import json
import pathlib
import re
import tempfile
from datetime import datetime, timedelta, timezone
from http import HTTPStatus
from unittest import mock

import pytest
import pytz
from PIL import Image
from pytz import utc
from sqlalchemy import text

from modules.archive import lib
from modules.archive.lib import ArchiveDownloaderConfigValidator
from modules.archive.lib import format_archive_filename
from modules.archive.lib import get_archive_downloader_config
from modules.archive.lib import get_or_create_domain_collection, get_new_archive_files, \
    delete_archives_by_file_group_ids, model_archive_result, get_domains
from modules.archive.models import Archive
from wrolpi.api_utils import CustomJSONEncoder
from wrolpi.cmd import CommandResult
from wrolpi.collections import Collection
from wrolpi.common import get_wrolpi_config
from wrolpi.db import get_db_session
from wrolpi.files.lib import count_file_groups
from wrolpi.files.models import FileGroup
from wrolpi.test.common import skip_ci


def make_fake_archive_result(readability=True, screenshot=True, title=True):
    with tempfile.NamedTemporaryFile(suffix='.png') as fh:
        Image.new('RGB', (1, 1), color='grey').save(fh)
        fh.seek(0)
        image = fh.read()
    singlefile = '<html><!--\n Page saved with SingleFile-->\ntest single-file\nジにてこちら\n<title>some title</title></html>'
    r = dict(
        content=f'<html>test readability content</html>',
        textContent='test readability textContent',
        title='ジにてこちら' if title else None,
    ) if readability else None
    s = image if screenshot else None
    return singlefile, r, s


@pytest.mark.asyncio
async def test_no_screenshot(async_client, test_directory, test_session):
    singlefile, readability, screenshot = make_fake_archive_result(screenshot=False)
    archive = await model_archive_result('https://example.com', singlefile, readability, screenshot)
    assert isinstance(archive.singlefile_path, pathlib.Path)
    assert isinstance(archive.readability_path, pathlib.Path)
    # Screenshot was empty
    assert archive.screenshot_path is None


@pytest.mark.asyncio
async def test_no_readability(async_client, test_directory, test_session):
    singlefile, readability, screenshot = make_fake_archive_result(readability=False)
    archive = await model_archive_result('https://example.com', singlefile, readability, screenshot)
    assert isinstance(archive.singlefile_path, pathlib.Path)
    assert isinstance(archive.screenshot_path, pathlib.Path)
    # Readability empty
    assert archive.readability_path is None
    assert archive.readability_txt_path is None


@pytest.mark.asyncio
async def test_dict(async_client, test_session, test_directory):
    singlefile, readability, screenshot = make_fake_archive_result()
    d = (await model_archive_result('https://example.com', singlefile, readability, screenshot)).dict()
    assert isinstance(d, dict)
    json.dumps(d, cls=CustomJSONEncoder)


@pytest.mark.asyncio
async def test_relationships(async_client, test_session, example_singlefile):
    with get_db_session(commit=True) as session:
        url = 'https://wrolpi.org:443'
        collection = get_or_create_domain_collection(session, url)
        archive = Archive.from_paths(test_session, example_singlefile)
        archive.url = url
        archive.collection_id = collection.id
        session.add(archive)
        session.flush()

    assert archive.collection == collection
    assert archive.domain == 'wrolpi.org'


@pytest.mark.asyncio
async def test_archive_title(async_client, test_session, archive_factory, singlefile_contents_factory, refresh_files):
    """An Archive's title can be fetched in multiple ways.  This tests from most to least reliable."""
    # Create some test files, delete all records for a fresh refresh.
    archive_factory(
        domain='example.com',
        url='https://example.com/json-url',
        title='json title',
        singlefile_contents=singlefile_contents_factory('singlefile title', 'https://example.com/singlefile-url'),
    )

    async def reset_and_get_archive():
        test_session.query(FileGroup).delete()
        test_session.commit()
        await refresh_files()
        return test_session.query(Archive).one()

    archive1: Archive = await reset_and_get_archive()

    assert test_session.query(FileGroup).count() == 1
    assert test_session.query(Archive).count() == 1

    # Readability JSON is trusted first.
    assert archive1.file_group.url == 'https://example.com/json-url'
    assert archive1.file_group.title == 'json title'

    # Delete JSON file.
    archive1.readability_json_path.unlink()
    archive1: Archive = await reset_and_get_archive()
    assert archive1.file_group.url == 'https://example.com/singlefile-url'
    assert archive1.file_group.title == 'singlefile title'

    # Clear out title and URL from Singlefile.
    archive1.singlefile_path.write_text(singlefile_contents_factory('', ''))
    archive1: Archive = await reset_and_get_archive()
    # No title/URL could be found.
    assert archive1.file_group.url is None
    assert archive1.file_group.title is None


@pytest.mark.asyncio
async def test_archive_refresh_deleted_archive(async_client, test_session, archive_directory, archive_factory):
    """Archives/domain collections should be deleted when archive files are deleted."""
    archive1 = archive_factory('example.com', 'https://example.com/1')
    archive2 = archive_factory('example.com', 'https://example.com/1')
    archive3 = archive_factory('example.com')
    archive4 = archive_factory('example.org')
    archive5 = archive_factory()
    test_session.commit()

    # Empty directories should be ignored.
    (archive_directory / 'empty').mkdir()

    def check_counts(archive_count, domain_count):
        assert test_session.query(Archive).count() == archive_count, 'Archive count does not match'
        assert test_session.query(Collection).filter_by(
            kind='domain').count() == domain_count, 'domain collection count does not match'

    # All 5 archives are already in the DB.
    check_counts(archive_count=5, domain_count=2)
    await async_client.post('/api/files/refresh')
    check_counts(archive_count=5, domain_count=1)

    # Delete archive2's files, it's the latest for 'https://example.com/1'
    for path in archive2.my_paths():
        path.unlink()
    await async_client.post('/api/files/refresh')
    check_counts(archive_count=4, domain_count=1)

    # Delete archive1's files, now the URL is empty.
    for path in archive1.my_paths():
        path.unlink()
    await async_client.post('/api/files/refresh')
    check_counts(archive_count=3, domain_count=0)

    # Delete archive3, now there is now example.com domain
    for path in archive3.my_paths():
        path.unlink()
    await async_client.post('/api/files/refresh')
    check_counts(archive_count=2, domain_count=0)

    # Delete all the rest of the archives
    for path in archive4.my_paths():
        path.unlink()
    for path in archive5.my_paths():
        path.unlink()
    await async_client.post('/api/files/refresh')
    check_counts(archive_count=0, domain_count=0)


@pytest.mark.asyncio
async def test_fills_contents_with_refresh(async_client, test_session, archive_factory, singlefile_contents_factory,
                                           refresh_files):
    """Refreshing archives fills in any missing contents."""
    archive1 = archive_factory('example.com', 'https://example.com/one')
    archive2 = archive_factory('example.com', 'https://example.com/one')
    # Title can be found in archive3's singlefile
    archive3 = archive_factory('example.org', singlefile_contents=singlefile_contents_factory('last title'))
    archive4 = archive_factory('example.org')  # has no contents

    # Clear the title, this will be filled in.
    archive1.file_group.title = archive2.file_group.title = archive3.file_group.title = archive4.file_group.title = None

    contents_title_archive = [
        ('foo bar qux', 'my archive', archive1),
        ('foo baz qux qux', 'other archive', archive2),
        ('baz qux qux qux', None, archive3),
    ]
    for contents, json_title, archive in contents_title_archive:
        archive: Archive
        readability_txt_path = archive.singlefile_path.with_suffix('.readability.txt')
        with readability_txt_path.open('wt') as fh:
            fh.write(contents)
        if json_title:
            readability_json_path = archive.singlefile_path.with_suffix('.readability.json')
            readability_json_path.write_text(json.dumps({'title': json_title, 'url': archive.url}))
    test_session.commit()

    # Contents are empty.  Archives are all missing the title.
    assert not archive1.file_group.d_text
    assert not archive2.file_group.d_text
    assert not archive3.file_group.d_text
    assert not archive4.file_group.d_text

    # Fill the contents.
    await refresh_files()
    # The archives will be renamed with their title.
    archive1, archive2, archive3, archive4 = test_session.query(Archive).order_by(Archive.id)
    assert not archive4.file_group.d_text

    # Missing titles are filled.
    assert archive1.file_group.title == 'my archive'
    assert archive2.file_group.title == 'other archive'
    assert archive3.file_group.title == 'last title'  # from singlefile HTML


@pytest.mark.asyncio
async def test_delete_archive(async_client, test_session, archive_factory):
    """Archives can be deleted."""
    archive1 = archive_factory('example.com', 'https://example.com/1')
    archive2 = archive_factory('example.com', 'https://example.com/1')
    archive3 = archive_factory('example.com', 'https://example.com/1')
    test_session.commit()

