python/hubspot_tasks.py

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

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

Direct CLI:
    # epoch ms
    python hubspot_tasks.py 1754025600000
    # ISO-8601
    python hubspot_tasks.py 2026-01-01T00:00:00Z
    # Back-compat date (floors to 00:00Z)
    python hubspot_tasks.py 2026-01-01
"""
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.tasks import ApiException as TasksApiException

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

# -----------------------------------------------------------------------------
# Constants
# -----------------------------------------------------------------------------
TASK_PROPERTIES = [
    "hs_created_by",
    "hs_created_by_user_id",
    "hs_createdate",
    "hs_date_entered_61bafb31_e7fa_46ed_aaa9_1322438d6e67_1866552342",
    "hs_engagement_source",
    "hs_object_id",
    "hs_object_source_label",
    "hs_pipeline",
    "hs_pipeline_stage",
    "hs_task_missed_due_date",
    "hs_task_status",
    "hs_task_subject",
    "hs_queue_membership_ids",
    "hs_lastmodifieddate",
]

TASK_ENUM_PROPERTIES = [
    "hs_object_source_label",
    "hs_task_status",
    "hs_queue_membership_ids",
]

# -----------------------------------------------------------------------------
# Logging
# -----------------------------------------------------------------------------
logging.basicConfig(
    filename=f"logs/hubspot_tasks_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_TASKS_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 and enum label maps
# -----------------------------------------------------------------------------
def get_enum_label_maps() -> Dict[str, Dict[str, str]]:
    """
    Retrieves label mappings for all enumeration properties on the task object.

    For each property in TASK_ENUM_PROPERTIES, calls the HubSpot Properties API
    to get option value-to-label mappings. Pipeline and stage labels are fetched
    separately from the Pipelines API and stored under 'hs_pipeline' and
    'hs_pipeline_stage'.

    :return: A dict of {property_name: {value: label}}. Empty dict per property
             on failure.
    """
    label_maps: Dict[str, Dict[str, str]] = {}

    for prop_name in TASK_ENUM_PROPERTIES:
        try:
            prop_info = hubspot_client.crm.properties.core_api.get_by_name(
                "tasks", prop_name
            )
            label_maps[prop_name] = {
                opt.value: opt.label for opt in (prop_info.options or [])
            }
        except Exception as e:
            logging.warning(
                "Failed to fetch options for property '%s': %s", prop_name, str(e)
            )
            label_maps[prop_name] = {}

    # Pipeline stages are nested under pipelines, not in property options.
    # Override hs_pipeline and hs_pipeline_stage with pipeline API data.
    try:
        pipelines_api = hubspot_client.crm.pipelines.pipelines_api
        response = pipelines_api.get_all(object_type="tasks")

        pipeline_map: Dict[str, str] = {}
        stage_map: Dict[str, str] = {}

        for pipeline in response.results:
            pipeline_map[pipeline.id] = pipeline.label
            for stage in pipeline.stages:
                stage_map[stage.id] = stage.label

        label_maps["hs_pipeline"] = pipeline_map
        label_maps["hs_pipeline_stage"] = stage_map

    except Exception as e:
        logging.warning("Failed to fetch task pipeline/stage mappings: %s", str(e))

    return label_maps


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

    for prop in props_to_try:
        try:
            ids = _search_task_ids_from(since_ms, prop)
            logging.info("Task 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 → map rows and track max cursor ts
# -----------------------------------------------------------------------------
def read_tasks_by_ids(
    task_ids: List[str],
    cursor_prop: str,
) -> Tuple[List[Dict], Optional[int]]:
    """
    Read tasks by ID from HubSpot and map them to Supabase row dicts.

