chore(teams): log role batch vs bot_top when edit_role_positions fails

A 50013 on positioning despite Manage Roles points to a role in the reorder
batch sitting at/above the bot's top role. Log the batch and bot_top so the
offending role is identifiable from the journal.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XPRsW4tazVtYi2jNzVQkre

.env.example

@@ -10,10 +10,28 @@ DISCORD_TOKEN=
 SHEET_ID=your-google-sheet-id-here
 
 # Separate spreadsheet holding tournament team registrations (Team Name + lineup
-# of Discord usernames). Optional; drives /teamsync + hourly team-role sync on
+# of Discord usernames). Optional; a FALLBACK source for /teamsync + hourly
-# the economy/community bot. Leave unset to disable team-role sync entirely.
+# team-role sync on the economy/community bot. Leave unset to disable it.
 TEAM_SHEET_ID=
 
+# Fienta ticketing - PRIMARY source for team-role sync. The registration collects
+# each competitor's Discord username + team name per ticket, so matching is by
+# real Discord handle (not game nickname). Get an API token from the Fienta admin
+# (organizer settings) and the event's numeric ID from its dashboard URL. Leave
+# unset to use only the sheet.
+FIENTA_API_TOKEN=
+FIENTA_EVENT_ID=
+
+# Where each game's team roles get positioned in the role list. The key suffix
+# is matched against the section's title row in the sheet ("TipiLAN 2026 CS2
+# Registration Log") - every underscore-separated part must appear in it, so
+# CS2_2026 matches only the 2026 CS2 block while a plain CS2 would match any
+# year. The value is the Discord role ID to place those teams under (matching by
+# ID means renaming the divider role never breaks positioning).
+# Optional: sections that match nothing still get their roles, just unpositioned.
+TEAM_DIVIDER_CS2_2026=1498736834656604251
+TEAM_DIVIDER_LOL_2026=1498736949706490017
+
 # Path to Google service account credentials JSON
 GOOGLE_CREDS_PATH=credentials.json
 

.gitignore

@@ -10,3 +10,6 @@ pocketbase
 pb_data/
 pb_migrations/
 logs/
+fientalog
+fientatickets
+fientaorders

README.md

@@ -10,8 +10,9 @@ Discord bot for the TipiLAN community. Manages member roles and nicknames via Go
 2. [Member Management](#member-management)
 3. [Admin Commands](#admin-commands)
 4. [Birthday System](#birthday-system)
-5. [TipiCOIN Economy](#tipicoin-economy)
+5. [Tournament Team Roles](#tournament-team-roles)
-6. [Project Structure](#project-structure)
+6. [TipiCOIN Economy](#tipicoin-economy)
+7. [Project Structure](#project-structure)
 
 ---
 
@@ -107,6 +108,8 @@ cp .env.example .env
 | `PB_ECONOMY_COLLECTION_DEV` | PocketBase collection used by `BOT_PROFILE=dev` |
 | `PB_ECONOMY_COLLECTION_ECONOMY` | PocketBase collection used by `BOT_PROFILE=economy` |
 | `PB_ECONOMY_COLLECTION` | Legacy fallback collection (optional) |
+| `TEAM_SHEET_ID` | Tournament registration spreadsheet (separate from the member sheet). Unset = team sync off |
+| `TEAM_DIVIDER_<GAME>_<YEAR>` | Divider role that a sheet section's teams get positioned under - see [Tournament team roles](#tournament-team-roles) |
 
 ### 6. Install & Run
 
@@ -168,6 +171,7 @@ Admins (bot lacks permission to modify them) are silently skipped and still mark
 |---|---|---|
 | `/check` | Manage Roles | Refreshes sheet data, backfills missing User IDs, syncs nicknames + roles for every member, reports stats |
 | `/member @user` | Manage Roles | Shows a member's full sheet data + calculated age |
+| `/teamsync` | Bot admin | Reloads the tournament registration sheet, grants/removes team roles, and repositions them under their dividers. See [Tournament Team Roles](#tournament-team-roles) |
 | `/sync` | Manage Guild | Re-registers slash commands with Discord |
 | `/restart` | Manage Guild | Gracefully restarts the bot process; posts ✅ in the same channel when back up |
 | `/shutdown` | Manage Guild | Shuts the bot down cleanly without restarting |
@@ -220,6 +224,46 @@ If a member joins and their birthday is within `BIRTHDAY_WINDOW_DAYS` days, a bi
 
 ---
 
+## Tournament Team Roles
+
+Runs on the **economy** profile only, driven by the separate `TEAM_SHEET_ID` spreadsheet. It is independent of the member roster - players are matched straight from the registration sheet by Discord username, and never need a row in the member sheet.
+
+Re-runs automatically **every hour**, and on demand via `/teamsync` (admin).
+
+### What it does
+
+- Grants each registered player the role named after their team, creating that role if it doesn't exist yet
+- Removes team roles they no longer hold (switched teams, dropped out)
+- Positions every team role directly beneath its game's divider role
+- **Only role names present in the sheet are ever added, removed, or moved** - no unrelated role is at risk
+
+### Divider placement
+
+The registration sheet stacks several game sections per tab, each under a merged title row:
+
+```
+[merged] TipiLAN 2026 CS2 Registration Log
+No | Team Name | Lineup (nickname, citizenship) | ...
+1  | GENESIS   | kapa (EST), neaQ (EST)         | ...
+```
+
+Set one `TEAM_DIVIDER_*` variable per section to say where those teams belong:
+
+```bash
+TEAM_DIVIDER_CS2_2026="====== COUNTER-STRIKE 2 2026 ======"
+TEAM_DIVIDER_LOL_2026="===== LEAGUE OF LEGENDS 2026 ====="
+```
+
+- **The key suffix** is matched against the section's title row. Every underscore-separated part must appear in it as a whole word, so `CS2_2026` matches only the 2026 CS2 block, while a plain `CS2` would match that game in any year. When several keys match, the most specific one (most parts) wins.
+- **The value** is the *exact* Discord role name, decoration included. Copy it from Server Settings → Roles.
+- Sections matching no key still get their team roles - they just aren't repositioned.
+
+Teams are ordered alphabetically downwards under their divider. Every sync re-checks placement, so roles that predate this feature get pulled into place on the next run.
+
+> The bot can only move roles **below its own** role. Drag the TipiBOT role above your dividers, or placement is skipped and reported as an error.
+
+---
+
 ## TipiCOIN Economy
 
 All economy data is stored in **PocketBase** (`economy_users` collection - see `core/pb_client.py`). The currency is **TipiCOIN** (⬡), displayed as a custom Discord emoji configured in `core/economy/store.py → COIN`.

bot.py

@@ -21,7 +21,7 @@ import psutil
 
 import config
 import strings as S
-from core import economy, pb_client, sheets
+from core import economy, fienta, pb_client, sheets
 from core.admin import is_bot_admin
 from core.member_sync import SyncResult, sync_all_team_roles
 from commands.dev_member_commands import register_dev_member_commands
@@ -339,31 +339,38 @@ async def before_birthday_daily():
 
 @tasks.loop(hours=1)
 async def team_sync_hourly():
-    """Reload the tournament registration sheet and re-apply team roles.
+    """Reload the tournament registration (Fienta + sheet) and re-apply team roles.
 