    # Files are real.
    assert archive1.my_paths() and all(i.is_file() for i in archive1.my_paths())
    assert archive2.my_paths() and all(i.is_file() for i in archive2.my_paths())
    assert archive3.my_paths() and all(i.is_file() for i in archive3.my_paths())

    assert test_session.query(Archive).count() == 3
    assert test_session.query(FileGroup).count() == 3

    # Save paths before deletion (archives become detached after delete)
    archive1_paths = list(archive1.my_paths())
    archive2_paths = list(archive2.my_paths())
    archive3_paths = list(archive3.my_paths())

    # Delete the oldest.
    delete_archives_by_file_group_ids(archive1.file_group_id, archive3.file_group_id)
    assert test_session.query(Archive).count() == 1
    assert test_session.query(FileGroup).count() == 1
    # Files were deleted.
    assert archive1_paths and not any(i.is_file() for i in archive1_paths)
    assert archive3_paths and not any(i.is_file() for i in archive3_paths)
    # Archive2 is untouched
    assert archive2_paths and all(i.is_file() for i in archive2_paths)

    # Delete the last archive.  The domain collection should also be deleted.
    delete_archives_by_file_group_ids(archive2.file_group_id)
    assert test_session.query(Archive).count() == 0
    domain = test_session.query(Collection).filter_by(kind='domain').one_or_none()
    assert domain is None

    # All Files were deleted.
    assert test_session.query(FileGroup).count() == 0
    assert archive2_paths and not any(i.is_file() for i in archive2_paths)


def test_get_domains(test_session, archive_factory):
    """
    `get_domains` gets only domain collections with Archives.
    """
    archive1 = archive_factory('example.com')
    archive2 = archive_factory('example.com')
    archive3 = archive_factory('example.org')
    test_session.commit()

    assert archive1.domain
    assert archive2.domain
    assert archive3.domain

    assert [i['domain'] for i in get_domains()] == ['example.com', 'example.org']

    archive2.delete()
    test_session.commit()
    assert [i['domain'] for i in get_domains()] == ['example.com', 'example.org']

    archive1.delete()
    test_session.commit()
    assert [i['domain'] for i in get_domains()] == ['example.org']

    archive3.delete()
    test_session.commit()
    assert [i['domain'] for i in get_domains()] == []


@skip_ci
@pytest.mark.asyncio
async def test_new_archive(test_session, test_directory, fake_now, async_client):
    singlefile, readability, screenshot = make_fake_archive_result()
    fake_now(datetime(2000, 1, 1))
    archive1 = await model_archive_result('https://example.com', singlefile, readability, screenshot)
    # Everything is filled out.
    assert isinstance(archive1, Archive)
    assert archive1.file_group.download_datetime is not None
    assert isinstance(archive1.singlefile_path, pathlib.Path)
    assert isinstance(archive1.readability_path, pathlib.Path)
    assert isinstance(archive1.readability_txt_path, pathlib.Path)
    assert isinstance(archive1.screenshot_path, pathlib.Path)
    assert archive1.file_group.title == 'ジにてこちら'
    assert archive1.url is not None
    assert archive1.domain is not None

    # The actual files were dumped and read correctly.  The HTML/JSON files are reformatted.
    # Note: We check for content presence rather than exact formatting because BeautifulSoup's
    # prettify() can produce different output on different platforms (Mac vs CI).
    singlefile_text = archive1.singlefile_path.read_text()
    assert 'Page saved with SingleFile' in singlefile_text
    assert 'test single-file' in singlefile_text
    assert 'ジにてこちら' in singlefile_text
    assert '<title>' in singlefile_text
    assert 'some title' in singlefile_text
    assert singlefile_text.count('\n') >= 5  # Verify it's properly formatted with newlines

    readability_text = archive1.readability_path.read_text()
    assert 'test readability content' in readability_text
    assert readability_text.count('\n') >= 3  # Verify it's properly formatted with newlines
    with open(archive1.readability_txt_path) as fh:
        assert fh.read() == 'test readability textContent'
    assert archive1.readability_json_path.read_text() == '''{
  "title": "\\u30b8\\u306b\\u3066\\u3053\\u3061\\u3089",
  "url": "https://example.com"
}'''

    fake_now(datetime(2000, 1, 2))
    archive2 = await model_archive_result('https://example.com', singlefile, readability, screenshot)
    # domain collection is reused.
    assert archive1.domain == archive2.domain


@pytest.mark.asyncio
async def test_get_title_from_html(test_directory, test_session, fake_now, async_client):
    fake_now(datetime(2000, 1, 1))
    singlefile, readability, screenshot = make_fake_archive_result()
    archive = await model_archive_result('https://example.com', singlefile, readability, screenshot)
    assert archive.file_group.title == 'ジにてこちら'

    singlefile = '<html>\ntest single-file\nジにてこちら\n<title>some title</title></html>'
    readability = dict(
        content=f'<html>test readability content</html>',
        textContent='<html>test readability textContent</html>',
    )
    screenshot = b'screenshot data'

    fake_now(datetime(2000, 1, 2))
    archive = await model_archive_result('https://example.com', singlefile, readability, screenshot)
    assert archive.file_group.title == 'some title'

    singlefile = '<html></html>'
    readability = dict(
        content=f'<html>missing a title</html>',
        textContent='',
    )
    screenshot = b'screenshot data'

    fake_now(datetime(2000, 1, 3))
    archive = await model_archive_result('https://example.com', singlefile, readability, screenshot)
    assert archive.file_group.title is None


@skip_ci
def test_get_new_archive_files_length(test_directory, fake_now):
    """Archive titles are truncated to fit file system BYTE limit (200 bytes - see MAXIMUM_FILE_BYTES)."""
    from wrolpi.common import MAXIMUM_FILE_BYTES

    fake_now(datetime(2001, 1, 1))
    archive_files = get_new_archive_files('https://example.com', 'a' * 500)
    # All filenames should be <= MAXIMUM_FILE_BYTES (bytes, not characters)
    assert len(str(archive_files.singlefile.name).encode('utf-8')) <= MAXIMUM_FILE_BYTES
    assert len(str(archive_files.readability.name).encode('utf-8')) <= MAXIMUM_FILE_BYTES
    assert len(str(archive_files.readability_txt.name).encode('utf-8')) <= MAXIMUM_FILE_BYTES
    assert len(str(archive_files.readability_json.name).encode('utf-8')) <= MAXIMUM_FILE_BYTES
    assert len(str(archive_files.screenshot.name).encode('utf-8')) <= MAXIMUM_FILE_BYTES


@skip_ci
def test_get_new_archive_files(test_directory, fake_now):
    """Archive files have a specific format, so they are sorted by datetime, and are near each other."""
    fake_now(datetime(2001, 1, 1))
    archive_files = get_new_archive_files('https://example.com/two', None)
    assert str(archive_files.singlefile).endswith('archive/example.com/2001-01-01-00-00-00_NA.html')
    assert str(archive_files.readability).endswith('archive/example.com/2001-01-01-00-00-00_NA.readability.html')
    assert str(archive_files.readability_txt).endswith('archive/example.com/2001-01-01-00-00-00_NA.readability.txt')
    assert str(archive_files.readability_json).endswith('archive/example.com/2001-01-01-00-00-00_NA.readability.json')
    assert str(archive_files.screenshot).endswith('archive/example.com/2001-01-01-00-00-00_NA.png')

    archive_files = get_new_archive_files('https://www.example.com/one', 'Title')
    # Leading www. is removed.
    assert str(archive_files.singlefile).endswith('archive/example.com/2001-01-01-00-00-00_Title.html')
    assert str(archive_files.readability).endswith('archive/example.com/2001-01-01-00-00-00_Title.readability.html')
    assert str(archive_files.readability_txt).endswith('archive/example.com/2001-01-01-00-00-00_Title.readability.txt')
    assert str(archive_files.readability_json).endswith(
        'archive/example.com/2001-01-01-00-00-00_Title.readability.json')
    assert str(archive_files.screenshot).endswith('archive/example.com/2001-01-01-00-00-00_Title.png')


@skip_ci
def test_get_new_archive_files_with_destination(test_directory, fake_now):
    """Archive files can be created in a custom destination directory instead of the default."""
    fake_now(datetime(2001, 1, 1))
    custom_destination = test_directory / 'archive/News/custom-domain.com'
    custom_destination.mkdir(parents=True, exist_ok=True)

    # When destination is provided, files should go there instead of archive/<domain>
    archive_files = get_new_archive_files('https://example.com/page', 'Title', destination=custom_destination)

