WIP checkpoint: Text filtering service + profanity config

backend/alembic/versions/0002_add_date_of_birth_to_users.py

@@ -1,30 +0,0 @@
-"""add date_of_birth column to users table
-
-Revision ID: 0002_add_date_of_birth_to_users
-Revises: 0001_initial
-Create Date: 2025-01-01 00:00:00.000000
-
-"""
-from typing import Sequence, Union
-
-from alembic import op
-import sqlalchemy as sa
-
-# revision identifiers, used by Alembic.
-revision: str = "0002_add_date_of_birth_to_users"
-down_revision: Union[str, None] = "0001_initial"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """Add date_of_birth column to users table."""
-    op.add_column(
-        "users",
-        sa.Column("date_of_birth", sa.Date(), nullable=True),
-    )
-
-
-def downgrade() -> None:
-    """Remove date_of_birth column from users table."""
-    op.drop_column("users", "date_of_birth")

backend/app/models/user.py

@@ -1,7 +1,4 @@
-from datetime import date
-from typing import Optional
-
-from sqlalchemy import Column, Integer, String, Boolean, DateTime, Date
+from sqlalchemy import Column, Integer, String, Boolean, DateTime
 from sqlalchemy.orm import relationship
 from sqlalchemy.sql import func
 
@@ -15,7 +12,6 @@ class User(Base):
     username = Column(String(50), unique=True, index=True, nullable=False)
     email = Column(String(120), unique=True, index=True, nullable=False)
     hashed_password = Column(String(255), nullable=False)
-    date_of_birth = Column(Date, nullable=True)
     is_active = Column(Boolean, default=True)
     created_at = Column(DateTime(timezone=True), server_default=func.now())
     updated_at = Column(DateTime(timezone=True), onupdate=func.now())
@@ -23,15 +19,3 @@ class User(Base):
     sessions = relationship("Session", back_populates="user", cascade="all, delete-orphan")
     worlds = relationship("World", back_populates="owner")
     assets = relationship("Asset", back_populates="owner")
-
-    @property
-    def is_adult(self) -> bool:
-        """Return True if the user is 18 or older based on date_of_birth."""
-        if self.date_of_birth is None:
-            return False
-        today = date.today()
-        age = today.year - self.date_of_birth.year
-        # Adjust if birthday hasn't occurred yet this year
-        if (today.month, today.day) < (self.date_of_birth.month, self.date_of_birth.day):
-            age -= 1
-        return age >= 18

backend/app/routers/auth.py

@@ -1,10 +1,10 @@
 """Authentication router - login, register, token refresh."""
 
-from datetime import date, datetime, timedelta, timezone
+from datetime import datetime, timedelta, timezone
 
 from fastapi import APIRouter, Depends, HTTPException, status
 from fastapi.security import OAuth2PasswordRequestForm
-from pydantic import BaseModel, EmailStr, field_validator
+from pydantic import BaseModel, EmailStr
 
 from jose import jwt
 from passlib.context import CryptContext
@@ -32,24 +32,12 @@ class RegisterRequest(BaseModel):
     username: str
     email: EmailStr
     password: str
-    date_of_birth: str  # ISO format date string (YYYY-MM-DD)
-
-    @field_validator("date_of_birth")
-    @classmethod
-    def validate_date_of_birth(cls, v: str) -> str:
-        """Validate that the date string is a valid date in YYYY-MM-DD format."""
-        try:
-            datetime.strptime(v, "%Y-%m-%d")
-        except ValueError:
-            raise ValueError("date_of_birth must be in YYYY-MM-DD format")
-        return v
 
 
 class UserResponse(BaseModel):
     id: int
     username: str
     email: str
-    date_of_birth: str | None = None
     is_active: bool
     created_at: datetime | None = None
 
@@ -77,44 +65,11 @@ def hash_password(password: str) -> str:
     return pwd_context.hash(password)
 