     Economy profile only (the tournament players live in the community guild).
     Runs the first iteration immediately on start, so this also covers the
-    initial load at boot. No-op when TEAM_SHEET_ID is unset.
+    initial load at boot. No-op when neither Fienta nor the team sheet is set.
     """
-    if IS_DEV_PROFILE or not config.TEAM_SHEET_ID:
+    if IS_DEV_PROFILE or not (config.TEAM_SHEET_ID or config.FIENTA_API_TOKEN):
         return
     try:
         rosters = await sheets.refresh_teams()
+        fienta_teams = await fienta.refresh_teams()
     except Exception as e:
-        log.error("team_sync_hourly: failed to load team sheet: %s", e)
+        log.error("team_sync_hourly: failed to load team data: %s", e)
         return
-    if not rosters:
+    if not rosters and not fienta_teams:
         return
     guild = bot.get_guild(config.GUILD_ID)
     if guild is None:
         log.warning("team_sync_hourly: guild %s not found", config.GUILD_ID)
         return
     summary = await sync_all_team_roles(guild, log)
-    if summary.assigned or summary.removed or summary.created:
+    if (summary.assigned or summary.removed or summary.created or summary.positioned
+            or summary.divider_assigned or summary.divider_removed or summary.errors):
         log.info(
-            "team_sync_hourly: assigned=%d, removed=%d, created=%d, errors=%d",
+            "team_sync_hourly: assigned=%d, removed=%d, created=%d, positioned=%d, "
-            summary.assigned, summary.removed, len(summary.created), len(summary.errors),
+            "divider_assigned=%d, divider_removed=%d, errors=%d",
+            summary.assigned, summary.removed, len(summary.created),
+            summary.positioned, summary.divider_assigned, summary.divider_removed,
+            len(summary.errors),
         )
+        for err in summary.errors:
+            log.warning("team_sync_hourly: %s", err)
 
 
 @team_sync_hourly.before_loop
@@ -469,7 +476,8 @@ async def on_ready():
         log.info("Birthday daily task started (fires 09:00 Tallinn time)")
 
     # Start hourly tournament team-role sync (economy/community guild)
-    if not IS_DEV_PROFILE and config.TEAM_SHEET_ID and not team_sync_hourly.is_running():
+    if (not IS_DEV_PROFILE and (config.TEAM_SHEET_ID or config.FIENTA_API_TOKEN)
+            and not team_sync_hourly.is_running()):
         team_sync_hourly.start()
         log.info("Team-role sync task started (hourly, from the registration sheet)")
 

commands/economy_team_commands.py

@@ -14,7 +14,7 @@ import logging
 import discord
 from discord import app_commands
 
-from core import sheets
+from core import fienta, sheets
 from core.admin import bot_admin_check
 from core.member_sync import sync_all_team_roles
 import strings as S
@@ -38,23 +38,28 @@ def register_economy_team_commands(
 
         try:
             rosters = await sheets.refresh_teams()
+            fienta_teams = await fienta.refresh_teams()
         except Exception as e:
             await interaction.followup.send(
                 S.TEAMSYNC_UI["refresh_error"].format(error=e), ephemeral=True
             )
             return
-        if not rosters:
+        if not rosters and not fienta_teams:
             await interaction.followup.send(S.TEAMSYNC_UI["disabled"], ephemeral=True)
             return
 
         summary = await sync_all_team_roles(guild, log)
         await interaction.followup.send(_format_summary(summary), ephemeral=True)
         log.info(
-            "/teamsync - scanned=%d, assigned=%d, removed=%d, created=%d, errors=%d",
+            "/teamsync - scanned=%d, assigned=%d, removed=%d, created=%d, "
+            "positioned=%d, divider_assigned=%d, divider_removed=%d, errors=%d",
             summary.scanned,
             summary.assigned,
             summary.removed,
             len(summary.created),
+            summary.positioned,
+            summary.divider_assigned,
+            summary.divider_removed,
             len(summary.errors),
         )
 
@@ -71,6 +76,12 @@ def _format_summary(summary) -> str:
         # several members in one run - de-dupe for the report.
         unique = list(dict.fromkeys(summary.created))
         lines.append(S.TEAMSYNC_UI["created"].format(roles=", ".join(unique)))
+    if summary.positioned:
+        lines.append(S.TEAMSYNC_UI["positioned"].format(count=summary.positioned))
+    if summary.divider_assigned:
+        lines.append(S.TEAMSYNC_UI["divider_assigned"].format(count=summary.divider_assigned))
+    if summary.divider_removed:
+        lines.append(S.TEAMSYNC_UI["divider_removed"].format(count=summary.divider_removed))
     if summary.errors:
         lines.append(S.TEAMSYNC_UI["errors"].format(count=len(summary.errors)))
 

config.py

@@ -28,6 +28,13 @@ SHEET_ID = os.getenv("SHEET_ID")
 TEAM_SHEET_ID = os.getenv("TEAM_SHEET_ID")
 GOOGLE_CREDS_PATH = os.getenv("GOOGLE_CREDS_PATH", "credentials.json")
 
+# Fienta ticketing: the registration collects each competitor's Discord username
+# (+ sometimes Discord user ID) and team name per ticket, giving a reliable
+# Discord-identity -> team mapping the nickname-only sheet cannot. Primary source
+# for team-role sync; the sheet stays as a fallback. Unset -> Fienta is skipped.
+FIENTA_API_TOKEN = os.getenv("FIENTA_API_TOKEN", "")
+FIENTA_EVENT_ID = os.getenv("FIENTA_EVENT_ID", "")
+
 _LEGACY_GUILD_ID = _env_int("GUILD_ID", 0)
 GUILD_ID_DEV = _env_int("GUILD_ID_DEV", _LEGACY_GUILD_ID)
 GUILD_ID_ECONOMY = _env_int("GUILD_ID_ECONOMY", _LEGACY_GUILD_ID)
@@ -69,6 +76,43 @@ def _parse_admin_roles(raw: str) -> dict[int, set[int]]:
 
 BOT_ADMIN_ROLES: dict[int, set[int]] = _parse_admin_roles(os.getenv("DISCORD_ADMIN_ROLES", ""))
 
+_TEAM_DIVIDER_PREFIX = "TEAM_DIVIDER_"
+
+
+def _parse_team_dividers() -> dict[str, int]:
+    """Collect TEAM_DIVIDER_<SUFFIX> env vars into {suffix: divider role ID}.
+
+    The suffix says which sheet sections the divider covers, the value is the
+    Discord role ID their teams get positioned under:
+
+        TEAM_DIVIDER_CS2_2026=1498736834656604251
+
+    Matching by ID (not name) means renaming the divider role in Discord never
+    breaks positioning. Every underscore-separated part of the suffix must appear
+    in the section's title row, so `CS2_2026` matches only "TipiLAN 2026 CS2
+    Registration Log" while a plain `CS2` would match that section in any year.
+    Defining a var is what switches positioning on for those sections; teams
+    whose section matches nothing are still granted their role, just never moved.
+    """
+    dividers: dict[str, int] = {}
+    for key, value in os.environ.items():
+        if not key.startswith(_TEAM_DIVIDER_PREFIX):
+            continue
+        suffix = key[len(_TEAM_DIVIDER_PREFIX):].strip().lower()
+        raw = value.strip()
+        if not suffix or not raw:
+            continue
+        try:
+            dividers[suffix] = int(raw)
+        except ValueError:
+            raise SystemExit(
+                f"{key}: expected a Discord role ID (integer), got {raw!r}"
+            )
+    return dividers
+
+
+TEAM_DIVIDERS: dict[str, int] = _parse_team_dividers()
+
 PB_URL = os.getenv("PB_URL", "http://127.0.0.1:8090")
 PB_ADMIN_EMAIL = os.getenv("PB_ADMIN_EMAIL", "")
 PB_ADMIN_PASSWORD = os.getenv("PB_ADMIN_PASSWORD", "")

