python/hubspot_contacts.py

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

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

Direct CLI:
    # epoch ms
    python hubspot_contacts.py 1754025600000
    # ISO-8601
    python hubspot_contacts.py 2025-08-01T09:30:00Z
    # Back-compat date (floors to 00:00Z)
    python hubspot_contacts.py 2025-08-01
"""

import os
import re
import time
import logging
import datetime
from typing import List, Dict, Tuple, Optional, Union

import httpx
import hubspot
from dotenv import load_dotenv
from supabase import create_client
from hubspot.crm.contacts import ApiException as ContactsApiException

from hubspot_utils import (
    try_parse_int, parse_ts, get_property_label_mapping,
)
from supabase_utils import (
    fetch_supabase_table, update_sync_metadata, enrich_supabase_row,
    upload_raw_json_to_supabase, batched_insert,
)

# -----------------------------------------------------------------------------
# Logging
# -----------------------------------------------------------------------------
logging.basicConfig(
    filename=f"logs/hubspot_contact_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_CONTACTS_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
# -----------------------------------------------------------------------------
CONTACT_PROPERTIES = [
    "full_name",
    "firstname",
    "lastname",
    "email",
    "phone",
    "job_title_level",
    "createdate",
    "lastmodifieddate",
    "hs_lastmodifieddate",
    "notes_last_updated",
    "associatedcompanyid",
    "jobtitle",
    "hs_object_source_label",
    "source_1__cont_",
    "source_2",
    "hs_object_source_detail_1",
    "hs_analytics_source",
    "hs_latest_source",
]

# -----------------------------------------------------------------------------
# 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

# -----------------------------------------------------------------------------
# Mapping & helpers
# -----------------------------------------------------------------------------
def map_contact_data_for_db(contacts: List[Dict]) -> List[Dict]:
    mapped = []
    for c in contacts:
        base_row = {
            "contact_id": try_parse_int(c["id"]),
            "full_name": c.get("full_name"),
            "first_name": c.get("firstname"),
            "last_name": c.get("lastname"),
            "email": c.get("email"),
            "phone_number": c.get("phone"),
            "job_title_level": c.get("job_title_level"),
            "hubspot_create_date": parse_ts(c.get("createdate")) or None,
            "hubspot_modified_date": parse_ts(c.get("lastmodifieddate") or c.get("hs_lastmodifieddate")) or None,
            "hubspot_last_activity_date": parse_ts(c.get("notes_last_updated")) or None,
            "number_of_associated_deals": c.get("number_of_associated_deals", 0),
            "associated_company_id": try_parse_int(c.get("associatedcompanyid")),
            "record_source": c.get("hs_object_source_label"),
            "source_1": c.get("source_1__cont_"),
            "source_2": c.get("source_2"),
            "record_source_detail_1": c.get("hs_object_source_detail_1"),
            "original_traffic_source": c.get("hs_analytics_source"),
            "latest_traffic_source": c.get("hs_latest_source"),
        }
        mapped.append(enrich_supabase_row(base_row))
    return mapped

def contacts_are_different(new_row: Dict, old_row: Dict) -> bool:
    compare_keys = [
        "full_name", "first_name", "last_name", "email", "phone_number",
        "hubspot_create_date", "hubspot_modified_date",
        "hubspot_last_activity_date", "number_of_associated_deals",
        "associated_company_id", "record_source", "source_1", "source_2",
        "record_source_detail_1", "original_traffic_source",
        "latest_traffic_source",
    ]
    for key in compare_keys:
        if str(new_row.get(key)) != str(old_row.get(key)):
            return True
    return False

