python/hubspot_invoices.py

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

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

Direct CLI:
    # epoch ms
    python hubspot_lineitems.py 1754025600000
    # ISO-8601
    python hubspot_lineitems.py 2025-08-01T09:30:00Z
    # Back-compat date (floors to 00:00Z)
    python hubspot_lineitems.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.crm.objects import ApiException as ObjectsApiException

from hubspot_utils import (
    parse_ts, try_parse_int, try_parse_float, deduplicate_by_key
)
from supabase_utils import (
    insert_into_supabase_table, update_sync_metadata
)

# -----------------------------------------------------------------------------
# Constants
# -----------------------------------------------------------------------------
INOVICE_PROPERTIES = [
    "hs_number",
    "hs_invoice_status",
    "hs_createdate",
    "hs_invoice_date",
    "hs_amount_billed",
    "hs_balance_due",
    "hs_amount_paid",
    "hs_due_date",
    "hubspot_owner_id",
    "hs_invoice_source",
    "hs_object_source_label",
    "hs_object_source_detail_1",
    "hs_comments",
]

# -----------------------------------------------------------------------------
# Logging
# -----------------------------------------------------------------------------
logging.basicConfig(
    filename=f"logs/hubspot_invoices_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_INVOICES_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)

# -----------------------------------------------------------------------------
# Time helpers
# -----------------------------------------------------------------------------
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 isinstance(value, str) and 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 / sec-epoch / ISO-8601; returns ms since epoch or None.
    """
    if value is None:
        return None
    try:
        v = int(str(value))
        if v < 10_000_000_000_000:  # seconds → ms
            v *= 1000
        return v
    except ValueError:
        pass
    try:
        return to_epoch_ms(str(value))
    except Exception:
        logging.warning("Could not parse timestamp value=%r", value)
        return None

# -----------------------------------------------------------------------------
# Search IDs (ts > since_ms) with property fallback
# -----------------------------------------------------------------------------
def _search_invoice_ids_from(since_ms: int, prop: str) -> List[str]:
    """
    Search line ite,s where {prop} > since_ms (epoch-ms).
    Sort ascending so we can advance the cursor monotonically.
    """
    url = "https://api.hubapi.com/crm/v3/objects/invoices/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"}],
    }

    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("Invoice search error for prop '%s': %s", prop, resp.json())
                except Exception:
                    logging.error("Invoice 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_invoice_ids_after_ms(since_ms: int) -> Tuple[List[str], str]:
    """
    Try these properties in order; return (ids, prop_used) for the first successful search:
        1) hs_lastmodifieddate
        2) lastmodifieddate
        3) createdate
        4) hs_createdate
    """
    props_to_try = ["createdate", "hs_createdate"]
    last_err = None

    for prop in props_to_try:
        try:
            ids = _search_invoice_ids_from(since_ms, prop)
            logging.info("Invoice 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_lastmodifieddate"

def read_invoice_by_ids(
    invoice_ids: List[str],
    cursor_prop: str,
) -> Tuple[List[Dict], List[Dict], Optional[int]]:
    """
    Read invoices by ID including deal associations.
    Returns: (all_invoices, max_ts_ms_for_cursor_prop)
    """
    if not invoice_ids:
        return [], [], None

    all_invoices: List[Dict] = []
    invoice_contact_links: List[Dict] = []

    assoc_types = ["contacts"]

    max_ts_ms: Optional[int] = None

    for i, did in enumerate(invoice_ids, start=1):
        try:
            record = hubspot_client.crm.objects.basic_api.get_by_id(
                object_type="invoices",
                object_id=did, properties=INOVICE_PROPERTIES,
                associations=assoc_types,archived=False
            )
            p = record.properties or {}

            # Track max timestamp for the cursor property we used in the search
            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
            
            # Contacts
            if getattr(record, "associations", None) and record.associations.get("contacts"):
                bucket = record.associations["contacts"]
                if getattr(bucket, "results", None):
                    for a in bucket.results:
                        if a.id and a.id.isdigit():
                            invoice_contact_links.append({
                                "invoice_id": try_parse_int(record.id),
                                "contact_id": try_parse_int(a.id),
                            })

            # Created date: accept either createdate or hs_createdate
            created_iso = p.get("createdate") or p.get("hs_createdate")

            all_invoices.append({
                "invoice_id": try_parse_int(record.id),
                "number": p.get("hs_number"),
                "invoice_status": p.get("hs_invoice_status"),
                "hubspot_createdate": parse_ts(created_iso),
                "invoice_date": parse_ts(p.get("hs_invoice_date")),
                "due_date": parse_ts(p.get("hs_due_date")),
                "amount_billed": try_parse_float(p.get("hs_amount_billed")),
                "balance_due": try_parse_float(p.get("hs_balance_due")),
                "amount_paid": try_parse_float(p.get("hs_amount_paid")),
                "hubspot_owner_id": try_parse_int(p.get("hubspot_owner_id")),
                "invoice_source": p.get("hs_invoice_source"),
                "record_source": p.get("hs_object_source_label"),
                "record_source_detail_1": p.get("hs_object_source_detail_1"),
                "comments": p.get("hs_comments"),
            })

            if i % 200 == 0:
                logging.info("Read %d deals...", i)
            time.sleep(0.05)

        except httpx.HTTPStatusError as e:
            logging.error("HTTP error reading invoice %s: %s", did, e)
        except (ObjectsApiException, httpx.HTTPError) as e:
            logging.error("Error reading invoice %s: %s", did, e)

    return all_invoices, invoice_contact_links, max_ts_ms

# -----------------------------------------------------------------------------
# Upsert
# -----------------------------------------------------------------------------
def upsert_invoices(invoices: List[Dict], invoice_contact_links: List[Dict]) -> None:
    if invoices:
        insert_into_supabase_table(supabase_client, "hubspot_invoices", invoices)
        print(f"Upserted {len(invoices)} invoices.")

    if invoice_contact_links:
        invoice_contact_links = deduplicate_by_key(invoice_contact_links, key=("invoice_id", "contact_id"))
        insert_into_supabase_table(
            supabase_client,
            "hubspot_invoice_contacts",
            invoice_contact_links,
            on_conflict=["invoice_id", "contact_id"],
        )
        print(f"Upserted {len(invoice_contact_links)} invoice-contact associations.")
    

# -----------------------------------------------------------------------------
# Main (timestamp cursor)
# -----------------------------------------------------------------------------
def main(since_ms: Optional[int] = None):
    """
    Orchestrates:
      1) Search invoice IDs with <cursor_prop> > since_ms (property fallback)
      2) Read full invoices with associations (track max timestamp for <cursor_prop>)
      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_INVOICES_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 invoices with timestamp > {since_ms} ...")
    ids, cursor_prop = search_invoice_ids_after_ms(since_ms)
    print(f"Search property: {cursor_prop}. Found {len(ids)} invoice IDs.")

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

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

    print("Reading invoices (with associations)...")
    invoices, invoice_contact_links, max_ts_ms = read_invoice_by_ids(ids, cursor_prop)

    print("Upserting into Supabase...")
    upsert_invoices(invoices, invoice_contact_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_invoices", now_iso)

    print(f"invoices 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()