    # Files should be in the custom destination, not archive/example.com
    assert str(archive_files.singlefile).endswith('archive/News/custom-domain.com/2001-01-01-00-00-00_Title.html')
    assert str(archive_files.readability).endswith(
        'archive/News/custom-domain.com/2001-01-01-00-00-00_Title.readability.html')
    assert str(archive_files.readability_txt).endswith(
        'archive/News/custom-domain.com/2001-01-01-00-00-00_Title.readability.txt')
    assert str(archive_files.readability_json).endswith(
        'archive/News/custom-domain.com/2001-01-01-00-00-00_Title.readability.json')
    assert str(archive_files.screenshot).endswith('archive/News/custom-domain.com/2001-01-01-00-00-00_Title.png')


@skip_ci
def test_get_new_archive_files_with_subdirectory_format(test_directory, fake_now):
    """Archive files use subdirectories when the config format contains subdirectories."""
    from modules.archive.lib import get_archive_downloader_config

    fake_now(datetime(2001, 1, 1))

    # Configure a file format with subdirectory
    config = get_archive_downloader_config()
    original_format = config._config['file_name_format']
    config._config['file_name_format'] = '%(download_year)s/%(download_datetime)s_%(title)s.%(ext)s'

    try:
        archive_files = get_new_archive_files('https://example.com/page', 'My Article')

        # Files should be created in a year subdirectory
        assert str(archive_files.singlefile).endswith('archive/example.com/2001/2001-01-01-00-00-00_My Article.html')
        assert str(archive_files.readability).endswith(
            'archive/example.com/2001/2001-01-01-00-00-00_My Article.readability.html')
        assert str(archive_files.readability_txt).endswith(
            'archive/example.com/2001/2001-01-01-00-00-00_My Article.readability.txt')
        assert str(archive_files.readability_json).endswith(
            'archive/example.com/2001/2001-01-01-00-00-00_My Article.readability.json')
        assert str(archive_files.screenshot).endswith('archive/example.com/2001/2001-01-01-00-00-00_My Article.png')

        # The subdirectory should exist
        subdir = test_directory / 'archive/example.com/2001'
        assert subdir.is_dir()
    finally:
        config._config['file_name_format'] = original_format


@skip_ci
@pytest.mark.asyncio
async def test_title_in_filename(async_client, test_session, fake_now, test_directory, image_bytes_factory):
    """
    The Archive files have the title in the path.
    """
    fake_now(datetime(2000, 1, 1))

    singlefile, readability, screenshot = make_fake_archive_result()
    archive1 = await model_archive_result('https://example.com', singlefile, readability, screenshot)

    assert archive1.file_group.title == 'ジにてこちら'

    assert str(archive1.singlefile_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_ジにてこちら.html'
    assert str(archive1.readability_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_ジにてこちら.readability.html'
    assert str(archive1.readability_txt_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_ジにてこちら.readability.txt'
    assert str(archive1.readability_json_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_ジにてこちら.readability.json'
    assert str(archive1.screenshot_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_ジにてこちら.png'

    singlefile = '<html>\ntest single-file\nジにてこちら\n</html>'  # no title in HTML
    readability = dict(
        # No Title from Readability.
        content=f'<html>test readability content</html>',
        textContent='test readability textContent',
    )
    screenshot = image_bytes_factory()

    archive2 = await model_archive_result('https://example.com', singlefile, readability, screenshot)

    assert str(archive2.singlefile_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_NA.html'
    assert str(archive2.readability_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_NA.readability.html'
    assert str(archive2.readability_txt_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_NA.readability.txt'
    assert str(archive2.readability_json_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_NA.readability.json'
    assert str(archive2.screenshot_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_NA.png'

    # Test with malformed HTML - <title> not properly closed with </title>
    # The parser includes </html></html> in the title, which gets escaped
    singlefile = '<html>\ntest single-file\nジにてこちら\n<title>dangerous ;\\//_title</html></html>'
    readability = dict(
        # No Title from Readability.
        content=f'<html>test readability content</html>',
        textContent='<html>test readability textContent</html>',
    )
    screenshot = image_bytes_factory()

    archive3 = await model_archive_result('https://example.com', singlefile, readability, screenshot)

    # Title includes </html></html> because the <title> tag is malformed (no </title>)
    # Slashes are escaped: </html> -> <⧸html>
    assert str(archive3.singlefile_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_dangerous ;⧸⧸_title⧸html⧸html.html'
    assert str(archive3.readability_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_dangerous ;⧸⧸_title⧸html⧸html.readability.html'
    assert str(archive3.readability_txt_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_dangerous ;⧸⧸_title⧸html⧸html.readability.txt'
    assert str(archive3.readability_json_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_dangerous ;⧸⧸_title⧸html⧸html.readability.json'
    assert str(archive3.screenshot_path.relative_to(test_directory)) == \
           'archive/example.com/2000-01-01-00-00-00_dangerous ;⧸⧸_title⧸html⧸html.png'


@pytest.mark.asyncio
async def test_refresh_archives(test_session, test_directory, async_client, make_files_structure):
    """Archives can be found and put in the database.  A single Archive will have multiple files."""
    # The start of a typical singlefile html file.
    singlefile_contents = '''<!DOCTYPE html> <html lang="en"><!--
 Page saved with SingleFile 
 url: https://example.com
 saved date: Mon May 16 2022 23:51:35 GMT+0000 (Coordinated Universal Time)
--><head><meta charset="utf-8">'''

    make_files_structure({
        # These are all for an individual Archive.
        'archive/example.com/2021-10-05-16-20-10_NA.html': singlefile_contents,
        'archive/example.com/2021-10-05-16-20-10_NA.png': None,
        'archive/example.com/2021-10-05-16-20-10_NA.readability.txt': 'article text contents',
        'archive/example.com/2021-10-05-16-20-10_NA.readability.json': '{"url": "https://example.com"}',
        'archive/example.com/2021-10-05-16-20-10_NA.readability.html': '<html></html>',
        # These should log an error because they are missing the singlefile path.
        'archive/example.com/2021-10-05 16:20:10.readability.html': '<html></html>',
        'archive/example.com/2021-10-05 16:20:10.readability.json': None,
        # Bad files should also be ignored.
        'archive/example.com/random file': 'hello',
        'archive/example.com/2021-10-05 16:20:10 no extension': None,
    })

    # The single archive is found.
    await async_client.post('/api/files/refresh')
    assert test_session.query(Archive).count() == 1

    # Cause a re-index of the archive.
    test_session.query(FileGroup).filter_by(
        primary_path=str(test_directory / 'archive/example.com/2021-10-05-16-20-10_NA.html')).one().indexed = False
    test_session.commit()

    # Running the refresh does not result in a new archive.
    await async_client.post('/api/files/refresh')
    assert test_session.query(Archive).count() == 1

    # Archives file format was changed, lets check the new formats are found.
    make_files_structure({
        'archive/example.com/2021-10-05-16-20-11_The Title.html': '<html></html>',
        'archive/example.com/2021-10-05-16-20-11_The Title.readability.json': '{"url": "https://example.com"}',
    })
    await async_client.post('/api/files/refresh')
    # The old formatted archive above is renamed.
    assert (test_directory / 'archive/example.com/2021-10-05-16-20-10_NA.html').is_file()
    assert test_session.query(Archive).count() == 2

    # "text" is only in the readability contents (d_text), so a deep search is required.
    content = json.dumps({'search_str': 'text', 'model': 'archive', 'deep': True})
    request, response = await async_client.post('/api/files/search', content=content)
    assert response.status_code == HTTPStatus.OK
    assert response.json['file_groups'], 'No files matched "text"'
    assert response.json['file_groups'][0]['model'] == 'archive', 'Returned file was not an archive'
    # data now stores just filenames (relative paths), not full paths
    assert response.json['file_groups'][0]['data'].get('readability_path') == \
           '2021-10-05-16-20-10_NA.readability.html', \
        'Could not find readability html file'
    assert response.json['file_groups'][0]['data'].get('readability_txt_path') == \
           '2021-10-05-16-20-10_NA.readability.txt', \
        'Could not find readability text file containing "text"'
    assert response.json['file_groups'][0]['download_datetime'], 'Archive has no datetime'


