from django.core.management import BaseCommand
import clickhouse_connect
from vacancies.main.vector_store import add_vectors, extract_features, qdrant_client
from vacancies.conf.settings import CLICKHOUSE_HOST, CLICKHOUSE_PORT

clickhouse_client = clickhouse_connect.create_client(host=CLICKHOUSE_HOST, port=CLICKHOUSE_PORT)

query = """
SELECT id, chat_username, telegram_id, message, timestamp
FROM telegram_parser_chatmessage
WHERE timestamp >= now() - INTERVAL 30 DAY
  AND length(message) > 150
  AND arrayCount(x -> position(message, x) > 0, [
    'вакансия', 'ищем', 'требуется', 'разработчик', 'будет плюсом',
    'зарплата', 'оклад', 'з/п', 'руб', 'опыт работы',
    'требования', 'обязанности', 'условия', 'компания', 'офис',
    'удаленно', 'гибкий график', 'полный день', 'частичная занятость',
    'резюме', 'собеседование', 'junior', 'middle', 'senior'
  ]) >= 5 AND position(message, 'О себе') = 0 AND position(message, 'Обо мне') = 0
  AND position(message, '#ищу') = 0 AND position(message, 'умею') = 0
  AND id NOT IN %(exist_points)s
"""


class Command(BaseCommand):
    help = "Collect vacancies from telegram messages"

    def handle(self, *args, **options):
        next_page_offset = 0
        exist_points_ids = [-1]
        while next_page_offset is not None:
            response = qdrant_client.scroll(
                collection_name="vacancies",
                limit=100_000,
                offset=next_page_offset,
                with_payload=False,
                with_vectors=False,
                timeout=30,
            )
            exist_points_ids.extend([point.id for point in response[0]])
            next_page_offset = response[1]
        exist_points_set = tuple(set(exist_points_ids))

        result_rows = clickhouse_client.query(query, parameters={"exist_points": exist_points_set}).result_rows
        result_rows_len = len(result_rows)
        for index, row in enumerate(result_rows):
            (id, chat_username, telegram_id, message, timestamp) = row

            link = f"https://t.me/{chat_username}/{telegram_id}"
            print(f"Processing {index+1}/{result_rows_len} link: {link}")
            features = extract_features(message)

            add_vectors(
                "vacancies",
                id,
                features.model_dump(),
                {'content': message, 'features_json': features.model_dump(), "link": link, "timestamp": timestamp},
            )