core/fienta.py

@@ -0,0 +1,190 @@
+"""Fienta ticketing integration - authoritative Discord -> team mapping.
+
+The tournament registration on Fienta collects, per competitor ticket, the
+player's Discord username (and sometimes Discord user ID), their team name
+(order-level, echoed onto each attendee), and the ticket type (which names the
+game). That gives a reliable Discord-identity -> team mapping the nickname-only
+registration sheet cannot, so this is the PRIMARY source for team-role sync,
+with :mod:`core.sheets` kept as a fallback.
+
+Enabled by ``FIENTA_API_TOKEN`` + ``FIENTA_EVENT_ID``; when either is unset all
+caches stay empty and every getter is a no-op, so team sync silently falls back
+to the sheet. Only real team members get roles - competitor, coach/manager and
+substitute tickets are included; visitors, supporters, LAN-access, early-bird
+and waiting-list tickets are excluded.
+"""
+
+from __future__ import annotations
+
+import logging
+
+import aiohttp
+
+import config
+
+log = logging.getLogger(__name__)
+
+_API_BASE = "https://fienta.com/api/v1"
+_PAGE_SIZE = 1000  # Fienta's max per page; one page covers a whole tournament
+
+# Ticket-type titles that represent an actual team member who should get a role.
+_INCLUDED_KEYWORDS = ("competitor", "coach", "manager", "substitute")
+# ...unless the title also matches one of these (visitors etc. are never players).
+_EXCLUDED_KEYWORDS = (
+    "visitor", "supporter", "lan area", "early bird", "waiting list", "waitlist",
+)
+
+# Attendee custom-field keys. Fienta appends the field id to the machine name;
+# these come from GET /events/{id}/custom-fields for event 176532.
+_F_DISCORD_USERNAME = "discord_username_134871"
+_F_DISCORD_USERID = "discord_user_id_135840"
+_F_TEAM_NAME = "team_name_134821"
+
+_GAME_KEYWORDS: list[tuple[str, tuple[str, ...]]] = [
+    ("CS2", ("counter-strike", "counter strike", "cs2", "csgo", "cs:go")),
+    ("LoL", ("league of legends", "league", "lol")),
+]
+
+# Caches, rebuilt by refresh_teams()/parse_tickets().
+_by_username: dict[str, str] = {}   # discord username (lower) -> team name
+_by_userid: dict[str, str] = {}     # discord user id (str)    -> team name
+_team_game: dict[str, str] = {}     # team name -> "CS2" / "LoL"
+_team_names: set[str] = set()
+
+
+def _norm_username(name: str) -> str:
+    """Normalise a Discord username for matching: lowercased, no leading @."""
+    return name.strip().lower().lstrip("@")
+
+
+def _detect_game(ticket_type_title: str) -> str | None:
+    """Return the game code for a ticket-type title, or None if unrecognised."""
+    title = ticket_type_title.lower()
+    for game, keywords in _GAME_KEYWORDS:
+        if any(k in title for k in keywords):
+            return game
+    return None
+
+
+def _is_included(ticket_type_title: str) -> bool:
+    """True when this ticket type is an actual team member (not a visitor etc.)."""
+    title = ticket_type_title.lower()
+    if any(k in title for k in _EXCLUDED_KEYWORDS):
+        return False
+    return any(k in title for k in _INCLUDED_KEYWORDS)
+
+
+def _game_divider_ids() -> dict[str, int]:
+    """Map game code -> divider role id, derived from ``config.TEAM_DIVIDERS``.
+
+    Reuses the same ``TEAM_DIVIDER_<SUFFIX>`` role IDs the sheet path uses: a
+    suffix like ``cs2_2026`` contributes its id to game ``CS2``.
+    """
+    out: dict[str, int] = {}
+    for suffix, rid in config.TEAM_DIVIDERS.items():
+        parts = suffix.split("_")
+        if any(p in ("cs2", "cs", "csgo") for p in parts):
+            out.setdefault("CS2", rid)
+        if any(p in ("lol", "league") for p in parts):
+            out.setdefault("LoL", rid)
+    return out
+
+
+def parse_tickets(tickets: list[dict]) -> None:
+    """Rebuild the caches from a list of Fienta ticket objects.
+
+    Pure/synchronous so it can be unit-tested without hitting the API.
+    """
+    global _by_username, _by_userid, _team_game, _team_names
+    by_username: dict[str, str] = {}
+    by_userid: dict[str, str] = {}
+    team_game: dict[str, str] = {}
+    for ticket in tickets:
+        rows = ticket.get("rows") or []
+        if not rows:
+            continue
+        row = rows[0]
+        title = (row.get("ticket_type") or {}).get("title", "")
+        if not _is_included(title):
+            continue
+        attendee = row.get("attendee") or {}
+        team = (attendee.get(_F_TEAM_NAME) or "").strip()
+        if not team:
+            continue
+        game = _detect_game(title)
+        # Keep the first non-None game seen for a team (all its tickets agree).
+        team_game[team] = game or team_game.get(team)
+        uname = _norm_username(attendee.get(_F_DISCORD_USERNAME) or "")
+        uid = (attendee.get(_F_DISCORD_USERID) or "").strip()
+        if uname:
+            by_username[uname] = team
+        if uid.isdigit():
+            by_userid[uid] = team
+    _by_username = by_username
+    _by_userid = by_userid
+    _team_game = team_game
+    _team_names = set(team_game)
+
+
+async def refresh_teams() -> set[str]:
+    """Fetch competitor tickets from Fienta and rebuild the caches.
+
+    No-op returning an empty set when ``FIENTA_API_TOKEN`` / ``FIENTA_EVENT_ID``
+    are unset, so the caller transparently falls back to the sheet.
+    """
+    if not config.FIENTA_API_TOKEN or not config.FIENTA_EVENT_ID:
+        parse_tickets([])
+        return set()
+
+    url = f"{_API_BASE}/events/{config.FIENTA_EVENT_ID}/tickets"
+    headers = {"Authorization": f"Bearer {config.FIENTA_API_TOKEN}"}
+    tickets: list[dict] = []
+    async with aiohttp.ClientSession() as session:
+        page = 1
+        while True:
+            params = {"attendees": "true", "per_page": str(_PAGE_SIZE), "page": str(page)}
+            async with session.get(url, headers=headers, params=params) as resp:
+                resp.raise_for_status()
+                data = await resp.json()
+            batch = data.get("tickets") or []
+            tickets.extend(batch)
+            if len(batch) < _PAGE_SIZE:
+                break
+            page += 1
+
+    parse_tickets(tickets)
+    log.info(
+        "Fienta: %d tickets -> %d teams, %d discord usernames, %d discord ids",
+        len(tickets), len(_team_names), len(_by_username), len(_by_userid),
+    )
+    return set(_team_names)
+
+
+def get_team_for_username(username: str) -> str | None:
+    """Team the given Discord username is registered on, or None."""
+    return _by_username.get(_norm_username(username))
+
+
+def get_team_for_userid(user_id: int) -> str | None:
+    """Team the given Discord user ID is registered on, or None (IDs are sparse)."""
+    return _by_userid.get(str(user_id))
+
+
+def all_team_names() -> set[str]:
+    """Every team name seen in the included Fienta tickets."""
+    return set(_team_names)
+
+
+def get_team_game(team: str) -> str | None:
+    """Game code ("CS2"/"LoL") for a team, or None."""
+    return _team_game.get(team)
+
+
+def get_team_dividers() -> dict[str, int]:
+    """{team name -> divider role id}, via each team's game and config dividers."""
+    game_div = _game_divider_ids()
+    return {
+        team: game_div[game]
+        for team, game in _team_game.items()
+        if game and game in game_div
+    }