@pytest.mark.asyncio
async def test_refresh_archives_index(test_session, make_files_structure, refresh_files):
    """Archives are indexed using ArchiveIndexer and archive_modeler."""
    # The start of a typical singlefile html file.
    singlefile_contents = '''<!DOCTYPE html> <html lang="en"><!--
 Page saved with SingleFile
 url: https://example.com
 saved date: Mon May 16 2022 23:51:35 GMT+0000 (Coordinated Universal Time)
--><head><meta charset="utf-8">
<script class="sf-hidden" type="application/ld+json">
 {"@context":"http://schema.org", "@type":"NewsArticle", "headline":"The headline",
  "datePublished":"2022-09-27T00:40:19.000Z", "dateModified":"2022-09-27T13:43:47.971Z",
  "author":{"@type":"Person", "name":"AUTHOR NAME", "jobTitle":""},
  "description": "The article description"}
</script>
</html>'''

    singlefile, *_ = make_files_structure({
        'archive/example.com/2021-10-05-16-20-10_NA.html': singlefile_contents,
        'archive/example.com/2021-10-05-16-20-10_NA.png': None,
        'archive/example.com/2021-10-05-16-20-10_NA.readability.txt': 'article text contents',
        'archive/example.com/2021-10-05-16-20-10_NA.readability.json':
            '{"url": "https://example.com", "title": "the title"}',
        'archive/example.com/2021-10-05-16-20-10_NA.readability.html': '<html></html>',
    })

    await refresh_files()

    archive: Archive = test_session.query(Archive).one()
    assert archive.singlefile_path == singlefile

    assert archive.file_group.author == 'AUTHOR NAME'
    assert archive.file_group.published_datetime
    assert archive.file_group.published_modified_datetime

    assert archive.file_group.a_text == 'the title'  # from readability json
    assert archive.file_group.b_text == 'The article description'  # From application/ld+json
    assert archive.file_group.d_text == 'article text contents'  # from readability txt
    assert archive.file_group.indexed is True


@pytest.mark.asyncio
@pytest.mark.parametrize('saved_date,expected', [
    ('Mon May 16 2022 23:51:35 GMT+0000 (Coordinated Universal Time)', datetime(2022, 5, 16, 23, 51, 35)),
    # The browser writes its local time; the offset must be applied, not dropped.
    ('Thu Nov 07 2024 18:36:23 GMT-0700 (Mountain Standard Time)', datetime(2024, 11, 8, 1, 36, 23)),
    ('Tue Mar 03 2026 09:15:00 GMT+0200 (Eastern European Standard Time)', datetime(2026, 3, 3, 7, 15, 0)),
])
async def test_singlefile_saved_date_timezone(test_session, make_files_structure, refresh_files, saved_date, expected):
    """An Archive's download_datetime is the SingleFile "saved date" converted to UTC."""
    make_files_structure({
        'archive/example.com/NA.html': f'''<!DOCTYPE html> <html lang="en"><!--
 Page saved with SingleFile
 url: https://example.com
 saved date: {saved_date}
--><head><meta charset="utf-8"></head></html>''',
    })

    await refresh_files()

    archive: Archive = test_session.query(Archive).one()
    assert archive.file_group.download_datetime == expected.replace(tzinfo=timezone.utc)


@pytest.mark.asyncio
async def test_singlefile_saved_date_after_long_url(test_session, make_files_structure, refresh_files):
    """The whole SingleFile header is read, so a long URL does not cut the saved date's offset off."""
    url = 'https://example.com/' + 'a' * 900
    make_files_structure({
        'archive/example.com/NA.html': f'''<!DOCTYPE html> <html lang="en"><!--
 Page saved with SingleFile
 url: {url}
 saved date: Mon May 26 2025 16:10:33 GMT-0600 (Mountain Daylight Time)
--><head><meta charset="utf-8"></head></html>''',
    })

    await refresh_files()

    archive: Archive = test_session.query(Archive).one()
    assert archive.file_group.url == url
    assert archive.file_group.download_datetime == datetime(2025, 5, 26, 22, 10, 33, tzinfo=timezone.utc)


@pytest.mark.parametrize('saved_date,expected', [
    ('Mon May 26 2025 16:10:33 GMT-0600 (Mountain Daylight Time)', datetime(2025, 5, 26, 22, 10, 33)),
    ('Mon May 26 2025 16:10:33 GMT+05:30', datetime(2025, 5, 26, 10, 40, 33)),
    # No offset at all is GMT.
    ('Mon May 26 2025 16:10:33', datetime(2025, 5, 26, 16, 10, 33)),
    # A cut-off offset is unknown, not GMT.
    ('Mon May 26 2025 16:10:33 GMT-06', None),
])
def test_parse_singlefile_saved_date(saved_date, expected):
    from modules.archive.models import parse_singlefile_saved_date
    expected = expected.replace(tzinfo=timezone.utc) if expected else None
    assert parse_singlefile_saved_date(saved_date) == expected


@pytest.mark.asyncio
async def test_repair_singlefile_saved_dates_unreadable(test_session, make_files_structure, refresh_files):
    """The repair is not marked done while an Archive could not be read, so it runs again."""
    misread = datetime(2024, 11, 7, 18, 36, 23, tzinfo=timezone.utc)
    singlefile, = make_files_structure({
        'archive/example.com/NA.html': '''<!DOCTYPE html> <html lang="en"><!--
 Page saved with SingleFile
 url: https://example.com
 saved date: Thu Nov 07 2024 18:36:23 GMT-0700 (Mountain Standard Time)
--><head><meta charset="utf-8"></head></html>''',
    })
    await refresh_files()
    test_session.query(Archive).one().file_group.download_datetime = misread
    test_session.commit()

    singlefile.chmod(0o000)
    try:
        assert lib.repair_singlefile_saved_dates() == 0
    finally:
        singlefile.chmod(0o644)

    assert lib.repair_singlefile_saved_dates() == 1
    test_session.expire_all()
    assert test_session.query(Archive).one().file_group.download_datetime == \
           datetime(2024, 11, 8, 1, 36, 23, tzinfo=timezone.utc)


@pytest.mark.asyncio
async def test_repair_singlefile_saved_dates(test_session, make_files_structure, refresh_files):
    """An Archive whose download_datetime is its SingleFile saved date's local time read as UTC is corrected, once.
    Any other download_datetime is never changed."""
    mountain = 'Thu Nov 07 2024 18:36:23 GMT-0700 (Mountain Standard Time)'
    make_files_structure({
        f'archive/example.com/{name}.html': f'''<!DOCTYPE html> <html lang="en"><!--
 Page saved with SingleFile
 url: https://example.com/{name}
 saved date: {saved_date}
--><head><meta charset="utf-8"></head></html>'''
        for name, saved_date in (('misread', mountain), ('downloaded', mountain),
                                 ('utc', 'Mon May 16 2022 23:51:35 GMT+0000 (Coordinated Universal Time)'))
    })
    await refresh_files()

    def file_groups() -> dict:
        test_session.expire_all()
        return {i.file_group.primary_path.stem: i.file_group for i in test_session.query(Archive)}

    # What the offset-dropping parser stored, and a time that came from elsewhere (e.g. WROLPi's own download).
    misread = datetime(2024, 11, 7, 18, 36, 23, tzinfo=timezone.utc)
    downloaded = datetime(2024, 11, 8, 1, 36, 50, tzinfo=timezone.utc)
    fgs = file_groups()
    fgs['misread'].download_datetime = misread
    fgs['downloaded'].download_datetime = downloaded
    test_session.commit()

    assert lib.repair_singlefile_saved_dates() == 1

    fgs = file_groups()
    assert fgs['misread'].download_datetime == datetime(2024, 11, 8, 1, 36, 23, tzinfo=timezone.utc)
    assert fgs['downloaded'].download_datetime == downloaded
    assert fgs['utc'].download_datetime == datetime(2022, 5, 16, 23, 51, 35, tzinfo=timezone.utc)

    # The repair is done; it does not run again.
    fgs['misread'].download_datetime = misread
    test_session.commit()
    assert lib.repair_singlefile_saved_dates() == 0
    assert file_groups()['misread'].download_datetime == misread


@pytest.mark.asyncio
async def test_archive_meta(async_client, test_session, make_files_structure, refresh_files):
    # The start of a typical singlefile html file.
    singlefile_contents = '''<!DOCTYPE html> <html lang="en">
<script data-vue-meta="ssr" type="application/ld+json">
   {"@context":"https://schema.org/","@type":"NewsArticle","headline":"The Headline","description":"The Description","authors":[{"name":"A.B.C.","@type":"Person"}],"datePublished":"2021-06-20T10:00"}
  </script>
</html>'''

    singlefile, *_ = make_files_structure({
        'archive/example.com/2021-10-05-16-20-10_NA.html': singlefile_contents,
        'archive/example.com/2021-10-05-16-20-10_NA.png': None,
        'archive/example.com/2021-10-05-16-20-10_NA.readability.txt': 'article text contents',
        'archive/example.com/2021-10-05-16-20-10_NA.readability.html': '<html></html>',
    })

    await refresh_files()

    archive: Archive = test_session.query(Archive).one()
    assert archive.singlefile_path == singlefile

    assert archive.file_group.author == 'A.B.C.'
    assert archive.file_group.published_datetime == datetime(2021, 6, 20, 10, 0, tzinfo=pytz.utc)

    assert archive.file_group.title == 'The Headline'
    assert archive.file_group.b_text == 'The Description'
    assert archive.file_group.indexed is True


