modules/utils.py

import os, re, json, base64, hmac, hashlib
from pathlib import Path
from typing import Iterable, List

DATA_DIR = Path("data")
LOG_PATH = DATA_DIR / "events.jsonl"

def ensure_dirs():
    DATA_DIR.mkdir(parents=True, exist_ok=True)

def chunk_text(text: str, max_len: int=600) -> Iterable[str]:
    lines = re.split(r"(\n+)", text)
    buff, count = [], 0
    for ln in lines:
        if count + len(ln) > max_len and buff:
            yield "".join(buff).strip()
            buff, count = [], 0
        buff.append(ln)
    if buff:
        yield "".join(buff).strip()

def truncate(text: str, n: int) -> str:
    return text if len(text) <= n else text[:n] + "..."

def log_event(event_type: str, payload: dict, meta: dict | None=None):
    ensure_dirs()
    rec = {"type": event_type, "payload": payload, "meta": meta or {}}
    with open(LOG_PATH, "a", encoding="utf-8") as f:
        f.write(json.dumps(rec, ensure_ascii=False) + "\n")

def verify_tracking_token(token: str):
    try:
        raw = base64.urlsafe_b64decode(token.encode())
        data, sig = raw.rsplit(b".",1)
        secret = os.getenv("TRACKING_SECRET", "dev").encode()
        expected = hmac.new(secret, data, hashlib.sha256).digest()
        if not hmac.compare_digest(sig, expected):
            return None
        return json.loads(data.decode())
    except Exception:
        return None