docs: remove PR/issue references from public docstrings

.github/workflows/publish-docs.yml

@@ -42,7 +42,9 @@ jobs:
       - name: 📝 Generate static HTML
         # `./spond` (not `spond`) forces pdoc to document the local checkout
         # instead of any same-named module that might be importable.
-        run: poetry run pdoc -o site/ ./spond
+        # `--docformat numpy` parses NumPy-style Parameters/Returns/Raises
+        # sections as structured lists rather than flat paragraphs.
+        run: poetry run pdoc --docformat numpy -o site/ ./spond
 
       - name: ⚙️ Configure Pages
         uses: actions/configure-pages@v5

README.md

@@ -104,7 +104,7 @@ install the dev dependencies and start the pdoc dev server:
 
 ```shell
 poetry install
-poetry run pdoc ./spond
+poetry run pdoc --docformat numpy ./spond
 ```
 
 A browser tab opens at `http://localhost:8080` with a searchable, navigable
@@ -114,9 +114,13 @@ next to each one. Pages update automatically when the docstrings change.
 To generate static HTML instead:
 
 ```shell
-poetry run pdoc -o docs/ ./spond
+poetry run pdoc --docformat numpy -o docs/ ./spond
 ```
 
+The `--docformat numpy` flag parses NumPy-style `Parameters`, `Returns`, and
+`Raises` sections as structured lists — omit it and the param list renders as
+one flat paragraph.
+
 The leading `./` is important when developing inside the repo — without it,
 pdoc would document the *installed* `spond` package from `site-packages`
 rather than your local checkout.

spond/__init__.py

@@ -17,7 +17,7 @@ class AuthenticationError(Exception):
 
     - Incorrect username/password.
     - 2FA enabled on the account (the library does not currently support
-      Spond's TOTP flow — see #205).
+      Spond's TOTP flow).
     - The account has hit Spond's login rate limit (`outOfLoginAttempts`).
     - The Spond login API has changed shape and the response no longer
       contains an `accessToken`.

spond/spond.py

@@ -82,7 +82,7 @@ def __init__(self, username: str, password: str) -> None:
             Spond account email address.
         password : str
             Spond account password. For accounts with 2FA enabled, login will
-            currently fail — see #205.
+            currently fail — Spond's TOTP flow is not yet supported.
         """
         super().__init__(username, password, self._API_BASE_URL)
         self._chat_url = None
@@ -449,13 +449,11 @@ async def get_events(
         and by visibility (scheduled, hidden). The full response is cached on
         `self.events`.
 
-        Note: `get_event(uid)` is a wrapper around this method via the cache,
-        so it inherits these defaults — an event that doesn't appear in the
+        Note: `get_event(uid)` looks up events via this method's cache, so
+        it inherits these defaults — an event that doesn't appear in the
         first `max_events` results or is excluded by `include_scheduled=False`
-        is unreachable through `get_event()` on current main. PR #236 changes
-        `get_event()` to fetch the singular `sponds/{uid}` endpoint directly,
-        removing that coupling; until it lands, pass appropriate filters here
-        when you need broader visibility.
+        is unreachable through `get_event()`. If you need broader visibility,
+        call this method directly with appropriate filters.
 
         Parameters
         ----------
@@ -542,11 +540,11 @@ async def get_events(
     async def get_event(self, uid: str) -> JSONDict:
         """Look up a single event by its unique id.
 
-        Currently routes through the cached events list (populated by
-        `get_events()`). Note this means events outside the `max_events=100`
-        default or with `scheduled=true` may not be findable — see #137 and
-        #138 (fix in PR #236 routes this through the singular `sponds/{uid}`
-        endpoint instead).
+        Routes through the cached events list (populated by `get_events()`),
+        which means events outside the `max_events=100` default or those
+        excluded by `include_scheduled=False` may not be findable. To reach
+        those events, call `get_events()` directly with appropriate filters
+        first to populate the cache, then call this method.
 
         Parameters
         ----------
@@ -614,9 +612,9 @@ async def get_event_attendance_xlsx(self, uid: str) -> bytes:
 
         Thin wrapper around Spond's own "Export attendance history" feature
         in the web UI. The columns and format are determined by Spond, not by
-        this library — for example, the export does not include member ids
-        (see closed issue #227). For a customisable CSV alternative built
-        from `get_event()` data, see `examples/attendance.py`.
+        this library — for example, the export does not include member ids.
+        For a customisable CSV alternative built from `get_event()` data,
+        see `examples/attendance.py`.
 
         Parameters
         ----------
@@ -680,9 +678,10 @@ async def _get_entity(self, entity_type: str, uid: str) -> JSONDict:
 
         Routes to the relevant cache (`self.events` or `self.groups`),
         triggers a fetch via `get_events()` / `get_groups()` if the cache is
-        empty, then linearly scans for a matching `id`. The empty-cache case
-        is handled explicitly to avoid the `TypeError: 'NoneType' object is
-        not iterable` that previously occurred (see #136, fixed in #235).
+        empty, then linearly scans for a matching `id`. Raises `KeyError`
+        cleanly (rather than `TypeError`) when the cache remains empty after
+        the fetch attempt — the underlying `get_*s()` method may legitimately
+        return `None` if the account has no events/groups available.
 
         Parameters
         ----------