@pytest.mark.asyncio
async def test_refresh_archives_deleted_singlefile(async_client, test_session, make_files_structure,
                                                   singlefile_contents_factory, refresh_files):
    """Removing a Singlefile file from a FileGroup makes that group no longer an Archive."""
    singlefile, readability = make_files_structure({
        '2022-09-04-16-20-11_The Title.html': singlefile_contents_factory(),
        '2022-09-04-16-20-11_The Title.readability.json': '{"url": "https://example.com"}',
    })
    await refresh_files()
    assert test_session.query(FileGroup).one().model == 'archive'
    assert test_session.query(Archive).count() == 1

    # Remove singlefile, FileGroup is no longer an Archive.
    singlefile.unlink()
    await refresh_files()
    assert test_session.query(FileGroup).one().model is None
    assert test_session.query(Archive).count() == 0


@pytest.mark.parametrize(
    'name,expected', [
        ('foo', False),
        ('2000-01-01-00-00-00_Some NA.html', True),
        ('2000-01-01-00-00-00_Some NA.readability.html', False),
        ('2000-01-01-00-00-00_ジにてこちら.html', True),
        ('2000-01-01-00-00-00_Some Title.readability.txt', False),
        ('foo.txt', False),
        ('foo.html', True),
    ]
)
def test_is_singlefile_file(name, expected, make_files_structure, singlefile_contents_factory):
    """A Singlefile may not be created by WROLPi, an HTML file created by Singlefile can also be an Archive."""
    path, = make_files_structure({name: singlefile_contents_factory()})
    assert lib.is_singlefile_file(path) == expected


def test_archive_files(test_directory):
    af = lib.ArchiveFiles(
        singlefile=test_directory / 'archive/singlefile.html',
    )
    assert repr(af) == \
           "<ArchiveFiles singlefile='singlefile.html' readability=None readability_txt=None readability_json=None" \
           " screenshot=None>"


def test_archive_order(test_session, test_directory, archive_factory):
    archive1 = archive_factory()
    archive2 = archive_factory()
    archive2.file_group.download_datetime += timedelta(seconds=10)
    test_session.commit()

    assert archive2 > archive1
    assert archive1 < archive2


@pytest.mark.asyncio
async def test_archive_download_index(test_session, test_directory, image_bytes_factory, async_client):
    """An Archive is indexed when it is downloaded."""
    singlefile = '<html><title>the singlefile</title></html>'
    readability = dict(
        content='<html>the readability</html>',
        textContent='the readability',
    )
    archive = await lib.model_archive_result('https://example.com', singlefile, readability, image_bytes_factory())

    assert isinstance(archive, lib.Archive), 'Did not get an archive'
    assert archive.singlefile_path and archive.singlefile_path.exists() and \
           archive.singlefile_path.read_text() == \
           '<html>\n <head>\n  <title>\n   the singlefile\n  </title>\n </head>\n</html>\n', \
        'Singlefile was not formatted'
    assert archive.readability_path and archive.readability_path.exists()
    assert archive.readability_json_path
    assert archive.readability_txt_path and archive.readability_txt_path.exists()
    assert archive.file_group.d_text == 'the readability', 'Readability text was not indexed'
    assert archive.file_group.title == 'the singlefile', 'Did not get the title from the singlefile'
    assert archive.screenshot_path and archive.screenshot_file and archive.screenshot_path.is_file(), \
        'Did not store the screenshot'


def test_archive_history(test_session, archive_factory, async_client):
    """Archive's can have a history of other archives.

    Archive's with an empty URL are not associated."""
    archive1 = archive_factory(url='https://example.com/1')
    archive2 = archive_factory(url='https://example.com/1')
    archive3 = archive_factory(url='https://example.com/2')
    archive4 = archive_factory()
    archive5 = archive_factory()

    assert archive1 in archive2.history, 'archive1 is a history for archive2'
    assert archive2 in archive1.history, 'archive2 is a history for archive2'
    assert archive3 not in archive1.history and archive3 not in archive2.history, 'archive3 is unique'
    assert not archive3.history, 'archive3 has no history'
    assert not archive4.history, 'archive4 has no history'
    assert not archive5.history, 'archive5 has no history'

    test_session.commit()

    assert archive1 in archive2.history, 'archive1 is a history for archive2'
    assert archive2 in archive1.history, 'archive2 is a history for archive2'
    assert archive3 not in archive1.history and archive3 not in archive2.history, 'archive3 is unique'
    assert not archive3.history, 'archive3 has no history'
    assert not archive4.history, 'archive4 has no history'
    assert not archive5.history, 'archive5 has no history'


METADATA_EXAMPLE_1 = b'''
<!DOCTYPE html>
<html lang="en-US">
</html>
'''

METADATA_EXAMPLE_2 = b'''
<!DOCTYPE html>
<html lang="en-US">
 <head>
  <meta charset="utf-8"/>
  <title>This is not a meta title and should be ignored</title>
  <meta content="en_US" property="og:locale"/>
  <meta content="article" property="og:type"/>
  <meta content="The Title" property="og:title"/>
  <meta name="author" content="Author Name"/>
  <meta content="2023-10-18T04:52:23+00:00" property="article:published_time"/>
  <meta content="2023-10-19T05:53:24+00:00" property="article:modified_time"/>
 </head>
</html>
'''

METADATA_EXAMPLE_3 = b'''
<!DOCTYPE html>
<html lang="en-US">
<title>This is not a meta title and should be ignored</title>
<script class="sf-hidden" type="application/ld+json">
 {"@context":"http://schema.org", "@type":"NewsArticle", "headline":"The headline",
  "datePublished":"2022-09-27T00:40:19.000Z", "dateModified":"2022-09-27T13:43:47.971Z",
  "author":{"@type":"Person", "name":"BOBBY", "jobTitle":""},
  "creator":{"@type":"Person", "name":"OTHER BOBBY", "jobTitle":""},
  "description": "The article description"}
</script>
</html>'''

METADATA_EXAMPLE_4 = b'''
<!DOCTYPE html>
<html lang="en-US">
<title>This is not a meta title and should be ignored</title>
<a href="https://example.com" rel="author">
    Link Author
</a>
</html>'''

METADATA_EXAMPLE_5 = b'''
<!DOCTYPE html>
<html lang="en-US">
<title>This is not a meta title and should be ignored</title>
<a class='g-profile' href='https://example.com' rel='author' title='author profile'>
    <span itemprop='name'>The Span Author</span>
</a>
<time class="post-date updated" itemprop="datePublished" datetime="2023-08-25"></time>
</html>
'''


def test_parse_article_html_metadata():
    """Singlefiles may contain metadata tags."""
    # Metadata can be empty.
    metadata = lib.parse_article_html_metadata(METADATA_EXAMPLE_1)
    assert metadata.title is None
    assert metadata.published_datetime is None
    assert metadata.modified_datetime is None
    assert metadata.description is None
    assert metadata.author is None

    # Data from <meta> elements.
    metadata = lib.parse_article_html_metadata(METADATA_EXAMPLE_2)
    assert metadata.title == 'The Title'
    assert metadata.published_datetime == datetime(2023, 10, 18, 4, 52, 23, tzinfo=utc)
    assert metadata.modified_datetime == datetime(2023, 10, 19, 5, 53, 24, tzinfo=utc)
    assert metadata.description is None
    assert metadata.author == 'Author Name'

    # Data from Article structured data
    # https://developers.google.com/search/docs/appearance/structured-data/article
    metadata = lib.parse_article_html_metadata(METADATA_EXAMPLE_3)
    assert metadata.title == 'The headline'
    assert metadata.published_datetime == datetime(2022, 9, 27, 0, 40, 19, tzinfo=utc)
    assert metadata.modified_datetime == datetime(2022, 9, 27, 13, 43, 47, 971000, tzinfo=utc)
    assert metadata.description == 'The article description'
    assert metadata.author == 'BOBBY'

    # Data from author anchor element.
    metadata = lib.parse_article_html_metadata(METADATA_EXAMPLE_4)
    assert metadata.author == 'Link Author'

    # author anchor may be a list.
    metadata = lib.parse_article_html_metadata(METADATA_EXAMPLE_5)
    assert metadata.author == 'The Span Author'
    assert metadata.published_datetime == datetime(2023, 8, 25, tzinfo=pytz.UTC)


