solorpower/crawler/tests/test_backward_backfill.py

import sqlite3
import tempfile
import unittest
from contextlib import closing
from pathlib import Path
from unittest.mock import patch

import backward_backfill


def make_plant(site_id="plant-a", start_date="2026-08-01"):
    return {
        "id": site_id,
        "name": site_id,
        "type": "fake",
        "start_date": start_date,
        "options": {},
    }


class BackwardBackfillTest(unittest.TestCase):
    def setUp(self):
        self.temp_dir = tempfile.TemporaryDirectory()
        self.db_path = str(Path(self.temp_dir.name) / "crawler_manager.db")
        self.db_path_patch = patch.object(backward_backfill, "DB_PATH", self.db_path)
        self.db_path_patch.start()

    def tearDown(self):
        self.db_path_patch.stop()
        self.temp_dir.cleanup()

    def read_state(self, site_id):
        with closing(sqlite3.connect(self.db_path)) as conn:
            return conn.execute(
                """
                SELECT last_backfilled_date, consecutive_zero_count, status
                FROM backfill_state
                WHERE site_id = ?
                """,
                (site_id,),
            ).fetchone()

    def init_sites(self, plants, first_target="2026-08-05"):
        with patch.object(backward_backfill, "get_all_plants", return_value=plants):
            backward_backfill.init_backfill_states(first_target)

    def run_process(self, plants, crawler, save_result=True, days=1):
        with (
            patch.object(backward_backfill, "get_all_plants", return_value=plants),
            patch.object(backward_backfill, "get_history_crawler", return_value=crawler),
            patch.object(backward_backfill, "save_history", return_value=save_result),
        ):
            backward_backfill.process_backfill(
                max_days_per_run=days,
                delay_sec=0,
            )

    def test_first_requested_date_is_not_skipped(self):
        plant = make_plant()
        requested_dates = []
        self.init_sites([plant])

        self.assertEqual(self.read_state("plant-a")[0], "2026-08-06")

        def crawler(_plant, start_date, _end_date):
            requested_dates.append(start_date)
            return [{
                "plant_id": "plant-a",
                "date": start_date,
                "generation_kwh": 12.0,
            }]

        self.run_process([plant], crawler)

        self.assertEqual(requested_dates, ["2026-08-05"])
        self.assertEqual(self.read_state("plant-a"), ("2026-08-05", 0, "RUNNING"))

    def test_fetch_failure_keeps_cursor_and_zero_count(self):
        plant = make_plant()
        self.init_sites([plant])
        backward_backfill.update_backfill_state("plant-a", "2026-08-06", 7, "RUNNING")

        self.run_process([plant], lambda *_args: None)

        self.assertEqual(self.read_state("plant-a"), ("2026-08-06", 7, "RUNNING"))

    def test_save_failure_keeps_cursor_and_zero_count(self):
        plant = make_plant()
        self.init_sites([plant])

        def crawler(_plant, start_date, _end_date):
            return [{
                "plant_id": "plant-a",
                "date": start_date,
                "generation_kwh": 5.0,
            }]

        self.run_process([plant], crawler, save_result=False)

        self.assertEqual(self.read_state("plant-a"), ("2026-08-06", 0, "RUNNING"))

    def test_successful_no_data_advances_without_counting_zero(self):
        plant = make_plant()
        self.init_sites([plant])
        backward_backfill.update_backfill_state("plant-a", "2026-08-06", 4, "RUNNING")

        self.run_process([plant], lambda *_args: [])

        self.assertEqual(self.read_state("plant-a"), ("2026-08-05", 0, "RUNNING"))

    def test_thirtieth_actual_zero_completes_site(self):
        plant = make_plant(start_date="2026-01-01")
        self.init_sites([plant], first_target="2026-08-30")
        backward_backfill.update_backfill_state("plant-a", "2026-08-31", 29, "RUNNING")

        def crawler(_plant, start_date, _end_date):
            return [{
                "plant_id": "plant-a",
                "date": start_date,
                "generation_kwh": 0.0,
            }]

        self.run_process([plant], crawler)

        self.assertEqual(self.read_state("plant-a"), ("2026-08-30", 30, "COMPLETED"))

    def test_one_site_failure_does_not_block_another_site(self):
        plants = [make_plant("plant-a"), make_plant("plant-b")]
        self.init_sites(plants)

        def crawler(plant, start_date, _end_date):
            if plant["id"] == "plant-a":
                return None
            return [{
                "plant_id": "plant-b",
                "date": start_date,
                "generation_kwh": 9.0,
            }]

        self.run_process(plants, crawler)

        self.assertEqual(self.read_state("plant-a")[0], "2026-08-06")
        self.assertEqual(self.read_state("plant-b")[0], "2026-08-05")

    def test_start_date_is_processed_before_completion(self):
        plant = make_plant(start_date="2026-08-05")
        self.init_sites([plant])

        def crawler(_plant, start_date, _end_date):
            return [{
                "plant_id": "plant-a",
                "date": start_date,
                "generation_kwh": 1.0,
            }]

        self.run_process([plant], crawler, days=2)

        self.assertEqual(self.read_state("plant-a"), ("2026-08-05", 0, "COMPLETED"))


if __name__ == "__main__":
    unittest.main()