python/hubspot_billing.py

"""
HubSpot Billing → Supabase (incremental since a millisecond cursor)

Usage from orchestrator:
    import hubspot_billing
    hubspot_billing.main(since_ms=<int milliseconds since epoch UTC>)

Direct CLI:
    # epoch ms
    python hubspot_billing.py 1754025600000
    # ISO-8601
    python hubspot_billing.py 2025-08-01T09:30:00Z
    # Back-compat date (floors to 00:00Z)
    python hubspot_billing.py 2025-08-01
"""
import os
import re
import time
import logging
import datetime
from typing import List, Dict, Optional, Tuple, Union

import httpx
import hubspot
from dotenv import load_dotenv
from supabase import create_client

from hubspot_utils import (
    parse_ts, try_parse_int, deduplicate_by_key,
)
from supabase_utils import (
    batched_insert, update_sync_metadata,
)

# -----------------------------------------------------------------------------
# Logging
# -----------------------------------------------------------------------------
logging.basicConfig(
    filename=f"logs/hubspot_billing_pipeline_{datetime.datetime.now().strftime('%Y-%m-%d')}.log",
    filemode="a",
    level=logging.INFO,
    format="%(asctime)s [%(levelname)s] %(message)s",
)

# -----------------------------------------------------------------------------
# Environment
# -----------------------------------------------------------------------------
load_dotenv()
HUBSPOT_TOKEN = os.getenv("HUBSPOT_TOKEN")
SUPABASE_URL = os.getenv("SUPABASE_URL")
SUPABASE_SERVICE_ROLE_KEY = os.getenv("SUPABASE_SERVICE_ROLE_KEY")
# Optional bootstrap cursor if orchestrator doesn't provide one
BOOTSTRAP_SINCE_MS_ENV = os.getenv("HUBSPOT_BILLING_SINCE_MS")

if not HUBSPOT_TOKEN:
    raise RuntimeError("HUBSPOT_TOKEN is not set")
if not SUPABASE_URL or not SUPABASE_SERVICE_ROLE_KEY:
    raise RuntimeError("Supabase env vars are not set")

hubspot_client = hubspot.Client.create(access_token=HUBSPOT_TOKEN)
supabase_client = create_client(SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY)

# -----------------------------------------------------------------------------
# Config
# -----------------------------------------------------------------------------

# Custom object type ID for Billing Services
BILLING_OBJECT_TYPE = "2-38060359"

BILLING_PROPERTIES = [
    "cli",
    "account_code",
    "service_name",
    "status",
    "type",
    "tariff_name",
    "contract_renewal_date",
    "supplier_network",
    "network_name",
    "product_type_name",
    "hs_created_by_user_id",
    "hs_createdate",
    "connection_date",
    "disconnect_date"
]

PROPERTY_RENAME = {
    "hs_created_by_user_id": "hubspot_created_by",
    "hs_createdate": "hubspot_created_at",
}

# -----------------------------------------------------------------------------
# Time helpers (mirroring hubspot_tickets)
# -----------------------------------------------------------------------------


def _ensure_utc(dt: datetime.datetime) -> datetime.datetime:
    if dt.tzinfo is None:
        dt = dt.replace(tzinfo=datetime.timezone.utc)
    return dt.astimezone(datetime.timezone.utc)


def floor_to_utc_midnight(dt: datetime.datetime) -> datetime.datetime:
    dt = _ensure_utc(dt)
    return dt.replace(hour=0, minute=0, second=0, microsecond=0)


def _parse_iso_like_to_dt(value: str) -> datetime.datetime:
    if value.endswith("Z"):
        value = value[:-1] + "+00:00"
    dt = datetime.datetime.fromisoformat(value)
    return _ensure_utc(dt)


def to_epoch_ms(dt_or_str: Union[str, datetime.datetime]) -> int:
    if isinstance(dt_or_str, str):
        dt = _parse_iso_like_to_dt(dt_or_str)
    elif isinstance(dt_or_str, datetime.datetime):
        dt = _ensure_utc(dt_or_str)
    else:
        raise TypeError(f"Unsupported type for to_epoch_ms: {type(dt_or_str)}")
    return int(dt.timestamp() * 1000)


def parse_any_ts_ms(value: Optional[Union[str, int, float]]) -> Optional[int]:
    """
    Accepts:
      - ms-epoch as str/int/float
      - seconds-epoch as str/int/float (auto *1000)
      - ISO-8601 string
    Returns ms since epoch or None.
    """
    if value is None:
        return None

    # numeric-ish string or number
    try:
        v = int(float(value))  # handles numeric strings
        # heuristics: treat 10-digit as seconds
        if v < 10_000_000_000_000:
            v *= 1000
        return v
    except Exception:
        pass

