, then return the non-empty path segments."""
+	raw = raw.strip()
+	if "://" not in raw:
+		raw = "https://" + raw  # allow "royalroad.com/fiction/118891"
+	try:
+		parts = urlsplit(raw)
+	except ValueError as e:
+		raise InvalidFeedUrlError("The URL could not be parsed") from e
+	if parts.scheme not in ("http", "https"):
+		raise InvalidFeedUrlError("The URL must use http or https")
+	# Comparing the whole netloc also rejects ports and user@host
+	if parts.netloc.lower().removeprefix("www.") != host:
+		raise WrongHostError(feed_type.value, host)
+	return [s for s in parts.path.split("/") if s]
+
+
+def _royalroad_feed(segments: list[str]) -> str:
+	# fiction/<id>[/<slug>]
+	if len(segments) >= 2 and segments[0] == "fiction":
+		story_id = segments[1]
+		if story_id.isascii() and story_id.isdigit():
+			return f"https://www.royalroad.com/fiction/syndication/{int(story_id)}"
+	raise UnsupportedPathError("expected e.g. royalroad.com/fiction/118891")
+
+
+def _gitea_feed(segments: list[str]) -> str:
+	# gitea/<repo>[/releases | /releases.rss]
+	if len(segments) >= 2 and segments[0] == "gitea":
+		repo, rest = segments[1], segments[2:]
+		repo_ok = repo[0] not in ".-" and set(repo) <= REPO_CHARS
+		if repo_ok and rest in ([], ["releases"], ["releases.rss"]):
+			return f"https://gitea.com/gitea/{repo}/releases.rss"
+	raise UnsupportedPathError("expected e.g. gitea.com/gitea/runner")
+
+
+# To support a new feed type: add an enum member, a function above, add a case tofetch_latest and a line here.
+_RULES = {
+	FeedEntryType.ROYALROAD: ("royalroad.com", _royalroad_feed),
+	FeedEntryType.GITEA: ("gitea.com", _gitea_feed),
+}
+
+
+class FeedEntry(BaseModel):
+	type: FeedEntryType
+	rss_feed_url: str = Field(max_length=512)
+	custom_name: str | None = Field(default=None, max_length=255)
+
+	@field_validator("custom_name")
+	@classmethod
+	def blank_name_is_none(cls, v: str | None) -> str | None:
+		return (v or "").strip() or None
+
+	@model_validator(mode="after")
+	def normalise_feed_url(self) -> Self:
+		host, build_feed_url = _RULES[self.type]
+		segments = _path_segments(self.rss_feed_url, self.type, host)
+		self.rss_feed_url = build_feed_url(segments)
+		return self