core/member_sync.py

@@ -11,9 +11,32 @@ from zoneinfo import ZoneInfo
 import discord
 
 import config
-from . import sheets
+from . import fienta, sheets
 
 log = logging.getLogger(__name__)
+
+
+def resolve_team(member: discord.Member) -> str | None:
+    """Team a member is registered on: Fienta first (by ID, then username),
+    then the sheet by username. Fienta is authoritative; the sheet is fallback."""
+    return (
+        fienta.get_team_for_userid(member.id)
+        or fienta.get_team_for_username(member.name)
+        or sheets.get_team_for_username(member.name)
+    )
+
+
+def all_managed_team_names() -> set[str]:
+    """Union of every team name from Fienta and the sheet - the only role names
+    team sync ever adds or removes."""
+    return fienta.all_team_names() | sheets.all_team_names()
+
+
+def team_dividers() -> dict[str, int]:
+    """{team -> divider role id} merged from both sources; Fienta wins on overlap."""
+    merged = dict(sheets.get_team_dividers())
+    merged.update(fienta.get_team_dividers())
+    return merged
 _PLACEHOLDER = {"-", "x", "n/a", "none", "ei"}
 _TZ = ZoneInfo("Europe/Tallinn")
 
@@ -65,11 +88,13 @@ class TeamSyncResult:
     added: str | None = None       # team role name granted, if any
     removed: list[str] = field(default_factory=list)  # stale team roles taken away
     created: str | None = None     # team role name auto-created in the guild, if any
+    divider_added: str | None = None  # game divider role granted as participant tag
+    divider_removed: list[str] = field(default_factory=list)  # stale divider roles taken away
     errors: list[str] = field(default_factory=list)
 
     @property
     def changed(self) -> bool:
-        return bool(self.added or self.removed)
+        return bool(self.added or self.removed or self.divider_added or self.divider_removed)
 
 
 @dataclass
@@ -79,6 +104,9 @@ class TeamSyncSummary:
     assigned: int = 0
     removed: int = 0
     created: list[str] = field(default_factory=list)
+    positioned: int = 0                               # team roles moved under a divider
+    divider_assigned: int = 0                         # members given their game divider role
+    divider_removed: int = 0                          # stale game divider roles taken away
     changes: list[str] = field(default_factory=list)  # human-readable per-member lines
     errors: list[str] = field(default_factory=list)
 
@@ -259,18 +287,21 @@ async def sync_team_role(
 
     * grants the role for the team they're registered on (auto-creating that
       role in the guild when it does not exist yet);
-    * removes any *other* team role they still carry (left / switched teams).
+    * removes any *other* team role they still carry (left / switched teams);
+    * grants their game's divider role as a participant tag (and strips any
+      other configured divider role they still carry, i.e. switched game).
 
-    Only role NAMES present in the team sheet are ever added or removed, so no
+    Only team role NAMES present in the team sheet and the configured divider
+    role IDs (``config.TEAM_DIVIDERS``) are ever added or removed, so no
     unrelated role is ever at risk. When ``TEAM_SHEET_ID`` is unset the caches
     are empty and this is a no-op returning an unchanged result.
     """
     result = TeamSyncResult()
 
-    team_name = sheets.get_team_for_username(member.name)
+    team_name = resolve_team(member)
-    all_teams = sheets.all_team_names()
+    all_teams = all_managed_team_names()
     if not all_teams:
-        return result  # feature switched off (no team sheet loaded)
+        return result  # feature switched off (no Fienta token and no team sheet)
 
     desired: discord.Role | None = None
     if team_name:
@@ -306,9 +337,164 @@ async def sync_team_role(
         except discord.HTTPException as e:
             result.errors.append(f"Tiimirolli eemaldamise viga kasutajale {member}: {e}")
 
+    # --- Participant divider role (the game's divider role doubles as a tag) ---
+    # Grant the divider role for the member's team's game, and strip any other
+    # configured divider role (switched game / dropped out). Matched by ID, so
+    # only the roles named in config.TEAM_DIVIDERS are ever touched.
+    divider_ids = set(config.TEAM_DIVIDERS.values())
+    want_divider_id = team_dividers().get(team_name) if team_name else None
+    want_divider = guild.get_role(want_divider_id) if want_divider_id else None
+
+    if want_divider is not None and want_divider not in member.roles:
+        try:
+            await member.add_roles(want_divider, reason="Team sync: mänguosaleja")
+            result.divider_added = want_divider.name
+        except discord.Forbidden:
+            log.debug("No permission to add divider role for %s, skipping", member)
+        except discord.HTTPException as e:
+            result.errors.append(f"Eraldajarolli viga kasutajale {member}: {e}")
+
+    stale_dividers = [
+        r for r in member.roles if r.id in divider_ids and r.id != want_divider_id
+    ]
+    if stale_dividers:
+        try:
+            await member.remove_roles(*stale_dividers, reason="Team sync: mäng vahetus")
+            result.divider_removed = [r.name for r in stale_dividers]
+        except discord.Forbidden:
+            log.debug("No permission to remove divider roles for %s, skipping", member)
+        except discord.HTTPException as e:
+            result.errors.append(f"Eraldajarolli eemaldamise viga kasutajale {member}: {e}")
+
     return result
 