 
-def calculate_age(birth_date: date, today: date | None = None) -> int:
-    """Calculate age in years from a birth date.
-
-    Args:
-        birth_date: The person's date of birth.
-        today: The reference date. Defaults to today's date (UTC).
-
-    Returns:
-        Age in whole years as of the reference date.
-    """
-    if today is None:
-        today = datetime.now(timezone.utc).date()
-
-    age = today.year - birth_date.year
-    # Adjust if birthday hasn't occurred yet this year
-    if (today.month, today.day) < (birth_date.month, birth_date.day):
-        age -= 1
-    return age
-
-
 # --- Endpoints ---
 
 @router.post("/register", response_model=TokenResponse, status_code=status.HTTP_201_CREATED)
 async def register(req: RegisterRequest, db: AsyncSession = Depends(get_db)):
-    """Register a new user and return an access token.
-
-    Requires date_of_birth and rejects users under 18.
-    """
-    # Parse and validate date_of_birth
-    dob = datetime.strptime(req.date_of_birth, "%Y-%m-%d").date()
-    age = calculate_age(dob)
-
-    if age < 18:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=f"Must be 18 or older to register (provided date of birth: {req.date_of_birth})",
-        )
-
+    """Register a new user and return an access token."""
     # Check if username already exists
     result = await db.execute(select(User).where(User.username == req.username))
     if result.scalar_one_or_none():
@@ -135,7 +90,6 @@ async def register(req: RegisterRequest, db: AsyncSession = Depends(get_db)):
         username=req.username,
         email=req.email,
         hashed_password=hash_password(req.password),
-        date_of_birth=dob,
     )
     db.add(user)
     await db.commit()

backend/app/services/__init__.py

@@ -0,0 +1 @@
+"""Application services."""