@pytest.mark.parametrize(
    'html,expected',
    [
        ('<meta content="2023-10-18T04:52:23+00:00" property="article:published_time"/>',
         datetime(2023, 10, 18, 4, 52, 23, tzinfo=utc)),
        ('<abbr class="published" itemprop="datePublished" title="2022-03-17T03:00:00-07:00">March 17, 2022</abbr>',
         datetime(2022, 3, 17, 10, 0, tzinfo=pytz.UTC)),
        ('<meta name="article.published" content="2023-04-04T21:52:00.000Z">',
         datetime(2023, 4, 4, 21, 52, tzinfo=pytz.UTC)),
        ('<meta itemprop="datePublished" content="2023-04-04T21:52:00.000Z">',
         datetime(2023, 4, 4, 21, 52, tzinfo=pytz.UTC)),
        (
                '<abbr class="published" itemprop="datePublished" title="2013-03-05T00:12:00-06:00">3/05/2013 12:12:00 AM</abbr>',
                datetime(2013, 3, 5, 6, 12, tzinfo=pytz.UTC)),
        ('<time class="post-date updated" itemprop="datePublished" datetime="2023-08-25"></time>',
         datetime(2023, 8, 25, tzinfo=pytz.UTC))
    ])
def test_published_datetime(html, expected):
    assert lib.parse_article_html_metadata(html).published_datetime == expected


@pytest.mark.parametrize(
    'html,expected',
    [
        ('<meta name="article.updated" content="2023-04-04T21:52:00.000Z">',
         datetime(2023, 4, 4, 21, 52, tzinfo=pytz.UTC)),
        ('<meta content="2023-10-19T05:53:24+00:00" property="article:modified_time"/>',
         datetime(2023, 10, 19, 5, 53, 24, tzinfo=utc)),
        (
                '<script class="sf-hidden" type="application/ld+json">{"@context":"http://schema.org","dateModified":"2022-09-27T13:43:47.971Z"}</script>',
                datetime(2022, 9, 27, 13, 43, 47, 971000, tzinfo=utc)),
    ])
def test_modified_datetime(html, expected):
    assert lib.parse_article_html_metadata(html).modified_datetime == expected


@pytest.mark.parametrize(
    'html,expected',
    [
        ('<a href="#" rel="author"><span itemprop="name">Linda</span></a></span>', 'Linda'),
        ('<meta content="Billy" property="article:author"/>', 'Billy'),
    ]
)
def test_author(html, expected):
    assert lib.parse_article_html_metadata(html).author == expected


SINGLEFILE_EXAMPLE_1 = b'''<!DOCTYPE html>
<html class="fonts-loaded" lang="en">
 <!--
 Page saved with SingleFile 
 url: https://www.example.com 
 saved date: Tue Oct 31 2023 15:57:19 GMT+0000 (Coordinated Universal Time)
-->
 <head>'''


def test_get_url_from_singlefile():
    assert lib.get_url_from_singlefile(SINGLEFILE_EXAMPLE_1) == 'https://www.example.com'


@pytest.mark.asyncio
async def test_get_custom_archive_directory(async_client, test_directory, test_wrolpi_config):
    """Custom directory can be used for archive directory."""
    # Default location.
    assert lib.get_archive_directory() == (test_directory / 'archive')

    get_wrolpi_config().archive_destination = 'custom/archives'

    assert lib.get_archive_directory() == (test_directory / 'custom/archives')


@pytest.mark.asyncio
async def test_detect_domain_directory_single_archive(async_client, test_directory, test_session, archive_factory):
    """detect_domain_directory should detect directory when all archives are in same location."""
    from modules.archive.lib import detect_domain_directory

    # Create an archive using the factory (which creates a domain collection and auto-detects directory)
    archive = archive_factory(domain='example.com', url='https://example.com/page1')
    test_session.flush()

    # Get the collection
    collection = archive.collection
    assert collection is not None
    assert collection.name == 'example.com'

    # Directory should have been auto-detected during collection creation
    assert collection.directory is not None
    assert 'example.com' in str(collection.directory)

    # Verify detect_domain_directory also returns the same result
    detected = detect_domain_directory(collection)
    assert detected is not None
    assert str(detected) == 'archive/example.com'


@pytest.mark.asyncio
async def test_detect_domain_directory_multiple_archives(async_client, test_directory, test_session, archive_factory):
    """detect_domain_directory should detect common directory for multiple archives."""
    from modules.archive.lib import detect_domain_directory

    # Create multiple archives in the same domain
    archive1 = archive_factory(domain='test.com', url='https://test.com/page1')
    archive2 = archive_factory(domain='test.com', url='https://test.com/page2')
    test_session.flush()

    # Get the collection
    collection = archive1.collection
    assert collection is not None
    assert collection.name == 'test.com'

    # Detect directory
    detected = detect_domain_directory(collection)
    assert detected is not None
    assert str(detected) == 'archive/test.com'


def test_detect_domain_directory_no_archives(test_directory, test_session):
    """detect_domain_directory should return None when collection has no archives."""
    from modules.archive.lib import detect_domain_directory

    # Create a domain collection without archives
    collection = Collection(name='empty.com', kind='domain', directory=None)
    test_session.add(collection)
    test_session.flush()

    # Detect directory - should return None
    detected = detect_domain_directory(collection)
    assert detected is None


@pytest.mark.asyncio
async def test_get_or_create_domain_collection_auto_detects_directory(async_client, test_directory, test_session,
                                                                      archive_factory):
    """get_or_create_domain_collection should auto-detect directory for existing archives."""
    from modules.archive.lib import get_or_create_domain_collection

    # Create an archive first
    archive = archive_factory(domain='autodetect.com', url='https://autodetect.com/page1')
    test_session.commit()

    # The collection should have been created with no directory initially
    # Call get_or_create_domain_collection again - should auto-detect
    collection = get_or_create_domain_collection(test_session, 'https://autodetect.com/page2')
    assert collection.directory is not None
    assert 'autodetect.com' in str(collection.directory)


@pytest.mark.asyncio
async def test_update_domain_directories(async_client, test_directory, test_session, archive_factory):
    """update_domain_directories should fix existing collections that lost their directory."""
    from modules.archive.lib import update_domain_directories

    # Create an archive - this creates a collection WITH directory (auto-detected)
    archive = archive_factory(domain='needsdir.com', url='https://needsdir.com/page1')
    test_session.commit()

    # Manually clear the directory to simulate legacy data
    collection = archive.collection
    collection.directory = None
    test_session.commit()

    # Verify collection has no directory
    assert collection.directory is None

    # Run update
    count = update_domain_directories(test_session)
    assert count == 1

    # Verify directory was set
    test_session.expire(collection)
    assert collection.directory is not None
    assert 'needsdir.com' in str(collection.directory)


def test_collection_unique_name_kind_constraint(test_directory, test_session, archive_factory):
    """Collections should have unique (name, kind) combinations.

    This prevents duplicate domain collections with the same name.
    """
    from sqlalchemy.exc import IntegrityError

    # Create an archive which creates a domain collection
    archive = archive_factory(domain='uniquetest.com')
    test_session.commit()

    # Get the original collection
    original_collection = archive.collection
    assert original_collection is not None

    # Attempting to create a duplicate domain collection should fail
    duplicate_collection = Collection(name='uniquetest.com', kind='domain', directory=None)
    test_session.add(duplicate_collection)

    with pytest.raises(IntegrityError) as exc_info:
        test_session.commit()

    # Verify it's the unique constraint violation.  SQLite reports the columns, not the
    # constraint name (Postgres reported 'uq_collection_name_kind').
    assert 'collection.name, collection.kind' in str(exc_info.value)
    test_session.rollback()


def test_search_archives_by_domain(test_directory, test_session, archive_factory):
    """search_archives should correctly filter by domain collection name."""
    from modules.archive.lib import search_archives

    # Create archives in different domains
    archive1 = archive_factory(domain='searchtest.com')
    archive2 = archive_factory(domain='searchtest.com')
    archive3 = archive_factory(domain='other.com')
    test_session.commit()

    # Search for archives in searchtest.com
    file_groups, total = search_archives(
        test_session,
        search_str=None,
        domain='searchtest.com',
        limit=10,
        offset=0,
        order=None,
        tag_names=None
    )

    # Should return only archives from searchtest.com
    assert total == 2
    assert len(file_groups) == 2

    # Search for other.com
    file_groups, total = search_archives(
        test_session,
        search_str=None,
        domain='other.com',
        limit=10,
        offset=0,
        order=None,
        tag_names=None
    )

    assert total == 1
    assert len(file_groups) == 1


def test_link_domain_and_downloads(test_session, test_download_manager):
    """Test that downloads are linked to domain collections."""
    from wrolpi.downloader import Download
    from modules.archive.lib import link_domain_and_downloads

    # Create a domain collection with a directory
    collection = Collection(name='example.com', kind='domain', directory='archive/example.com')
    test_session.add(collection)
    test_session.commit()