 
+def plan_team_positions(
+    ordered_names: list[str],
+    managed: set[str],
+    placements: dict[str, list[str]],
+) -> dict[str, int]:
+    """Work out the new position of every team role that needs to move.
+
+    ``ordered_names`` is every role name ascending by Discord position, so index
+    0 is the bottom of the role list (@everyone) and the last entry is the top.
+    ``placements`` maps a divider role name to the team roles that belong under
+    it. Roles in ``managed`` are never moved (integration-managed roles, and
+    anything at or above the bot's own role, cannot be repositioned).
+
+    Teams are pulled out of the list and re-inserted immediately below their
+    divider, sorted so they read alphabetically top-to-bottom in the Discord UI.
+    Everything else keeps its relative order; a role only appears in the result
+    when its position actually changed. Roles are matched by name, mirroring the
+    rest of the team sync - with duplicate role names the lowest one wins.
+    """
+    present = set(ordered_names)
+    placeable: dict[str, list[str]] = {}
+    for divider, teams in placements.items():
+        if divider not in present or divider in managed:
+            continue
+        # Descending here because the list is bottom-up: reversing it renders
+        # alphabetically downwards from the divider.
+        block = sorted({t for t in teams if t in present and t not in managed}, reverse=True)
+        if block:
+            placeable[divider] = block
+    if not placeable:
+        return {}
+
+    movable = {t for block in placeable.values() for t in block}
+    remaining = [n for n in ordered_names if n not in movable]
+    for divider, block in placeable.items():
+        idx = remaining.index(divider)
+        remaining[idx:idx] = block
+
+    old_pos = {name: i for i, name in enumerate(ordered_names)}
+    return {
+        name: i
+        for i, name in enumerate(remaining)
+        if name not in managed and old_pos.get(name) != i
+    }
+
+
+async def apply_team_role_positions(
+    guild: discord.Guild,
+    log: logging.Logger = log,
+) -> tuple[int, list[str]]:
+    """Move every team role directly beneath its configured divider role.
+
+    Driven by the ``TEAM_DIVIDER_*`` config: teams whose sheet section matched
+    one are placed under that divider role, the rest are left exactly where they
+    are. Dividers are configured by role ID, so a rename never breaks placement;
+    the ID is resolved to the role's current name here, and the ordering maths
+    downstream is name-based. Returns ``(roles_moved, errors)``; a no-op returns
+    ``(0, [])``.
+    """
+    team_divider_ids = team_dividers()  # {team: divider role ID}, both sources
+    if not team_divider_ids:
+        return 0, []  # no dividers configured, or nothing matched
+
+    errors: list[str] = []
+    # Resolve each configured divider ID to its role once, then key placements by
+    # that role's current name for the name-based positioning maths below.
+    placements: dict[str, list[str]] = {}
+    resolved: dict[int, discord.Role | None] = {}
+    for team, divider_id in team_divider_ids.items():
+        if divider_id not in resolved:
+            resolved[divider_id] = guild.get_role(divider_id)
+            if resolved[divider_id] is None:
+                errors.append(f"Eraldajarolli ID {divider_id} ei leitud serverist")
+        divider = resolved[divider_id]
+        if divider is not None:
+            placements.setdefault(divider.name, []).append(team)
+    if not placements:
+        return 0, errors
+
+    by_name: dict[str, discord.Role] = {}
+    for role in sorted(guild.roles, key=lambda r: r.position):
+        by_name.setdefault(role.name, role)
+
+    # The bot can only reorder roles strictly below its own highest role.
+    bot_top = max((r.position for r in guild.me.roles), default=0)
+    for divider in list(placements):
+        role = by_name.get(divider)
+        if role is None:
+            errors.append(f"Eraldajarolli '{divider}' ei leitud serverist")
+            placements.pop(divider)
+        elif role.position >= bot_top:
+            errors.append(f"Eraldaja '{divider}' on boti rollist kõrgemal - ei saa liigutada")
+            placements.pop(divider)
+    if not placements:
+        return 0, errors
+
+    ordered = sorted(guild.roles, key=lambda r: r.position)
+    ordered_names = [r.name for r in ordered]
+    managed = {r.name for r in ordered if r.managed or r.position >= bot_top}
+    plan = plan_team_positions(ordered_names, managed, placements)
+    if not plan:
+        return 0, errors  # already in the right place
+
+    positions = {by_name[n]: p for n, p in plan.items() if n in by_name}
+    try:
+        await guild.edit_role_positions(positions=positions)
+    except discord.Forbidden:
+        errors.append("Tiimirollide järjestamiseks puudub õigus")
+        return 0, errors
+    except discord.HTTPException as e:
+        errors.append(f"Tiimirollide järjestamine ebaõnnestus: {e}")
+        # 50013 here despite Manage Roles usually means a role in the batch sits
+        # at/above the bot's top role. Log the batch vs bot_top to pinpoint it.
+        log.warning(
+            "edit_role_positions failed (%s); bot_top=%d; batch=%s",
+            e, bot_top,
+            sorted(
+                ((r.name, r.position, target) for r, target in positions.items()),
+                key=lambda x: -x[1],
+            ),
+        )
+        return 0, errors
+    log.info("Positioned %d team role(s) under their dividers", len(positions))
+    return len(positions), errors
+
+
 async def sync_all_team_roles(
     guild: discord.Guild,
     log: logging.Logger = log,
@@ -332,13 +518,26 @@ async def sync_all_team_roles(
             summary.assigned += 1
         if res.removed:
             summary.removed += len(res.removed)
+        if res.divider_added:
+            summary.divider_assigned += 1
+        if res.divider_removed:
+            summary.divider_removed += len(res.divider_removed)
         if res.changed:
             bits: list[str] = []
             if res.added:
                 bits.append(f"+{res.added}")
             if res.removed:
                 bits.append("-" + ", -".join(res.removed))
+            if res.divider_added:
+                bits.append(f"+[{res.divider_added}]")
+            if res.divider_removed:
+                bits.append("-[" + "], -[".join(res.divider_removed) + "]")
             summary.changes.append(f"{member.display_name}: {', '.join(bits)}")
+
+    # Placement runs after the grant/remove pass so roles created this run are
+    # positioned in the same sweep rather than waiting for the next one.
+    summary.positioned, position_errors = await apply_team_role_positions(guild, log)
+    summary.errors.extend(position_errors)
     return summary
 
 

core/sheets.py

@@ -9,6 +9,7 @@ Pure-cache helpers (get_cache, find_*) remain sync.
 import asyncio
 import logging
 import re
+from dataclasses import dataclass, field
 
 import gspread
 from google.oauth2.service_account import Credentials
@@ -279,6 +280,19 @@ _LINEUP_HEADER_PREFIX = "lineup"
 _team_roster: dict[str, list[str]] = {}   # team name -> [Discord username, ...]
 _team_by_username: dict[str, str] = {}     # normalized username -> team name
 _team_names: set[str] = set()              # universe of all team names
+_team_divider: dict[str, int] = {}         # team name -> divider role ID
+
+
+@dataclass(frozen=True)
+class TeamSection:
+    """One game's block of teams within a tab, plus the title row above it.
+
+    The title ("TipiLAN 2026 CS2 Registration Log") is the only thing that says
+    which game and year a block belongs to - the header and team rows below it
+    carry neither - so it is what :func:`resolve_divider` matches against.
+    """
+    title: str = ""
+    rosters: dict[str, list[str]] = field(default_factory=dict)
 
 
 def parse_lineup(cell: str) -> list[str]:
@@ -321,22 +335,41 @@ def _cell(row: list, idx: int) -> str:
     return str(row[idx]) if 0 <= idx < len(row) else ""
 
 
-def parse_team_rosters(rows: list[list]) -> dict[str, list[str]]:
+def _merged_title(row: list) -> str | None:
-    """Extract {team_name: [nickname, ...]} from a tab's raw rows.
+    """Return a row's lone non-empty cell - a merged section title - else None.
+
+    A fully blank row returns None rather than "", so blank separators between
+    sections do not wipe the title we are holding for the next header row.
+    """
+    values = [str(c).strip() for c in row if str(c).strip()]
+    return values[0] if len(values) == 1 else None
+
+
+def parse_team_sections(rows: list[list]) -> list[TeamSection]:
+    """Extract each game's block of teams from a tab's raw rows, with its title.
 