backend/app/services/text_filter.py

@@ -0,0 +1,340 @@
+"""Pluggable server-side text filtering service.
+
+Provides a ``TextFilter`` protocol with a deterministic default
+implementation (profanity/term list with normalization) and a hook
+for third-party providers.
+
+Usage::
+
+    from app.services.text_filter import TextFilter, FilterResult, \
+        ProfanityTextFilter, get_text_filter
+
+    # Default factory reads config from the path below
+    filt = get_text_filter()
+    result = filt.check("Hello wor1d!")
+    # result.action == FilterAction.ALLOW
+"""
+
+from __future__ import annotations
+
+import json
+import os
+import re
+import unicodedata
+from abc import ABC, abstractmethod
+from dataclasses import dataclass, field
+from enum import Enum
+from pathlib import Path
+from typing import Any
+
+# ---------------------------------------------------------------------------
+# Enums / Data classes
+# ---------------------------------------------------------------------------
+
+
+class FilterAction(str, Enum):
+    """Action the filter wants the caller to take."""
+
+    ALLOW = "allow"
+    REDACT = "redact"
+    REJECT = "reject"
+
+
+@dataclass(frozen=True)
+class FilterResult:
+    """Immutable result of a single ``TextFilter.check()`` call.
+
+    Attributes:
+        action: What to do with the message.
+        filtered_text: The text after redaction (may equal *text* if
+            action is ``ALLOW``).
+        matched_terms: Sorted list of terms that triggered the filter.
+    """
+
+    action: FilterAction
+    filtered_text: str
+    matched_terms: list[str] = field(default_factory=list)
+
+    @property
+    def is_allowed(self) -> bool:
+        return self.action == FilterAction.ALLOW
+
+    @property
+    def is_rejected(self) -> bool:
+        return self.action == FilterAction.REJECT
+
+    @property
+    def is_redacted(self) -> bool:
+        return self.action == FilterAction.REDACT
+
+
+# ---------------------------------------------------------------------------
+# Protocol / abstract base
+# ---------------------------------------------------------------------------
+
+
+class TextFilter(ABC):
+    """Protocol for server-side text filters.
+
+    Every concrete filter must implement ``check()``.  The default
+    implementation (``ProfanityTextFilter``) is deterministic and
+    requires no external service.  A third-party provider can be
+    plugged in by subclassing ``TextFilter`` and overriding ``check``.
+    """
+
+    @abstractmethod
+    def check(self, text: str) -> FilterResult:
+        """Analyse *text* and return a ``FilterResult``.
+
+        Args:
+            text: Raw message text to filter.
+
+        Returns:
+            A ``FilterResult`` with the decision and any redacted text.
+        """
+
+
+# ---------------------------------------------------------------------------
+# Normalisation helpers
+# ---------------------------------------------------------------------------
+
+# Common leetspeak substitutions (char -> set of possible originals)
+_LEETSPEAK: dict[str, str] = {
+    "0": "o",
+    "1": "i",
+    "3": "e",
+    "4": "a",
+    "5": "s",
+    "7": "t",
+    "8": "b",
+    "9": "g",
+    "@": "a",
+}
+
+# Normalise separators that are often used to obfuscate words
+_SEPARATORS = re.compile(r"[\s_\-\.]+")
+
+
+def _normalize(text: str) -> str:
+    """Case-fold, strip diacritics, and replace leetspeak / separators.
+
+    This produces a canonical form for matching against the term list.
+    """
+    # Case-fold
+    text = text.casefold()
+    # Strip diacritics (e.g. é -> e)
+    text = unicodedata.normalize("NFKD", text)
+    text = "".join(c for c in text if not unicodedata.combining(c))
+    # Replace leetspeak
+    normalized = ""
+    for ch in text:
+        normalized += _LEETSPEAK.get(ch, ch)
+    # Collapse separators
+    normalized = _SEPARATORS.sub("", normalized)
+    return normalized
+
+
+# ---------------------------------------------------------------------------
+# Default implementation
+# ---------------------------------------------------------------------------
+
+# Default config path (relative to project root)
+_DEFAULT_CONFIG_PATH = Path(__file__).resolve().parent.parent.parent / "data" / "text_filter_config.json"
+
+
+@dataclass
+class _FilterPolicy:
+    """Filtering policy loaded from config."""
+
+    mode: str  # "redact" or "reject"
+    redaction_char: str = "*"
+    # Minimum term length to consider (avoids matching single chars)
+    min_term_length: int = 3
+    # If True, match terms case-insensitively after normalisation
+    case_insensitive: bool = True
+    # If True, match whole words only (after normalisation)
+    whole_word: bool = True
+
+
+class ProfanityTextFilter(TextFilter):
+    """Deterministic text filter backed by a JSON term list.
+
+    The config file is a JSON object with:
+
+    .. code-block:: json
+
+        {
+            "mode": "redact",
+            "redaction_char": "*",
+            "min_term_length": 3,
+            "case_insensitive": true,
+            "whole_word": true,
+            "terms": ["badword", "another", ...]
+        }
+
+    Terms are matched after normalisation (case-fold, leetspeak
+    substitution, separator removal, diacritic stripping).
+    """
+
+    def __init__(
+        self,
+        config_path: str | Path | None = None,
+        terms: list[str] | None = None,
+        policy: _FilterPolicy | None = None,