    # 1. Download with matching destination directory (exact match)
    download1 = Download(
        url='https://other.com/rss',
        downloader='rss',
        sub_downloader='archive',
        frequency=86400,
        settings={'destination': 'archive/example.com'}
    )
    # 2. Download with destination in subdirectory
    download2 = Download(
        url='https://other.com/rss2',
        downloader='rss',
        sub_downloader='archive',
        frequency=86400,
        settings={'destination': 'archive/example.com/2025/01'}
    )
    # 3. RSS download with archive sub_downloader (matches by URL domain)
    download3 = Download(
        url='https://example.com/feed.xml',
        downloader='rss',
        sub_downloader='archive',
        frequency=86400
    )
    # 4. Download without frequency (should NOT be linked)
    download4 = Download(url='https://example.com/once', downloader='archive')
    test_session.add_all([download1, download2, download3, download4])
    test_session.commit()

    assert not any(d.collection_id for d in [download1, download2, download3, download4])

    link_domain_and_downloads(test_session)

    assert download1.collection_id == collection.id  # matched by directory (exact)
    assert download2.collection_id == collection.id  # matched by subdirectory
    assert download3.collection_id == collection.id  # matched by URL domain
    assert download4.collection_id is None  # no frequency = no link


def test_link_domain_and_downloads_by_destination(test_session, test_directory, test_download_manager):
    """A Download whose destination column is in a domain's directory is linked to that domain (the destination was
    moved out of the settings)."""
    from wrolpi.downloader import Download
    from modules.archive.lib import link_domain_and_downloads

    collection = Collection(name='example.com', kind='domain', directory=test_directory / 'archive/example.com')
    test_session.add(collection)
    exact = Download(url='https://other.com/rss', downloader='rss', sub_downloader='archive', frequency=86400,
                     destination=test_directory / 'archive/example.com')
    below = Download(url='https://other.com/rss2', downloader='rss', sub_downloader='archive', frequency=86400,
                     destination=test_directory / 'archive/example.com/2025/01')
    similar = Download(url='https://other.com/rss3', downloader='rss', sub_downloader='archive', frequency=86400,
                       destination=test_directory / 'archive/example.com.au')
    test_session.add_all([exact, below, similar])
    test_session.commit()

    link_domain_and_downloads(test_session)

    assert exact.collection_id == collection.id
    assert below.collection_id == collection.id
    assert similar.collection_id is None


@skip_ci
@pytest.mark.asyncio
async def test_archive_download_uses_domain_collection_directory(async_client, test_session, test_directory, fake_now,
                                                                 monkeypatch):
    """
    Test that archive downloads use the domain collection's directory when no explicit destination is set.

    If a domain collection exists with a directory (e.g., archive/News/example.com), new archives for that
    domain should be created in that directory instead of the default archive/<domain> directory.
    """
    from modules.archive import archive_downloader
    from wrolpi.downloader import Download

    fake_now(datetime(2024, 1, 15))

    # Mock request_archive to return fake archive data
    singlefile, readability, screenshot = make_fake_archive_result()

    async def mock_request_archive(url, compress=False):
        return singlefile, readability, screenshot

    # Patch in both lib and archive module (imported at module level)
    import modules.archive
    monkeypatch.setattr(lib, 'request_archive', mock_request_archive)
    monkeypatch.setattr(modules.archive, 'request_archive', mock_request_archive)

    # Create a domain collection with a custom directory
    custom_dir = test_directory / 'archive/News/example.com'
    custom_dir.mkdir(parents=True, exist_ok=True)
    collection = Collection(name='example.com', kind='domain', directory=custom_dir)
    test_session.add(collection)
    test_session.commit()

    # Create a download without an explicit destination
    download = Download(
        url='https://example.com/some-article',
        downloader='archive',
        settings={},  # No destination set
    )
    test_session.add(download)
    test_session.commit()

    # Drive the phase-split API directly: prepare → execute → finalize.
    from modules.archive.conftest import make_test_ctx
    prepared = archive_downloader.prepare_download(test_session, download)
    executed = await archive_downloader.execute_download(prepared, make_test_ctx())
    result = archive_downloader.finalize_download(test_session, download, executed)
    test_session.commit()

    # Verify the download was successful
    assert result.success is True

    # Verify the archive was created in the domain collection's directory
    archives = test_session.query(Archive).all()
    assert len(archives) == 1
    archive = archives[0]

    # The archive should be in the custom directory, not archive/example.com
    assert str(archive.singlefile_path).startswith(str(custom_dir)), \
        f"Archive should be in domain collection directory {custom_dir}, but was in {archive.singlefile_path.parent}"


@skip_ci
@pytest.mark.asyncio
async def test_archive_download_explicit_destination_overrides_domain_collection(
        async_client, test_session, test_directory, fake_now, monkeypatch):
    """
    Test that an explicit destination in settings overrides the domain collection's directory.

    Even if a domain collection exists with a directory, if the download has an explicit
    settings['destination'], that should be used instead.
    """
    from modules.archive import archive_downloader
    from wrolpi.downloader import Download

    fake_now(datetime(2024, 1, 15))

    # Mock request_archive to return fake archive data
    singlefile, readability, screenshot = make_fake_archive_result()

    async def mock_request_archive(url, compress=False):
        return singlefile, readability, screenshot

    # Patch in both lib and archive module (imported at module level)
    import modules.archive
    monkeypatch.setattr(lib, 'request_archive', mock_request_archive)
    monkeypatch.setattr(modules.archive, 'request_archive', mock_request_archive)

    # Create a domain collection with a custom directory
    collection_dir = test_directory / 'archive/News/example.com'
    collection_dir.mkdir(parents=True, exist_ok=True)
    collection = Collection(name='example.com', kind='domain', directory=collection_dir)
    test_session.add(collection)
    test_session.commit()

    # Create an explicit destination that's different from the collection directory
    explicit_dir = test_directory / 'archive/Special/my-archives'
    explicit_dir.mkdir(parents=True, exist_ok=True)

    # Create a download with an explicit destination
    download = Download(
        url='https://example.com/some-article',
        downloader='archive',
        settings={'destination': str(explicit_dir)},  # Explicit destination
    )
    test_session.add(download)
    test_session.commit()

    # Drive the phase-split API directly: prepare → execute → finalize.
    from modules.archive.conftest import make_test_ctx
    prepared = archive_downloader.prepare_download(test_session, download)
    executed = await archive_downloader.execute_download(prepared, make_test_ctx())
    result = archive_downloader.finalize_download(test_session, download, executed)
    test_session.commit()

    # Verify the download was successful
    assert result.success is True

    # Verify the archive was created in the explicit destination, NOT the collection directory
    archives = test_session.query(Archive).all()
    assert len(archives) == 1
    archive = archives[0]

    assert str(archive.singlefile_path).startswith(str(explicit_dir)), \
        f"Archive should be in explicit destination {explicit_dir}, but was in {archive.singlefile_path.parent}"
    assert not str(archive.singlefile_path).startswith(str(collection_dir)), \
        f"Archive should NOT be in collection directory {collection_dir}"


@pytest.mark.asyncio
async def test_archive_downloader_config_import(test_directory, async_client):
    """ArchiveDownloaderConfig imports successfully and provides correct defaults."""
    config = get_archive_downloader_config()

    # Default format should end with .%(ext)s
    assert config.file_name_format.endswith('.%(ext)s')
    assert config.file_name_format == '%(download_datetime)s_%(title)s.%(ext)s'

    # Create a valid config file
    config_path = config.get_file()
    config_path.parent.mkdir(parents=True, exist_ok=True)
    config_path.write_text('version: 0\nfile_name_format: "%(download_datetime)s_%(title)s.%(ext)s"\n')

    # Import should succeed and set successful_import
    config.import_config()
    assert config.successful_import is True


def test_archive_downloader_config_validation():
    """ArchiveDownloaderConfig validator rejects invalid formats."""
    # Valid format
    ArchiveDownloaderConfigValidator(file_name_format='%(download_date)s_%(title)s.%(ext)s')

    # Invalid format (missing .%(ext)s)
    with pytest.raises(ValueError, match='must end with'):
        ArchiveDownloaderConfigValidator(file_name_format='%(download_date)s_%(title)s')

    # Valid browser_args (JSON array)
    ArchiveDownloaderConfigValidator(
        file_name_format='%(title)s.%(ext)s',
        browser_args='["--no-sandbox", "--headless"]'
    )

    # Invalid browser_args (not valid JSON)
    with pytest.raises(ValueError, match='must be valid JSON'):
        ArchiveDownloaderConfigValidator(
            file_name_format='%(title)s.%(ext)s',
            browser_args='not valid json'
        )

    # Invalid browser_args (not an array)
    with pytest.raises(ValueError, match='must be a JSON array'):
        ArchiveDownloaderConfigValidator(
            file_name_format='%(title)s.%(ext)s',
            browser_args='{"key": "value"}'
        )


def test_format_archive_filename_default(test_directory, fake_now):
    """format_archive_filename uses default format."""
    result = format_archive_filename('My Article')