    # ISO-8601
    try:
        return to_epoch_ms(str(value))
    except Exception:
        return None


# -----------------------------------------------------------------------------
# Search (by timestamp cursor)
# -----------------------------------------------------------------------------


def _search_billing_ids_from(since_ms: int, prop: str) -> List[str]:
    """
    Search billing custom object IDs where {prop} > since_ms.
    Sort ascending so we can advance cursor monotonically.
    """
    url = f"https://api.hubapi.com/crm/v3/objects/{BILLING_OBJECT_TYPE}/search"
    headers = {
        "Authorization": f"Bearer {HUBSPOT_TOKEN}",
        "Content-Type": "application/json",
        "Accept": "application/json",
    }
    payload = {
        "filterGroups": [{
            "filters": [
                {"propertyName": prop, "operator": "GT", "value": str(since_ms)},
            ]
        }],
        "limit": 100,
        "sorts": [{"propertyName": prop, "direction": "ASCENDING"}],
        "properties": ["hs_object_id"],
    }

    ids: List[str] = []
    after: Optional[str] = None
    with httpx.Client(timeout=30.0) as client:
        while True:
            body = dict(payload)
            if after:
                body["after"] = after

            resp = client.post(url, headers=headers, json=body)
            if resp.status_code >= 400:
                try:
                    logging.error(
                        "Billing search error for prop '%s': %s", prop, resp.json())
                except Exception:
                    logging.error(
                        "Billing search error for prop '%s': %s", prop, resp.text)
                resp.raise_for_status()

            data = resp.json()
            ids.extend([obj["id"] for obj in data.get("results", []) or []])

            after = (data.get("paging") or {}).get("next", {}).get("after")
            if not after:
                break
            time.sleep(0.1)

    return ids


def search_billing_ids_after_ms(since_ms: int) -> Tuple[List[str], str]:
    """
    Search billing IDs where {prop} > since_ms (strictly greater),
    sorted ASC so the max timestamp at the end is monotonic.
    Returns a tuple of (ids, prop_used).
    """

    props_to_try = ["hs_createdate"]
    last_err = None

    for prop in props_to_try:
        try:
            ids = _search_billing_ids_from(since_ms, prop)
            logging.info(
                "Billing search with '%s' returned %d IDs.", prop, len(ids))
            return ids, prop
        except httpx.HTTPStatusError as e:
            last_err = e
            continue

    if last_err:
        raise last_err
    return [], "hs_createdate"


# -----------------------------------------------------------------------------
# Read-by-ID (with associations)
# -----------------------------------------------------------------------------


def read_billing_by_ids(
    billing_ids: List[str],
    cursor_prop: str,
) -> Tuple[List[Dict], List[Dict], List[Dict], Optional[int]]:
    """
    Read billing services by ID with properties and associations (companies/deals).
    Returns: services, company_links, deal_links, max_ts_ms_for_cursor_prop
    """
    if not billing_ids:
        return [], [], [], None

    services: List[Dict] = []
    company_links: List[Dict] = []
    deal_links: List[Dict] = []

    max_ts_ms: Optional[int] = None

    assoc_types = ["companies", "deals"]

    for i, bid in enumerate(billing_ids, start=1):
        try:
            record = hubspot_client.crm.objects.basic_api.get_by_id(
                object_type=BILLING_OBJECT_TYPE,
                object_id=bid,
                properties=BILLING_PROPERTIES,
                associations=assoc_types,
                archived=False,
            )

            p = record.properties or {}

            # Track max timestamp based on cursor_prop
            cursor_val = p.get(cursor_prop)
            ts_ms = parse_any_ts_ms(cursor_val)
            if ts_ms is not None and (max_ts_ms is None or ts_ms > max_ts_ms):
                max_ts_ms = ts_ms

            # Build service row
            row = {
                "billing_id": str(record.id),
                "cli": p.get("cli"),
                "account_code": p.get("account_code"),
                "service_name": p.get("service_name"),
                "status": p.get("status"),
                "type": p.get("type"),
                "contract_renewal_date": parse_ts(p.get("contract_renewal_date")),
                "supplier_network": p.get("supplier_network"),
                "network_name": p.get("network_name"),
                "product_type_name": p.get("product_type_name"),
                "hubspot_created_by": try_parse_int(p.get("hs_created_by_user_id")),
                "hubspot_created_at": parse_ts(p.get("hs_createdate")),
                "connection_date": parse_ts(p.get("connection_date")),
                "disconnect_date": parse_ts(p.get("disconnect_date")),
            }
            services.append(row)