    :param task_ids: List of HubSpot task object IDs to fetch.
    :param cursor_prop: The property name used in the search (to track max ts).
    :return: (all_tasks, max_ts_ms) where max_ts_ms is the highest value of
             cursor_prop seen across all fetched tasks, or None if unavailable.
    """
    if not task_ids:
        return [], None

    all_tasks: List[Dict] = []
    label_maps = get_enum_label_maps()
    max_ts_ms: Optional[int] = None

    def get_label(prop_name: str, value: Optional[str]) -> str:
        """Return the human-readable label for an enumeration value, or ''."""
        if value is None:
            return ""
        return label_maps.get(prop_name, {}).get(value, "")

    for i, tid in enumerate(task_ids, start=1):
        try:
            record = hubspot_client.crm.objects.basic_api.get_by_id(
                object_type="tasks",
                object_id=tid,
                properties=TASK_PROPERTIES,
                archived=False,
            )
            props = record.properties or {}

            # Track max timestamp for the cursor property used in the search
            cursor_val = props.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

            # hs_task_missed_due_date comes back as the string "true"/"false"
            missed_due_raw = props.get("hs_task_missed_due_date")
            if missed_due_raw is None:
                missed_due = None
            else:
                missed_due = missed_due_raw.lower() == "true"

            all_tasks.append({
                "task_id": try_parse_int(props.get("hs_object_id")),
                "hubspot_created_by": try_parse_int(props.get("hs_created_by")),
                "hubspot_created_by_user_id": try_parse_int(props.get("hs_created_by_user_id")),
                "hubspot_createdate": parse_ts(props.get("hs_createdate")),
                "date_entered_completed": parse_ts(props.get(
                    "hs_date_entered_61bafb31_e7fa_46ed_aaa9_1322438d6e67_1866552342"
                )),
                "engagement_source": props.get("hs_engagement_source"),
                "record_source": props.get("hs_object_source_label"),
                "record_source_label": get_label(
                    "hs_object_source_label", props.get("hs_object_source_label")
                ),
                "pipeline": props.get("hs_pipeline"),
                "pipeline_label": get_label("hs_pipeline", props.get("hs_pipeline")),
                "pipeline_stage": props.get("hs_pipeline_stage"),
                "pipeline_stage_label": get_label(
                    "hs_pipeline_stage", props.get("hs_pipeline_stage")
                ),
                "task_missed_due_date": missed_due,
                "task_status": props.get("hs_task_status"),
                "task_status_label": get_label("hs_task_status", props.get("hs_task_status")),
                "task_subject": props.get("hs_task_subject"),
                "queue_membership_ids": props.get("hs_queue_membership_ids"),
            })

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

        except httpx.HTTPStatusError as e:
            logging.error("HTTP error reading task %s: %s", tid, e)
        except (TasksApiException, httpx.HTTPError) as e:
            logging.error("Error reading task %s: %s", tid, e)

    return all_tasks, max_ts_ms


# -----------------------------------------------------------------------------
# Upsert
# -----------------------------------------------------------------------------
def upsert_tasks(tasks: List[Dict]) -> None:
    if tasks:
        insert_into_supabase_table(supabase_client, "hubspot_tasks", tasks)
        print(f"Upserted {len(tasks)} tasks.")


# -----------------------------------------------------------------------------
# Main (timestamp cursor)
# -----------------------------------------------------------------------------
def main(since_ms: Optional[int] = None):
    """
    Orchestrates:
      1) Search task IDs with <cursor_prop> > since_ms (property fallback)
      2) Read full tasks (track max timestamp for <cursor_prop>)
      3) Upsert into Supabase
      4) Update sync metadata with last_sync_time
    """
    # 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_TASKS_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 tasks with timestamp > {since_ms} ...")
    ids, cursor_prop = search_task_ids_after_ms(since_ms)
    print(f"Search property: {cursor_prop}. Found {len(ids)} task IDs.")

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

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

    print("Reading tasks...")
    tasks, max_ts_ms = read_tasks_by_ids(ids, cursor_prop)

    print("Upserting into Supabase...")
    upsert_tasks(tasks)

    # 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_tasks", now_iso)

    print(f"Tasks 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()