     Scans for every header row that has both a 'Team Name' and a 'Lineup...'
     column, then reads the rows beneath it (using that section's own column
     positions) until the team-name column goes blank or a new header appears.
-    Handles multiple stacked sections with differing layouts in one tab.
+    Handles multiple stacked sections with differing layouts in one tab, and
+    tags each with the most recent merged title row seen above it.
     """
-    rosters: dict[str, list[str]] = {}
+    sections: list[TeamSection] = []
+    title = ""
     i, n = 0, len(rows)
     while i < n:
         name_col = _find_col(rows[i], lambda c: c.strip().lower() == _TEAM_NAME_HEADER)
         lineup_col = _find_col(rows[i], lambda c: c.strip().lower().startswith(_LINEUP_HEADER_PREFIX))
         if name_col is None or lineup_col is None:
+            if (text := _merged_title(rows[i])) is not None:
+                # Accumulate every single-cell row above the header, not just the
+                # last one: sheets stack the "TipiLAN 2026 CS2" title above notice
+                # rows ("If a team withdraws..."), and the later notices must not
+                # clobber the title whose keywords resolve_divider needs.
+                title = f"{title} {text}".strip() if title else text
             i += 1
             continue
+        rosters: dict[str, list[str]] = {}
         i += 1  # move past the header into the data block
         while i < n:
             team = _cell(rows[i], name_col).strip()
@@ -346,9 +379,44 @@ def parse_team_rosters(rows: list[list]) -> dict[str, list[str]]:
             if players:
                 rosters.setdefault(team, []).extend(players)
             i += 1
+        if rosters:
+            sections.append(TeamSection(title=title, rosters=rosters))
+        title = ""  # consumed - do not leak it onto the next section
+    return sections
+
+
+def parse_team_rosters(rows: list[list]) -> dict[str, list[str]]:
+    """Flatten every section in a tab into {team_name: [nickname, ...]}."""
+    rosters: dict[str, list[str]] = {}
+    for section in parse_team_sections(rows):
+        for team, players in section.rosters.items():
+            rosters.setdefault(team, []).extend(players)
     return rosters
 
 
+def resolve_divider(title: str, dividers: dict[str, int] | None = None) -> int | None:
+    """Return the divider role ID configured for a section title, if any.
+
+    A ``TEAM_DIVIDER_<SUFFIX>`` entry matches when every underscore-separated
+    part of its suffix appears as a whole word in the title, so ``CS2`` matches
+    a CS2 section from any year while ``CS2_2026`` matches only the 2026 one.
+    The most specific match (most parts) wins, which lets a year-scoped entry
+    override a general one for the same game.
+    """
+    if dividers is None:
+        dividers = config.TEAM_DIVIDERS
+    haystack = title.lower()
+    best_id: int | None = None
+    best_parts = 0
+    for suffix, role_id in dividers.items():
+        parts = [p for p in suffix.split("_") if p]
+        if not parts or len(parts) <= best_parts:
+            continue
+        if all(re.search(rf"\b{re.escape(p)}\b", haystack) for p in parts):
+            best_id, best_parts = role_id, len(parts)
+    return best_id
+
+
 def build_username_index(rosters: dict[str, list[str]]) -> dict[str, str]:
     """Invert {team: [usernames]} into {normalized username: team}.
 
@@ -371,19 +439,28 @@ def build_username_index(rosters: dict[str, list[str]]) -> dict[str, str]:
 
 
 def _refresh_teams_sync() -> dict[str, list[str]]:
-    global _team_roster, _team_by_username, _team_names
+    global _team_roster, _team_by_username, _team_names, _team_divider
     creds = Credentials.from_service_account_file(config.GOOGLE_CREDS_PATH, scopes=SCOPES)
     client = gspread.authorize(creds)
     spreadsheet = client.open_by_key(config.TEAM_SHEET_ID)
 
     rosters: dict[str, list[str]] = {}
+    dividers: dict[str, int] = {}
     for ws in spreadsheet.worksheets():
-        for team, players in parse_team_rosters(ws.get_all_values()).items():
+        for section in parse_team_sections(ws.get_all_values()):
-            rosters.setdefault(team, []).extend(players)
+            # Match on the tab name too ("CS2"/"LoL"): it names the game reliably
+            # even when the game/year title row is shadowed by notice rows, while
+            # the section title still supplies the year for year-scoped dividers.
+            divider = resolve_divider(f"{ws.title} {section.title}")
+            for team, players in section.rosters.items():
+                rosters.setdefault(team, []).extend(players)
+                if divider:
+                    dividers[team] = divider
 
     _team_roster = rosters
     _team_by_username = build_username_index(rosters)
     _team_names = set(rosters)
+    _team_divider = dividers
     return rosters
 
 
@@ -411,3 +488,12 @@ def all_team_names() -> set[str]:
 def get_team_rosters() -> dict[str, list[str]]:
     """Current {team: [usernames]} cache (mainly for diagnostics/tests)."""
     return _team_roster
+
+
+def get_team_dividers() -> dict[str, int]:
+    """Current {team: divider role ID} cache.
+
+    Only teams whose section title matched a configured TEAM_DIVIDER_* entry
+    appear here, so an empty dict means positioning is switched off.
+    """
+    return dict(_team_divider)

docs/DEV_NOTES.md

@@ -19,7 +19,7 @@ The codebase is split into **`core/`** (domain logic), **`commands/`** (Discord
 | `core/economy/` | Economy business logic **package**, re-exported via `core/economy/__init__.py` so callers use `from core import economy` + attribute access (`economy.do_daily`, `economy.SHOP`, ...). Submodules: `store.py` (user records, per-user locks, `COOLDOWNS`, `JAIL_DURATION`, `COIN`, `get_user`/`_commit`/`_txn`), `income.py`, `gambling.py`, `fishing.py`, `jail.py`, `heist.py`, `prestige.py`, `shop.py`, `levels.py`, `quests.py`, `leaderboards.py`, `house.py`, `admin.py` |
 | `core/pb_client.py` | Async PocketBase REST client - auth token cache, CRUD on `economy_users` collection |
 | `core/sheets.py` | Google Sheets integration (member sync) |
-| `core/member_sync.py` | Birthday/member sync helpers |
+| `core/member_sync.py` | Birthday/member sync helpers, plus tournament team-role sync + divider placement |
 
 ### `commands/` - one slash-command group per file
 
@@ -148,6 +148,37 @@ Gambling EXP is bet-scaled via `gamble_exp(bet)`; fish EXP is per-species in `FI
 
 ---
 
+## Tournament Team Roles
+
+Economy profile only. Roster-independent: matches Discord usernames straight against `TEAM_SHEET_ID`, never the member sheet. Entry points are `/teamsync` (`commands/economy_team_commands.py`) and the hourly `team_sync_hourly` task in `bot.py`; both call `sheets.refresh_teams()` then `member_sync.sync_all_team_roles()`.
+
+### Pipeline
+
+| Step | Where | Notes |
+|---|---|---|
+| Split a tab into per-game blocks | `sheets.parse_team_sections` | Returns `TeamSection(title, rosters)`. The merged title row above each header is the **only** thing identifying game + year |
+| Flatten to `{team: [players]}` | `sheets.parse_team_rosters` | Thin wrapper over `parse_team_sections` |
+| Section title → divider role name | `sheets.resolve_divider` | Matches `config.TEAM_DIVIDERS`; most specific key wins |
+| Cache | `sheets._refresh_teams_sync` | Populates `_team_roster`, `_team_by_username`, `_team_names`, `_team_divider` |
+| Grant/remove roles per member | `member_sync.sync_team_role` | Only names in `all_team_names()` are ever touched |
+| Compute new role positions | `member_sync.plan_team_positions` | **Pure** - list of names ascending + managed set + placements → `{name: new position}`. Unit-tested without Discord |
+| Apply | `member_sync.apply_team_role_positions` | One `guild.edit_role_positions` call per sync |
+
+### Adding a game to the divider config
+
+1. `.env` - add `TEAM_DIVIDER_<GAME>_<YEAR>="<exact Discord role name>"`. No code change; `config._parse_team_dividers` discovers any var with that prefix at startup.
+2. Restart the bot (env is read once at import).
+
+### Position maths gotchas
+
+- Discord positions are **ascending from the bottom** (`@everyone` = 0), so "under the divider in the UI" means a *lower* number. `plan_team_positions` therefore inserts each block reverse-sorted.
+- The bot can only reorder roles strictly below its own highest role. Dividers at or above it are skipped with an error rather than attempted.
+- `managed` roles (integration/bot/booster) are never emitted in the plan.
+- Roles are matched **by name** throughout the feature; with duplicate names the lowest-positioned one wins.
+- Only roles whose position actually changed are submitted, so a settled guild costs zero API calls.
+
+---
+
 ## Role Hierarchy (Discord)
 
