Fuugi-API / app /research.py
2Btainment's picture
Upgrade Fuugi chat, search and anonymous research
45b7965 verified
Raw History Blame Contribute Delete
6.52 kB
from __future__ import annotations
import asyncio
import os
import re
import sqlite3
from collections.abc import Callable
from datetime import UTC, date, datetime, timedelta
from enum import StrEnum
from pathlib import Path
class ResearchTopic(StrEnum):
ARTISTS_RELEASES = "artists_releases"
AUDIO_RADIO = "audio_radio"
VIDEO = "video"
SHOP_DOWNLOADS = "shop_downloads"
STUDIO_SERVICES = "studio_services"
CONTACT_BOOKING = "contact_booking"
PRIVACY = "privacy"
OTHER = "other"
class ResearchIntent(StrEnum):
INFORMATION = "information"
NAVIGATION = "navigation"
AVAILABILITY = "availability"
CONSUME = "consume"
DOWNLOAD_PURCHASE = "download_purchase"
CONTACT_COLLABORATE = "contact_collaborate"
SUPPORT = "support"
OTHER = "other"
TOPIC_RULES = (
(ResearchTopic.PRIVACY, ("datenschutz", "privacy", "speichern", "erinnerung", "memory")),
(
ResearchTopic.CONTACT_BOOKING,
("kontakt", "contact", "booking", "buchen", "einreichen", "email", "zusammenarbeit"),
),
(
ResearchTopic.STUDIO_SERVICES,
("2b studio", "studio", "service", "grafik", "produktion", "mixing", "mastering"),
),
(
ResearchTopic.SHOP_DOWNLOADS,
("shop", "merch", "kaufen", "bestellen", "download", "wav", "mp3"),
),
(ResearchTopic.VIDEO, ("2b tube", "video", "clip", "visual", "ansehen", "anschauen")),
(
ResearchTopic.AUDIO_RADIO,
("2bfy", "radio", "musik", "song", "anhören", "hoeren", "stream", "playlist"),
),
(
ResearchTopic.ARTISTS_RELEASES,
("artist", "roster", "release", "album", "ep", "track", "evil crow", "ktn", "zmureo"),
),
)
INTENT_RULES = (
(ResearchIntent.SUPPORT, ("fehler", "funktioniert nicht", "problem", "hilfe", "support")),
(
ResearchIntent.CONTACT_COLLABORATE,
("kontakt", "contact", "booking", "buchen", "einreichen", "mitmachen", "zusammenarbeiten"),
),
(
ResearchIntent.DOWNLOAD_PURCHASE,
("download", "herunterladen", "kaufen", "bestellen", "preis", "kosten"),
),
(
ResearchIntent.CONSUME,
("anhören", "hoeren", "abspielen", "ansehen", "anschauen", "streamen"),
),
(ResearchIntent.NAVIGATION, ("wo finde", "link", "seite", "öffnen", "oeffnen", "hin")),
(
ResearchIntent.AVAILABILITY,
("verfügbar", "verfuegbar", "wann", "neu", "aktuell", "online", "live"),
),
(ResearchIntent.INFORMATION, ("was", "wer", "wie", "warum", "welche", "info", "erklär")),
)
class ResearchWriteError(RuntimeError):
pass
def _normalized(text: str) -> str:
return " " + re.sub(r"[^\wäöüß]+", " ", text.casefold()).strip() + " "
def _matches(normalized: str, terms: tuple[str, ...]) -> bool:
return any(f" {_normalized(term).strip()} " in normalized for term in terms)
def classify_research(text: str) -> tuple[ResearchTopic, ResearchIntent]:
normalized = _normalized(text)
topic = next(
(category for category, terms in TOPIC_RULES if _matches(normalized, terms)),
ResearchTopic.OTHER,
)
intent = next(
(category for category, terms in INTENT_RULES if _matches(normalized, terms)),
ResearchIntent.OTHER,
)
return topic, intent
def _utc_day() -> date:
return datetime.now(UTC).date()
SCHEMA = """
CREATE TABLE IF NOT EXISTS research_daily_counts (
day_utc TEXT NOT NULL,
dimension TEXT NOT NULL CHECK (dimension IN ('topic', 'intent')),
category TEXT NOT NULL,
count INTEGER NOT NULL CHECK (typeof(count) = 'integer' AND count > 0),
PRIMARY KEY (day_utc, dimension, category),
CHECK (day_utc GLOB '[0-9][0-9][0-9][0-9]-[0-9][0-9]-[0-9][0-9]'),
CHECK (
(dimension = 'topic' AND category IN (
'artists_releases', 'audio_radio', 'video', 'shop_downloads',
'studio_services', 'contact_booking', 'privacy', 'other'
)) OR
(dimension = 'intent' AND category IN (
'information', 'navigation', 'availability', 'consume',
'download_purchase', 'contact_collaborate', 'support', 'other'
))
)
) WITHOUT ROWID;
"""
class ResearchStore:
def __init__(
self,
database_path: str,
retention_days: int = 90,
utc_day: Callable[[], date] = _utc_day,
) -> None:
self._path = Path(database_path).expanduser().resolve()
self._retention_days = retention_days
self._utc_day = utc_day
self._prepare()
def _prepare(self) -> None:
try:
self._path.parent.mkdir(parents=True, exist_ok=True, mode=0o700)
os.chmod(self._path.parent, 0o700)
descriptor = os.open(self._path, os.O_CREAT | os.O_WRONLY, 0o600)
os.close(descriptor)
os.chmod(self._path, 0o600)
with sqlite3.connect(self._path) as connection:
connection.execute("PRAGMA trusted_schema = OFF")
connection.executescript(SCHEMA)
os.chmod(self._path, 0o600)
except (OSError, sqlite3.Error) as error:
raise ResearchWriteError("Could not initialize anonymous research counters.") from error
async def increment(self, topic: ResearchTopic, intent: ResearchIntent) -> None:
try:
await asyncio.to_thread(self._increment_sync, topic, intent)
except (OSError, sqlite3.Error) as error:
raise ResearchWriteError("Could not update anonymous research counters.") from error
def _increment_sync(self, topic: ResearchTopic, intent: ResearchIntent) -> None:
current_day = self._utc_day()
cutoff = current_day - timedelta(days=self._retention_days - 1)
rows = (("topic", topic.value), ("intent", intent.value))
with sqlite3.connect(self._path, timeout=5) as connection:
connection.execute("BEGIN IMMEDIATE")
connection.execute(
"DELETE FROM research_daily_counts WHERE day_utc < ?",
(cutoff.isoformat(),),
)
connection.executemany(
"""
INSERT INTO research_daily_counts (day_utc, dimension, category, count)
VALUES (?, ?, ?, 1)
ON CONFLICT(day_utc, dimension, category)
DO UPDATE SET count = count + 1
""",
((current_day.isoformat(), dimension, category) for dimension, category in rows),
)
connection.commit()