# -----------------------------------------------------------------------------
# Search IDs (ts > since_ms) with property fallback
# -----------------------------------------------------------------------------
def _search_contact_ids_from(since_ms: int, prop: str) -> List[str]:
    """
    Search contacts where {prop} > since_ms (epoch-ms).
    Sort ascending so we can advance the cursor monotonically.
    """
    url = "https://api.hubapi.com/crm/v3/objects/contacts/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("Contacts search error for prop '%s': %s", prop, resp.json())
                except Exception:
                    logging.error("Contacts 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_contact_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
    """
    props_to_try = ["createdate"]
    last_err = None

    for prop in props_to_try:
        try:
            ids = _search_contact_ids_from(since_ms, prop)
            logging.info("Contacts 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) → enrich & track max cursor ts
# -----------------------------------------------------------------------------
def _enrich_contact_data_from_record(record, job_level_map: Optional[Dict[str, str]]) -> Dict:
    props = record.properties or {}

    contact_data: Dict[str, Optional[str]] = {"id": record.id}
    for p in CONTACT_PROPERTIES:
        contact_data[p] = props.get(p)

    # Full name fallback
    if not contact_data.get("full_name") or not str(contact_data["full_name"]).strip():
        first = contact_data.get("firstname") or ""
        last = contact_data.get("lastname") or ""
        contact_data["full_name"] = f"{first.strip()} {last.strip()}".strip()

    # Associations: deals count
    num_deals = 0
    if getattr(record, "associations", None) and record.associations.get("deals"):
        bucket = record.associations["deals"]
        if getattr(bucket, "results", None):
            num_deals = len({a.id for a in bucket.results if getattr(a, "id", None)})

    contact_data["number_of_associated_deals"] = num_deals

    # Map job_title_level code → label (if configured)
    if job_level_map:
        code = contact_data.get("job_title_level")
        contact_data["job_title_level"] = job_level_map.get(code) if code in job_level_map else None

    return contact_data

def read_contacts_by_ids(contact_ids: List[str], cursor_prop: str) -> Tuple[List[Dict], Optional[int]]:
    if not contact_ids:
        return [], None

    contacts: List[Dict] = []
    assoc_types = ["deals"]

    # Fetch property label map once
    try:
        job_level_map = get_property_label_mapping(hubspot_client, "contacts", "job_title_level")
    except Exception as e:
        logging.warning("Failed to fetch job_title_level map: %s", e)
        job_level_map = None

    max_ts_ms: Optional[int] = None

    for i, cid in enumerate(contact_ids, start=1):
        try:
            record = hubspot_client.crm.contacts.basic_api.get_by_id(
                contact_id=cid,
                properties=CONTACT_PROPERTIES,
                associations=assoc_types,
                archived=False,
            )

            # Track max timestamp for the chosen cursor property
            cursor_val = (record.properties or {}).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.append(_enrich_contact_data_from_record(record, job_level_map))

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

            time.sleep(0.05)  # gentle pacing

        except httpx.HTTPStatusError as e:
            logging.error("HTTP error reading contact %s: %s", cid, e)
        except (ContactsApiException, httpx.HTTPError) as e:
            logging.error("Error reading contact %s: %s", cid, e)

    return contacts, max_ts_ms

# -----------------------------------------------------------------------------
# Upsert flow
# -----------------------------------------------------------------------------
def upsert_contacts(contacts: List[Dict]) -> None:
    if not contacts:
        print("No contacts to upsert.")
        return

    existing = fetch_supabase_table(supabase_client, "hubspot_contacts", "contact_id")
    rows_to_upsert: List[Dict] = []

    for c in contacts:
        contact_id = try_parse_int(c.get("id"))
        if not contact_id:
            continue

        mapped_row = map_contact_data_for_db([c])[0]
        existing_row = existing.get(str(contact_id))

        if not existing_row or contacts_are_different(mapped_row, existing_row):
            rows_to_upsert.append(mapped_row)

    print(f"{len(rows_to_upsert)} contacts to insert/update (out of {len(contacts)} read).")

    if rows_to_upsert:
        # upload_raw_json_to_supabase(supabase_client, rows_to_upsert, object_type="contacts")
        batched_insert(supabase_client, "hubspot_contacts", rows_to_upsert, batch_size=1000)

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

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

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

    print("Reading contacts (with associations)...")
    contacts, max_ts_ms = read_contacts_by_ids(ids, cursor_prop)

    print("Upserting into Supabase...")
    upsert_contacts(contacts)

    # 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, "contacts", now_iso)

    print(f"Contacts 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()