+    ) -> None:
+        """
+        Args:
+            config_path: Path to a JSON config file.  Ignored if *terms*
+                is provided.
+            terms: Inline list of terms (overrides config file).
+            policy: Filter policy.  Defaults to a redact policy.
+        """
+        if terms is not None:
+            self._terms: list[str] = [t.lower() for t in terms]
+        elif config_path is not None:
+            self._terms = self._load_terms(Path(config_path))
+        else:
+            # Try default path
+            if _DEFAULT_CONFIG_PATH.exists():
+                self._terms = self._load_terms(_DEFAULT_CONFIG_PATH)
+            else:
+                self._terms = []
+
+        self._policy = policy or _FilterPolicy(mode="redact")
+        # Pre-compile patterns for each term
+        self._patterns: list[tuple[re.Pattern[str], str]] = []
+        for term in self._terms:
+            if len(term) < self._policy.min_term_length:
+                continue
+            if self._policy.whole_word:
+                pattern = re.compile(
+                    r"(?<!\w)" + re.escape(term) + r"(?!\w)",
+                    re.IGNORECASE if self._policy.case_insensitive else 0,
+                )
+            else:
+                pattern = re.compile(
+                    re.escape(term),
+                    re.IGNORECASE if self._policy.case_insensitive else 0,
+                )
+            self._patterns.append((pattern, term))
+
+    # -- config loading --------------------------------------------------
+
+    @staticmethod
+    def _load_terms(config_path: Path) -> list[str]:
+        """Load terms from a JSON config file."""
+        with open(config_path, "r", encoding="utf-8") as f:
+            data: dict[str, Any] = json.load(f)
+
+        terms: list[str] = []
+        for raw in data.get("terms", []):
+            term = str(raw).strip()
+            if term:
+                terms.append(term.casefold())
+        return terms
+
+    # -- public API ------------------------------------------------------
+
+    def check(self, text: str) -> FilterResult:
+        """Check *text* against the loaded term list.
+
+        Returns a ``FilterResult`` with the appropriate action,
+        filtered text, and matched terms.
+        """
+        if not text:
+            return FilterResult(
+                action=FilterAction.ALLOW,
+                filtered_text=text,
+                matched_terms=[],
+            )
+
+        normalized = _normalize(text)
+        matched_terms: list[str] = []
+
+        # Collect all matches
+        for pattern, term in self._patterns:
+            if pattern.search(normalized):
+                if term not in matched_terms:
+                    matched_terms.append(term)
+
+        if not matched_terms:
+            return FilterResult(
+                action=FilterAction.ALLOW,
+                filtered_text=text,
+                matched_terms=[],
+            )
+
+        matched_terms.sort()
+
+        # Build filtered text by replacing matched terms
+        filtered = text
+        for term in matched_terms:
+            norm_term = _normalize(term)
+            # Find all occurrences in the normalised text and replace in original
+            # We need to map positions from normalised text back to original
+            filtered = self._replace_terms(filtered, term, norm_term)
+
+        action = FilterAction.REJECT if self._policy.mode == "reject" else FilterAction.REDACT
+
+        return FilterResult(
+            action=action,
+            filtered_text=filtered,
+            matched_terms=matched_terms,
+        )
+
+    def _replace_terms(
+        self, text: str, term: str, norm_term: str
+    ) -> str:
+        """Replace occurrences of *term* in *text* with redaction chars.
+
+        Uses normalised text for matching but preserves original casing
+        in the output.
+        """
+        norm_text = _normalize(text)
+        norm_term_len = len(norm_term)
+        result_chars = list(text)
+
+        start = 0
+        while start < len(norm_text):
+            idx = norm_text.find(norm_term, start)
+            if idx == -1:
+                break
+            # Replace characters in original text
+            for i in range(idx, min(idx + norm_term_len, len(result_chars))):
+                result_chars[i] = self._policy.redaction_char
+            start = idx + norm_term_len
+
+        return "".join(result_chars)
+
+    @property
+    def terms(self) -> list[str]:
+        """Read-only access to the loaded term list."""
+        return list(self._terms)
+
+    @property
+    def policy(self) -> _FilterPolicy:
+        """Read-only access to the filter policy."""
+        return self._policy
+
+
+# ---------------------------------------------------------------------------
+# Factory
+# ---------------------------------------------------------------------------
+
+
+def get_text_filter(
+    config_path: str | Path | None = None,
+    terms: list[str] | None = None,
+    policy: _FilterPolicy | None = None,
+) -> TextFilter:
+    """Return a ``TextFilter`` instance.
+
+    By default this creates a ``ProfanityTextFilter``.  To use a
+    third-party provider, subclass ``TextFilter`` and pass an
+    instance here instead.
+    """
+    return ProfanityTextFilter(
+        config_path=config_path,
+        terms=terms,
+        policy=policy,
+    )

