python/hubspot_deals.py

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

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

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

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
# -----------------------------------------------------------------------------
DEAL_TO_COMPANY_ASSOC_TYPE = "deal_to_company"

HUBSPOT_TEAM_MAP = {
    "1322863": "Fulfilment Team",
    "1322864": "Customer Services Team",
    "1322865": "Sales Team",
    "1448134": "Accounts & Billing Team",
    "3557793": "Marketing Team",
    "57348939": "MDR Team",
}

DEAL_PROPERTIES = [
    "dealname",
    "hubspot_owner_id",
    "pipeline",
    "dealstage",
    "createdate",            # use createdate; hs_createdate is also seen in some portals
    "hs_createdate",
    "closedate",
    "contract_signed_date",
    "contract_end_date",
    "notes_last_updated",
    "num_contacted_notes",
    "hubspot_team_id",
    "hs_analytics_source",
    "number_of_cli",
    "amount",
    "hs_acv",
    "hs_tcv",
    "margin",
    "source_of_deal_2___migration",
    "hs_primary_associated_company",
    "hs_object_source_label",
    "hs_object_source_detail_1",
    "dealtype",
    "hs_lastmodifieddate",   # for mapping/debugging
    "lastmodifieddate",
]

# -----------------------------------------------------------------------------
# Logging
# -----------------------------------------------------------------------------
logging.basicConfig(
    filename=f"logs/hubspot_deals_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_DEALS_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

# -----------------------------------------------------------------------------
# Pipeline / Stage labels
# -----------------------------------------------------------------------------
def get_pipeline_and_stage_mappings() -> Tuple[Dict[str, str], Dict[str, str]]:
    """Retrieve pipeline and stage label mappings for deals."""
    try:
        resp = hubspot_client.crm.pipelines.pipelines_api.get_all(object_type="deals")
        pipeline_mapping: Dict[str, str] = {}
        stage_mapping: Dict[str, str] = {}
        for p in resp.results:
            pipeline_mapping[p.id] = p.label
            for s in p.stages:
                stage_mapping[s.id] = s.label
        return pipeline_mapping, stage_mapping
    except Exception as e:
        logging.error("Failed to fetch pipeline/stage mappings: %s", e)
        return {}, {}

# -----------------------------------------------------------------------------
# Search IDs (ts > since_ms) with property fallback
# -----------------------------------------------------------------------------
def _search_deal_ids_from(since_ms: int, prop: str) -> List[str]:
    """
    Search deals where {prop} > since_ms (epoch-ms).
    Sort ascending so we can advance the cursor monotonically.
    """
    url = "https://api.hubapi.com/crm/v3/objects/deals/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("Deal search error for prop '%s': %s", prop, resp.json())
                except Exception:
                    logging.error("Deal 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_deal_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_deal_ids_from(since_ms, prop)
            logging.info("Deal 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"

# -----------------------------------------------------------------------------
# Read-by-ID (with associations) → map rows and track max cursor ts
# -----------------------------------------------------------------------------
def _extract_primary_company_id(record, props: Dict) -> Optional[int]:
    """
    Prefer hs_primary_associated_company property. If missing, fall back to
    associations of type 'deal_to_company' (if available).
    """
    primary = props.get("hs_primary_associated_company")
    if primary:
        return try_parse_int(primary)

    assoc = getattr(record, "associations", None)
    if assoc and assoc.get("companies") and getattr(assoc["companies"], "results", None):
        for a in assoc["companies"].results:
            if getattr(a, "type", None) == DEAL_TO_COMPANY_ASSOC_TYPE or not hasattr(a, "type"):
                return try_parse_int(a.id)
    return None

def read_deals_by_ids(
    deal_ids: List[str],
    cursor_prop: str,
) -> Tuple[List[Dict], List[Dict], Optional[int]]:
    """
    Read deals by ID including contacts/companies associations.
    Returns: (all_deals, deal_contact_links, max_ts_ms_for_cursor_prop)
    """
    if not deal_ids:
        return [], [], None

    all_deals: List[Dict] = []
    deal_contact_links: List[Dict] = []

    assoc_types = ["contacts", "companies"]
    pipeline_map, stage_map = get_pipeline_and_stage_mappings()

    max_ts_ms: Optional[int] = None

    for i, did in enumerate(deal_ids, start=1):
        try:
            record = hubspot_client.crm.deals.basic_api.get_by_id(
                deal_id=did, properties=DEAL_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():
                            deal_contact_links.append({
                                "deal_id": try_parse_int(record.id),
                                "contact_id": try_parse_int(a.id),
                            })