    # Default format: %(download_datetime)s_%(title)s.%(ext)s
    assert result == '2000-01-01-00-00-00_My Article.html'


def test_format_archive_filename_with_domain(test_directory, fake_now):
    """format_archive_filename includes domain variable."""
    # Temporarily change the format to include domain
    config = get_archive_downloader_config()
    original_format = config._config['file_name_format']
    config._config['file_name_format'] = '%(domain)s_%(title)s.%(ext)s'

    try:
        result = format_archive_filename('My Article', domain='example.com')
        assert result == 'example.com_My Article.html'
    finally:
        config._config['file_name_format'] = original_format


def test_format_archive_filename_with_year_subdirectory(test_directory, fake_now):
    """format_archive_filename can include subdirectories in format."""
    config = get_archive_downloader_config()
    original_format = config._config['file_name_format']
    config._config['file_name_format'] = '%(download_year)s/%(download_date)s_%(title)s.%(ext)s'

    try:
        result = format_archive_filename('My Article')
        # Should include year subdirectory
        assert result == '2000/2000-01-01_My Article.html'
    finally:
        config._config['file_name_format'] = original_format


def test_format_archive_filename_escapes_special_chars(test_directory, fake_now):
    """format_archive_filename escapes special characters in title."""
    result = format_archive_filename('My/Article:With*Special?Chars')

    # Special characters should be escaped
    assert '/' not in result.split('/')[-1]  # No slashes in filename part
    assert '2000-01-01' in result
    assert result.endswith('.html')


@pytest.mark.asyncio
async def test_get_statistics(test_session, archive_factory):
    """get_statistics returns archive and domain statistics."""
    from modules.archive.lib import get_statistics

    # Can get statistics in empty DB.
    result = await get_statistics(test_session)
    assert 'statistics' in result
    assert 'archives' in result['statistics']
    assert 'domains' in result['statistics']
    assert 'historical' in result['statistics']

    # Empty stats should have zero values.
    stats = result['statistics']
    assert stats['archives']['archives'] == 0
    assert stats['domains']['domains'] == 0

    # Create some archives.
    archive1 = archive_factory('example.com')
    archive2 = archive_factory('example.com')
    archive3 = archive_factory('example.org')
    archive4 = archive_factory()  # No domain
    test_session.commit()

    result = await get_statistics(test_session)
    stats = result['statistics']

    # Should count all 4 archives.
    assert stats['archives']['archives'] == 4
    # Should count 2 domain collections (example.com and example.org).
    assert stats['domains']['domains'] == 2


@pytest.mark.asyncio
async def test_singlefile_to_archive_result_outlives_its_session(async_client, test_session, test_directory,
                                                                  singlefile_contents_factory):
    """`singlefile_to_archive`'s Session is gone once it returns (production frees it), so its result must not
    need it."""
    import sqlalchemy.orm
    from unittest import mock

    singlefile = singlefile_contents_factory(title='upload title', url='https://example.com/upload').encode()
    artifacts = lib.SinglefileArtifacts(url='https://example.com/upload', singlefile=singlefile,
                                        readability=None, screenshot=None)
    with mock.patch('modules.archive.lib.extract_singlefile_artifacts', return_value=artifacts):
        uploaded = await lib.singlefile_to_archive(singlefile)
    sqlalchemy.orm.close_all_sessions()

    location, name, url = uploaded.location, uploaded.name, uploaded.url
    archive = test_session.query(Archive).one()
    assert location == archive.location
    assert name == 'upload title'
    assert url == 'https://example.com/upload'
    assert uploaded.file_group_id == archive.file_group_id


@pytest.mark.parametrize('deep', [False, True])
def test_search_archives_count_is_driven_by_the_search_index(test_directory, test_session, archive_factory, deep):
    """A search's total starts from the FTS matches.  Starting from `archive` runs the MATCH once per Archive,
    which took ~20 seconds on a Pi with 128k Archives."""
    archive_factory(domain='example.com', title='solar panels', contents='solar panels')
    archive_factory(domain='example.com', title='wind turbines', contents='wind turbines')
    test_session.commit()

    counts = []

    def capture_count(session, statement, params):
        counts.append((statement, params))
        return count_file_groups(session, statement, params)

    with mock.patch('modules.archive.lib.count_file_groups', capture_count):
        file_groups, total = lib.search_archives(test_session, search_str='solar', domain=None, limit=10, offset=0,
                                                 order=None, tag_names=None, deep=deep)
    assert total == 1 and len(file_groups) == 1

    (statement, params), = counts
    plan = [row[3] for row in test_session.execute(text(f'EXPLAIN QUERY PLAN {statement}'), params).fetchall()]
    assert plan[0].startswith('SCAN fts VIRTUAL TABLE'), plan


@pytest.mark.asyncio
async def test_html_to_readability_data_uris(test_directory):
    """readability-extractor never receives inline base64 data, which makes it take minutes on a page of
    inline images; the data is restored in the readability output."""
    png = 'iVBORw0KGgo' + 'A' * 5000 + '=='
    webp = 'UklGRiQAAABXRUJQ' + 'B' * 5000
    html = f'<html><head><style>.x {{ background: url(data:image/png;base64,{png}) }}</style></head>' \
           f'<body><article><p>Some text</p>' \
           f'<img src="data:image/png;base64,{png}">' \
           f'<img src=\'data:image/webp;base64,{webp}\'>' \
           f'<img src="data:image/svg+xml,%3Csvg%3E%3C/svg%3E">' \
           f'</article></body></html>'

    received = list()

    async def fake_run_command(cmd, *args, **kwargs):
        # Echo the images back like readability-extractor does in `content`.
        page = pathlib.Path(cmd[3]).read_text()
        received.append(page)
        content = ''.join(re.findall(r'<img[^>]*>', page))
        stdout = json.dumps({'title': 'Title', 'content': content, 'textContent': 'Some text'}).encode()
        return CommandResult(return_code=0, cancelled=False, stdout=stdout, stderr=b'', elapsed=0)

    with mock.patch('modules.archive.lib.run_command', fake_run_command):
        readability = await lib.html_to_readability(html, 'https://example.com')

    # readability-extractor got the page without the base64 payloads.
    assert png not in received[0] and webp not in received[0]
    assert 'Some text' in received[0]
    assert 'data:image/svg+xml,%3Csvg%3E%3C/svg%3E' in received[0], 'non-base64 data URI should be untouched'

    # The images were restored in the output.
    assert readability['content'] == f'<img src="data:image/png;base64,{png}">' \
                                     f'<img src=\'data:image/webp;base64,{webp}\'>' \
                                     f'<img src="data:image/svg+xml,%3Csvg%3E%3C/svg%3E">'
    assert readability['textContent'] == 'Some text'
    assert readability['title'] == 'Title'


@pytest.mark.asyncio
async def test_html_to_readability_forged_placeholder():
    """Placeholder-like text already in a page is not treated as a placeholder; it cannot index past the
    payloads or duplicate an image into the output."""
    png = 'iVBORw0KGgo' + 'A' * 5000
    html = f'<p>WROLPIDATAURI0X WROLPIDATAURI99X</p><img src="data:image/png;base64,{png}">'

    async def fake_run_command(cmd, *args, **kwargs):
        page = pathlib.Path(cmd[3]).read_text()
        stdout = json.dumps({'content': page, 'textContent': 'WROLPIDATAURI0X WROLPIDATAURI99X'}).encode()
        return CommandResult(return_code=0, cancelled=False, stdout=stdout, stderr=b'', elapsed=0)

    with mock.patch('modules.archive.lib.run_command', fake_run_command):
        readability = await lib.html_to_readability(html, 'https://example.com')

    assert readability['content'] == html
    assert readability['content'].count(png) == 1
    assert readability['textContent'] == 'WROLPIDATAURI0X WROLPIDATAURI99X'


def test_strip_data_uris_placeholder_length():
    """Readability's _fixLazyImages removes an <img> `src` whose base64 payload is shorter than 133 characters
    when another attribute names an image file (e.g. alt="cat.png"), so a placeholder must be at least that
    long or the image is lost before it can be restored."""
    png = 'iVBORw0KGgo' + 'A' * 4000
    html, payloads = lib.strip_data_uris(f'<img src="data:image/png;base64,{png}" alt="cat.png">'.encode())

    (placeholder,) = payloads
    assert len(placeholder) >= 133
    assert html == f'<img src="data:image/png;base64,{placeholder}" alt="cat.png">'.encode()
    assert lib.restore_data_uris({'content': html.decode()}, payloads)['content'] \
           == f'<img src="data:image/png;base64,{png}" alt="cat.png">'