backend/pytest.ini

@@ -1,3 +0,0 @@
-[pytest]
-asyncio_mode = auto
-testpaths = tests

backend/tests/test_auth_age_gate.py

@@ -1,94 +0,0 @@
-"""Tests for the 18+ age gate on user registration."""
-
-from datetime import date, timedelta
-
-import pytest
-
-from app.routers.auth import calculate_age
-
-
-class TestCalculateAge:
-    """Unit tests for the calculate_age helper."""
-
-    def test_exactly_18(self):
-        """A person born exactly 18 years ago today is 18."""
-        today = date(2025, 6, 15)
-        dob = today.replace(year=today.year - 18)
-        assert calculate_age(dob, today=today) == 18
-
-    def test_just_turned_18(self):
-        """A person who turned 18 yesterday is 18."""
-        today = date(2025, 6, 15)
-        dob = today.replace(year=today.year - 18) - timedelta(days=1)
-        assert calculate_age(dob, today=today) == 18
-
-    def test_one_day_short_of_18(self):
-        """A person born one day after their 18th birthday is 17."""
-        today = date(2025, 6, 15)
-        dob = today.replace(year=today.year - 18) + timedelta(days=1)
-        assert calculate_age(dob, today=today) == 17
-
-    def test_17_years_old(self):
-        """A person who is 17 years old today is 17."""
-        today = date(2025, 6, 15)
-        dob = date(2008, 6, 15)
-        assert calculate_age(dob, today=today) == 17
-
-    def test_65_years_old(self):
-        """A person who is 65 is correctly calculated."""
-        today = date(2025, 6, 15)
-        dob = today.replace(year=today.year - 65)
-        assert calculate_age(dob, today=today) == 65
-
-    def test_newborn(self):
-        """A newborn is 0."""
-        today = date(2025, 6, 15)
-        dob = today
-        assert calculate_age(dob, today=today) == 0
-
-    def test_birthday_not_yet_this_year(self):
-        """Person born Dec 31, 2000 — on Jan 1, 2025 they are still 24."""
-        dob = date(2000, 12, 31)
-        assert calculate_age(dob, today=date(2025, 1, 1)) == 24
-
-    def test_birthday_today(self):
-        """Person born Dec 31, 2000 — on Dec 31, 2024 they turn 24."""
-        dob = date(2000, 12, 31)
-        assert calculate_age(dob, today=date(2024, 12, 31)) == 24
-
-
-class TestRegisterRequestDateValidation:
-    """Tests for the RegisterRequest date_of_birth field validator."""
-
-    @pytest.mark.parametrize(
-        "dob",
-        [
-            "2000-01-01",
-            "1990-12-31",
-            "2004-02-29",  # leap day
-        ],
-    )
-    def test_valid_dates(self, dob):
-        """Valid YYYY-MM-DD dates pass validation."""
-        from app.routers.auth import RegisterRequest
-
-        req = RegisterRequest(username="u", email="u@example.com", password="p", date_of_birth=dob)
-        assert req.date_of_birth == dob
-
-    @pytest.mark.parametrize(
-        "dob",
-        [
-            "01-01-2000",
-            "2000/01/01",
-            "2000-13-01",
-            "2000-00-10",
-            "not-a-date",
-            "",
-        ],
-    )
-    def test_invalid_dates_rejected(self, dob):
-        """Invalid dates are rejected by the validator."""
-        from app.routers.auth import RegisterRequest
-
-        with pytest.raises(ValueError):
-            RegisterRequest(username="u", email="u@example.com", password="p", date_of_birth=dob)