 Order top to bottom in server roles:

strings/member.py

@@ -109,6 +109,9 @@ TEAMSYNC_UI: dict[str, str] = {
     "assigned":       "✅ Tiimirolle antud: {count}",
     "removed":        "➖ Tiimirolle eemaldatud: {count}",
     "created":        "🆕 Loodud uusi tiimirolle: {roles}",
+    "positioned":     "📍 Eraldaja alla paigutatud: {count}",
+    "divider_assigned": "🏷️ Mängurolle (eraldaja) antud: {count}",
+    "divider_removed":  "➖ Mängurolle (eraldaja) eemaldatud: {count}",
     "errors":         "⚠️ Vead: {count}",
     "no_changes":     "✨ Kõik tiimirollid olid juba korras.",
     "changes_header": "**Muudatused:**",

tests/test_fienta.py

@@ -0,0 +1,101 @@
+"""Tests for the Fienta ticket parsing that feeds team-role sync.
+
+Covers the risky bit: turning raw Fienta ticket JSON into a reliable
+{discord identity -> team} + {team -> game} mapping, including which ticket
+types count as team members and how the game is detected.
+"""
+
+from __future__ import annotations
+
+import sys
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).parent.parent))
+
+import config  # noqa: E402
+from core import fienta  # noqa: E402
+
+
+def _ticket(ttype: str, *, team="", discord="", discord_id="", nick=""):
+    return {
+        "order_id": 1,
+        "rows": [{
+            "ticket_type": {"title": ttype},
+            "attendee": {
+                fienta._F_TEAM_NAME: team,
+                fienta._F_DISCORD_USERNAME: discord,
+                fienta._F_DISCORD_USERID: discord_id,
+                fienta._F_DISCORD_USERNAME.replace("discord", "x"): "",
+                "nickname_134815": nick,
+            },
+        }],
+    }
+
+
+CS2 = "Counter-Strike 2 Tournament - competitor ticket"
+LOL = "League of Legends Tournament - competitor ticket"
+
+
+def test_parse_maps_username_and_userid_to_team():
+    fienta.parse_tickets([
+        _ticket(CS2, team="KONE", discord="ar7enchik", discord_id="123"),
+    ])
+    assert fienta.get_team_for_username("ar7enchik") == "KONE"
+    assert fienta.get_team_for_username("AR7ENCHIK") == "KONE"  # case-insensitive
+    assert fienta.get_team_for_userid(123) == "KONE"
+    assert fienta.get_team_game("KONE") == "CS2"
+
+
+def test_parse_detects_game_from_ticket_type():
+    fienta.parse_tickets([
+        _ticket(CS2, team="KONE", discord="a"),
+        _ticket(LOL, team="Ööbik", discord="b"),
+    ])
+    assert fienta.get_team_game("KONE") == "CS2"
+    assert fienta.get_team_game("Ööbik") == "LoL"
+
+
+def test_parse_excludes_non_player_ticket_types():
+    fienta.parse_tickets([
+        _ticket("Visitor's Ticket", team="", discord=""),
+        _ticket("Early Bird - visitor ticket", team="X", discord="ghost"),
+        _ticket("Counter-Strike 2 Tournament Waiting List", team="WL", discord="waiter"),
+        _ticket("LAN area - Access Ticket", team="", discord=""),
+        _ticket(CS2, team="KONE", discord="real"),
+    ])
+    assert fienta.all_team_names() == {"KONE"}
+    assert fienta.get_team_for_username("ghost") is None
+    assert fienta.get_team_for_username("waiter") is None
+    assert fienta.get_team_for_username("real") == "KONE"
+
+
+def test_parse_includes_coach_and_substitute():
+    fienta.parse_tickets([
+        _ticket("Counter-Strike 2 Coach/Manager - competitor ticket", team="KONE", discord="coach"),
+        _ticket("CS2 Substitute Player - competitor ticket", team="KONE", discord="sub"),
+    ])
+    assert fienta.get_team_for_username("coach") == "KONE"
+    assert fienta.get_team_for_username("sub") == "KONE"
+
+
+def test_parse_skips_tickets_without_team_or_discord():
+    fienta.parse_tickets([
+        _ticket(CS2, team="", discord="noteam"),     # no team -> skipped
+        _ticket(CS2, team="KONE", discord=""),        # team but no discord -> team known, no user
+    ])
+    assert fienta.get_team_for_username("noteam") is None
+    assert "KONE" in fienta.all_team_names()
+
+
+def test_get_team_dividers_maps_via_game(monkeypatch):
+    monkeypatch.setattr(config, "TEAM_DIVIDERS", {"cs2_2026": 100, "lol_2026": 200})
+    fienta.parse_tickets([
+        _ticket(CS2, team="KONE", discord="a"),
+        _ticket(LOL, team="Ööbik", discord="b"),
+    ])
+    assert fienta.get_team_dividers() == {"KONE": 100, "Ööbik": 200}
+
+
+def test_strips_leading_at_from_discord_username():
+    fienta.parse_tickets([_ticket(CS2, team="KONE", discord="@handle")])
+    assert fienta.get_team_for_username("handle") == "KONE"

tests/test_team_sync.py

@@ -13,6 +13,7 @@ from pathlib import Path
 
 sys.path.insert(0, str(Path(__file__).parent.parent))
 
+import config  # noqa: E402
 from core import member_sync, sheets  # noqa: E402
 from tests.conftest import run  # noqa: E402
 
@@ -87,6 +88,124 @@ def test_parse_team_rosters_ignores_non_table_content():
     assert sheets.parse_team_rosters([["just", "some", "prose"], ["more"]]) == {}
 