            # Associations
            assoc = record.associations or {}

            if assoc.get("companies") and getattr(assoc["companies"], "results", None):
                for a in assoc["companies"].results:
                    if a.id:
                        company_links.append({
                            "billing_id": str(record.id),
                            "company_id": str(a.id),
                        })

            if assoc.get("deals") and getattr(assoc["deals"], "results", None):
                for a in assoc["deals"].results:
                    if a.id:
                        deal_links.append({
                            "billing_id": str(record.id),
                            "deal_id": str(a.id),
                        })

            if i % 200 == 0:
                logging.info("Read %d billing services...", i)

            time.sleep(0.05)

        except httpx.HTTPStatusError as e:
            logging.error("HTTP error reading billing %s: %s", bid, e)
        except Exception as e:
            logging.error("Error reading billing %s: %s", bid, e)

    return services, company_links, deal_links, max_ts_ms


# -----------------------------------------------------------------------------
# Upsert
# -----------------------------------------------------------------------------


def upsert_billing(
    services: List[Dict],
    company_links: List[Dict],
    deal_links: List[Dict],
) -> None:
    if services:
        services = deduplicate_by_key(services, key="billing_id")
        batched_insert(
            supabase_client, "hubspot_billing_services", services,
            batch_size=1000, on_conflict=["billing_id"]
        )
        print(f"Upserted {len(services)} billing services.")

    if company_links:
        company_links = deduplicate_by_key(
            company_links, key=("billing_id", "company_id"))
        batched_insert(
            supabase_client, "hubspot_billing_companies", company_links,
            batch_size=1000, on_conflict=["billing_id", "company_id"]
        )
        print(f"Upserted {len(company_links)} billing-company associations.")

    if deal_links:
        deal_links = deduplicate_by_key(
            deal_links, key=("billing_id", "deal_id"))
        batched_insert(
            supabase_client, "hubspot_billing_deals", deal_links,
            batch_size=1000, on_conflict=["billing_id", "deal_id"]
        )
        print(f"Upserted {len(deal_links)} billing-deal associations.")


# -----------------------------------------------------------------------------
# Main (timestamp cursor)
# -----------------------------------------------------------------------------


def main(since_ms: Optional[int] = None):
    """
    Orchestrates:
      1) Search billing IDs with <cursor_prop> > since_ms
      2) Read full records with associations (track max timestamp)
      3) Upsert into Supabase
      4) Update sync metadata with { last_sync_metadata, last_sync_time, cursor_prop }
    """
    # Resolve since_ms
    if since_ms is None and BOOTSTRAP_SINCE_MS_ENV:
        try:
            since_ms = int(BOOTSTRAP_SINCE_MS_ENV)
        except ValueError:
            raise RuntimeError(
                "HUBSPOT_BILLING_SINCE_MS must be an integer (ms) if set.")

    if since_ms is None:
        # Default: today@00:00:00Z for first run
        today0 = floor_to_utc_midnight(
            datetime.datetime.now(datetime.timezone.utc))
        since_ms = to_epoch_ms(today0)

    print(f"Searching billing services with timestamp > {since_ms} ...")
    ids, cursor_prop = search_billing_ids_after_ms(since_ms)
    print(f"Search property: {cursor_prop}. Found {len(ids)} billing IDs.")

    now_iso = datetime.datetime.now(datetime.timezone.utc).isoformat()

    if not ids:
        print("No billing services beyond the cursor. Updating sync metadata and exiting.")
        update_sync_metadata(supabase_client, "hubspot_billing_services", now_iso)
        return

    print("Reading billing services (with associations)...")
    services, company_links, deal_links, max_ts_ms = read_billing_by_ids(
        ids, cursor_prop)

    print("Upserting into Supabase...")
    upsert_billing(services, company_links, deal_links)

    # Advance cursor to max timestamp we actually ingested for the chosen property
    new_cursor_ms = max_ts_ms if max_ts_ms is not None else since_ms

    update_sync_metadata(supabase_client, "hubspot_billing_services", now_iso)

    print(
        f"Billing sync complete. Advanced cursor to {new_cursor_ms} using prop '{cursor_prop}'.")


# -----------------------------------------------------------------------------
# CLI
# -----------------------------------------------------------------------------


def _parse_cli_arg_to_ms(arg: str) -> int:
    """
    Accept:
      - integer epoch ms
      - ISO-8601 (Z or offset)
      - YYYY-MM-DD (floors to 00:00Z)
    """
    # epoch ms or seconds
    if re.fullmatch(r"\d{10,13}", arg):
        v = int(arg)
        if v < 10_000_000_000_000:  # seconds -> ms
            v *= 1000
        return v

    # YYYY-MM-DD
    if re.fullmatch(r"\d{4}-\d{2}-\d{2}", arg):
        d = datetime.datetime.strptime(
            arg, "%Y-%m-%d").replace(tzinfo=datetime.timezone.utc)
        return to_epoch_ms(floor_to_utc_midnight(d))

    # ISO-8601
    return to_epoch_ms(arg)


if __name__ == "__main__":
    import sys
    if len(sys.argv) > 1:
        try:
            since = _parse_cli_arg_to_ms(sys.argv[1])
        except Exception as e:
            print(
                f"Invalid timestamp. Provide epoch ms, ISO-8601, or YYYY-MM-DD. Error: {e}"
            )
            sys.exit(1)
        main(since_ms=since)
    else:
        main()