            # Company
            company_id = _extract_primary_company_id(record, p)

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

            all_deals.append({
                "deal_id": try_parse_int(record.id),
                "dealname": p.get("dealname"),
                "hubspot_owner_id": try_parse_int(p.get("hubspot_owner_id")),
                "pipeline_id": try_parse_int(p.get("pipeline")),
                "pipeline_label": pipeline_map.get(p.get("pipeline"), ""),
                "dealstage": p.get("dealstage"),
                "dealstage_label": stage_map.get(p.get("dealstage"), ""),
                "hubspot_createdate": parse_ts(created_iso),
                "closedate": parse_ts(p.get("closedate")) or None,
                "contract_signed_date": p.get("contract_signed_date") or None,
                "contract_end_date": p.get("contract_end_date") or None,
                "hubspot_last_activity_date": parse_ts(p.get("notes_last_updated")),
                "num_contacted_notes": p.get("num_contacted_notes"),
                "hubspot_team_id": try_parse_int(p.get("hubspot_team_id")),
                "hubspot_team_label": HUBSPOT_TEAM_MAP.get(p.get("hubspot_team_id"), ""),
                "hs_analytics_source": p.get("hs_analytics_source"),
                "number_of_cli": try_parse_int(p.get("number_of_cli")),
                "amount": try_parse_int(p.get("amount")),
                "annual_contract_value": try_parse_float(p.get("hs_acv")),
                "total_contract_value": try_parse_float(p.get("hs_tcv")),
                "margin": try_parse_float(p.get("margin")),
                "source_of_deal": p.get("source_of_deal_2___migration"),
                "record_source": p.get("hs_object_source_label"),
                "record_source_detail_1": p.get("hs_object_source_detail_1"),
                "dealtype": p.get("dealtype"),
                "company_id": company_id,
            })

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

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

    return all_deals, deal_contact_links, max_ts_ms

# -----------------------------------------------------------------------------
# Upsert
# -----------------------------------------------------------------------------
def upsert_deals(deals: List[Dict], deal_contact_links: List[Dict]) -> None:
    if deals:
        insert_into_supabase_table(supabase_client, "hubspot_deals", deals)
        print(f"Upserted {len(deals)} deals.")

    if deal_contact_links:
        deal_contact_links = deduplicate_by_key(deal_contact_links, key=("deal_id", "contact_id"))
        insert_into_supabase_table(
            supabase_client,
            "hubspot_deal_contacts",
            deal_contact_links,
            on_conflict=["deal_id", "contact_id"],
        )
        print(f"Upserted {len(deal_contact_links)} deal-contact associations.")

# -----------------------------------------------------------------------------
# Main (timestamp cursor)
# -----------------------------------------------------------------------------
def main(since_ms: Optional[int] = None):
    """
    Orchestrates:
      1) Search deal IDs with <cursor_prop> > since_ms (property fallback)
      2) Read full deals 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_DEALS_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 deals with timestamp > {since_ms} ...")
    ids, cursor_prop = search_deal_ids_after_ms(since_ms)
    print(f"Search property: {cursor_prop}. Found {len(ids)} deal IDs.")

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

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

    print("Reading deals (with associations)...")
    deals, deal_contact_links, max_ts_ms = read_deals_by_ids(ids, cursor_prop)

    print("Upserting into Supabase...")
    upsert_deals(deals, deal_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_deals", now_iso)

    print(f"Deals 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()