 
+# --- parse_team_sections (which game/year a block of teams came from) ------
+
+def test_parse_team_sections_tags_each_block_with_its_title():
+    sections = sheets.parse_team_sections(SHEET_ROWS)
+    assert [(s.title, sorted(s.rosters)) for s in sections] == [
+        ("[merged] TipiLAN 2026 CS2 Registration Log", ["GENESIS", "Piirivalvurid"]),
+        ("[merged] TipiLAN 2026 LoL Registration Log", ["Pushing 30s"]),
+    ]
+
+
+def test_parse_team_sections_title_does_not_leak_to_untitled_block():
+    rows = [
+        ["[merged] TipiLAN 2026 CS2 Registration Log"] + [""] * 2,
+        ["No", "Team Name", "Lineup (nickname)"],
+        ["1", "GENESIS", "kapa (EST)"],
+        ["", "", ""],
+        ["No", "Team Name", "Lineup (nickname)"],  # second block, no title above it
+        ["1", "Nameless", "someone (EST)"],
+    ]
+    titles = [s.title for s in sheets.parse_team_sections(rows)]
+    assert titles == ["[merged] TipiLAN 2026 CS2 Registration Log", ""]
+
+
+# --- resolve_divider (section title -> configured divider role) ------------
+
+DIVIDERS = {
+    "cs2_2026": 1498736834656604251,
+    "lol_2026": 1498736949706490017,
+}
+
+
+def test_resolve_divider_matches_game_and_year():
+    assert sheets.resolve_divider(
+        "[merged] TipiLAN 2026 CS2 Registration Log", DIVIDERS
+    ) == 1498736834656604251
+    assert sheets.resolve_divider(
+        "[merged] TipiLAN 2026 LoL Registration Log", DIVIDERS
+    ) == 1498736949706490017
+
+
+def test_resolve_divider_year_scoped_key_ignores_other_years():
+    assert sheets.resolve_divider("TipiLAN 2025 CS2 Registration Log", DIVIDERS) is None
+
+
+def test_resolve_divider_prefers_the_most_specific_match():
+    dividers = {"cs2": 111, "cs2_2026": 222}
+    assert sheets.resolve_divider("TipiLAN 2026 CS2 Log", dividers) == 222
+    assert sheets.resolve_divider("TipiLAN 2025 CS2 Log", dividers) == 111
+
+
+def test_resolve_divider_no_config_or_no_title_is_none():
+    assert sheets.resolve_divider("TipiLAN 2026 CS2 Log", {}) is None
+    assert sheets.resolve_divider("", DIVIDERS) is None
+
+
+def test_title_survives_notice_rows_below_it():
+    # Real sheets stack the game/year title above single-cell notice rows; the
+    # title's keywords must survive so resolve_divider still matches (regression:
+    # notice rows used to overwrite the title, yielding a None divider).
+    rows = [
+        ["TipiLAN 2026 CS2 Registration Log"] + [""] * 4,
+        ["This log is updated automatically."] + [""] * 4,
+        ["If a team from the Top 32 withdraws, ..."] + [""] * 4,
+        ["No", "Team Name", "Lineup (nickname, citizenship)", "", ""],
+        ["1", "GENESIS", "kapa (EST)", "", ""],
+    ]
+    section = sheets.parse_team_sections(rows)[0]
+    assert "CS2" in section.title and "2026" in section.title
+    assert sheets.resolve_divider(section.title, DIVIDERS) == 1498736834656604251
+
+
+# --- plan_team_positions (pure role-ordering maths) ------------------------
+
+def test_plan_moves_teams_directly_under_their_divider():
+    # Ascending = bottom-up, so the Discord UI renders this list reversed.
+    ordered = ["@everyone", "60hz", "999", "Raid", "CS2-2026", "LOL-2026", "Bot"]
+    plan = member_sync.plan_team_positions(
+        ordered, {"Bot"}, {"CS2-2026": ["60hz", "999", "Raid"]}
+    )
+    final = sorted(set(ordered), key=lambda n: plan.get(n, ordered.index(n)))
+    # Teams end up just below the divider, alphabetical downwards in the UI.
+    assert final == ["@everyone", "Raid", "999", "60hz", "CS2-2026", "LOL-2026", "Bot"]
+
+
+def test_plan_groups_each_divider_separately():
+    ordered = ["@everyone", "Alpha", "Pushing 30s", "CS2-2026", "LOL-2026", "Bot"]
+    plan = member_sync.plan_team_positions(
+        ordered,
+        {"Bot"},
+        {"CS2-2026": ["Alpha"], "LOL-2026": ["Pushing 30s"]},
+    )
+    final = sorted(set(ordered), key=lambda n: plan.get(n, ordered.index(n)))
+    assert final == ["@everyone", "Alpha", "CS2-2026", "Pushing 30s", "LOL-2026", "Bot"]
+
+
+def test_plan_is_a_noop_when_already_positioned():
+    ordered = ["@everyone", "Raid", "999", "60hz", "CS2-2026", "Bot"]
+    assert member_sync.plan_team_positions(
+        ordered, {"Bot"}, {"CS2-2026": ["60hz", "999", "Raid"]}
+    ) == {}
+
+
+def test_plan_skips_missing_divider_and_leaves_its_teams_alone():
+    ordered = ["@everyone", "60hz", "CS2-2026", "Bot"]
+    plan = member_sync.plan_team_positions(
+        ordered, {"Bot"}, {"LOL-2026": ["60hz"]}  # divider not in the guild
+    )
+    assert plan == {}
+
+
+def test_plan_never_moves_managed_roles():
+    ordered = ["@everyone", "60hz", "CS2-2026", "Nitro", "Bot"]
+    plan = member_sync.plan_team_positions(
+        ordered, {"Nitro", "Bot"}, {"CS2-2026": ["60hz", "Nitro"]}
+    )
+    assert "Nitro" not in plan and "Bot" not in plan
+
+
 def test_build_username_index_is_case_insensitive():
     index = sheets.build_username_index({"GENESIS": ["Kapa", "neaQ"]})
     assert index == {"kapa": "GENESIS", "neaq": "GENESIS"}
@@ -135,6 +254,9 @@ class FakeGuild:
         self.created.append(name)
         return role
 
+    def get_role(self, rid):
+        return next((r for r in self.roles if r.id == rid), None)
+
 
 def test_sync_creates_missing_team_role_and_removes_old_one(monkeypatch):
     old_team = FakeRole(1, "OldTeam")
@@ -174,6 +296,43 @@ def test_sync_uses_existing_team_role(monkeypatch):
     assert genesis in member.roles
 
 
+def test_sync_grants_game_divider_role(monkeypatch):
+    genesis = FakeRole(3, "GENESIS")
+    cs2_div = FakeRole(100, "====== CS2 2026 ======")
+    member = FakeMember(1, "kapa", roles=[])
+    guild = FakeGuild([genesis, cs2_div])
+
+    monkeypatch.setattr(sheets, "get_team_for_username", lambda n: "GENESIS")
+    monkeypatch.setattr(sheets, "all_team_names", lambda: {"GENESIS"})
+    monkeypatch.setattr(sheets, "get_team_dividers", lambda: {"GENESIS": 100})
+    monkeypatch.setattr(config, "TEAM_DIVIDERS", {"cs2_2026": 100, "lol_2026": 200})
+
+    result = run(member_sync.sync_team_role(member, guild))
+
+    assert result.divider_added == "====== CS2 2026 ======"
+    assert cs2_div in member.roles
+
+
+def test_sync_swaps_divider_role_on_game_switch(monkeypatch):
+    genesis = FakeRole(3, "GENESIS")
+    cs2_div = FakeRole(100, "CS2")
+    lol_div = FakeRole(200, "LoL")
+    member = FakeMember(1, "kapa", roles=[lol_div])  # was LoL, now on a CS2 team
+    guild = FakeGuild([genesis, cs2_div, lol_div])
+
+    monkeypatch.setattr(sheets, "get_team_for_username", lambda n: "GENESIS")
+    monkeypatch.setattr(sheets, "all_team_names", lambda: {"GENESIS"})
+    monkeypatch.setattr(sheets, "get_team_dividers", lambda: {"GENESIS": 100})
+    monkeypatch.setattr(config, "TEAM_DIVIDERS", {"cs2_2026": 100, "lol_2026": 200})
+
+    result = run(member_sync.sync_team_role(member, guild))
+
+    assert result.divider_added == "CS2"
+    assert result.divider_removed == ["LoL"]
+    ids = {r.id for r in member.roles}
+    assert 100 in ids and 200 not in ids
+
+
 def test_sync_strips_team_role_when_not_registered(monkeypatch):
     old_team = FakeRole(1, "OldTeam")
     member = FakeMember(1, "ghost", roles=[old_team])