diff --git a/modules/archive/lib.py b/modules/archive/lib.py index 6f8921ff..36c1051f 100644 --- a/modules/archive/lib.py +++ b/modules/archive/lib.py @@ -5,8 +5,9 @@ import pathlib import re import subprocess import tempfile +from datetime import datetime from itertools import groupby -from typing import Iterator +from typing import Iterator, Optional import requests from bs4 import BeautifulSoup @@ -14,8 +15,8 @@ from selenium import webdriver from sqlalchemy.orm import Session from modules.archive.models import Domain, Archive -from wrolpi.common import get_media_directory, logger, chunks, extract_domain, chdir -from wrolpi.dates import now, strptime_ms, strftime_ms, Seconds +from wrolpi.common import get_media_directory, logger, chunks, extract_domain, chdir, escape_file_name, walk +from wrolpi.dates import now, Seconds, local_timezone from wrolpi.db import get_db_session, get_db_curs, get_ranked_models from wrolpi.errors import InvalidDomain, UnknownURL, InvalidArchive from wrolpi.vars import DOCKERIZED, PYTEST @@ -47,18 +48,20 @@ def get_domain_directory(url: str) -> pathlib.Path: return directory -def get_new_archive_files(url: str): +def get_new_archive_files(url: str, title: Optional[str]): """ Create a list of archive files using a shared name schema. Raise an error if any of them exist. """ directory = get_domain_directory(url) - dt = strftime_ms(now()) + dt = now().strftime('%Y-%m-%d-%H-%M-%S') - singlefile_path = directory / f'{dt}.html' - readability_path = directory / f'{dt}-readability.html' - readability_txt_path = directory / f'{dt}-readability.txt' - readability_json_path = directory / f'{dt}-readability.json' - screenshot_path = directory / f'{dt}.png' + title = escape_file_name(title or 'NA') + prefix = f'{dt}_{title}' + singlefile_path = directory / f'{prefix}.html' + readability_path = directory / f'{prefix}-readability.html' + readability_txt_path = directory / f'{prefix}-readability.txt' + readability_json_path = directory / f'{prefix}-readability.json' + screenshot_path = directory / f'{prefix}.png' ret = (singlefile_path, readability_path, readability_txt_path, readability_json_path, screenshot_path) @@ -196,8 +199,6 @@ def do_archive(url: str) -> Archive: record in the DB. Create Domain/URL if missing. """ logger.info(f'Archiving {url}') - singlefile_path, readability_path, readability_txt_path, readability_json_path, screenshot_path = \ - get_new_archive_files(url) if DOCKERIZED or PYTEST: # Perform the archive in the Archive docker container. (Typically in the development environment). @@ -206,6 +207,30 @@ def do_archive(url: str) -> Archive: # Perform the archive using locally installed executables. singlefile, readability, screenshot = local_archive(url) + # First try to get the title from Readability. + title = readability.get('title') if readability else None + + if not title and singlefile: + # Try to get the title ourselves from the HTML. + title = get_title_from_html(singlefile, url=url) + if readability: + # Readability could not find title, lets use ours. + readability['title'] = title + + singlefile_path, readability_path, readability_txt_path, readability_json_path, screenshot_path = \ + get_new_archive_files(url, title) + + if readability: + # Write the readability parts to their own files. Write what is left after pops to the JSON file. + with readability_path.open('wt') as fh: + fh.write(readability.pop('content')) + with readability_txt_path.open('wt') as fh: + readability_txt = readability.pop('textContent') + fh.write(readability_txt) + else: + # No readability was returned, so there are no files. + readability_txt = readability_path = readability_txt_path = None + # Store the single-file HTML in its own file. with singlefile_path.open('wt') as fh: fh.write(singlefile) @@ -215,27 +240,8 @@ def do_archive(url: str) -> Archive: else: screenshot_path = None - # Store the Readability into separate files. This allows the user to view text-only or html articles. - title = None - readability_txt = None - if readability: - title = readability.get('title') - - # Write the readability parts to their own files. Write what is left after pops to the JSON file. - with readability_path.open('wt') as fh: - fh.write(readability.pop('content')) - with readability_txt_path.open('wt') as fh: - readability_txt = readability.pop('textContent') - fh.write(readability_txt) - else: - readability_path = readability_txt_path = None - # Always write a JSON file that contains at least the URL. readability = readability or {} - # Use the Readability title, or try and extract one from singlefile. - if not title and singlefile: - title = get_title_from_html(singlefile, url=url) - readability['title'] = title readability['url'] = url with readability_json_path.open('wt') as fh: fh.write(json.dumps(readability)) @@ -304,6 +310,13 @@ def delete_archive(archive_id: int): archive.delete() +def archive_strptime(dt: str) -> datetime: + try: + return local_timezone(datetime.strptime(dt, '%Y-%m-%d-%H-%M-%S')) + except ValueError: + return local_timezone(datetime.strptime(dt, '%Y-%m-%d %H:%M:%S')) + + def group_archive_files(files: Iterator[pathlib.Path]) -> groupby: """ Group archive files by their timestamp. @@ -311,10 +324,10 @@ def group_archive_files(files: Iterator[pathlib.Path]) -> groupby: # groupby requires the files to be sorted. files = sorted(files) # Group archive files by their datetime at the beginning of the file. - groups = groupby(files, key=lambda i: i.name[:26]) + groups = groupby(files, key=lambda i: i.name[:19]) for dt, files in groups: try: - dt = strptime_ms(dt) + dt = archive_strptime(dt) except ValueError: logger.info(f'Ignoring invalid archives of {dt=}') continue @@ -345,15 +358,22 @@ def group_archive_files(files: Iterator[pathlib.Path]) -> groupby: ARCHIVE_MATCHER = re.compile(r'\d{4}-\d\d-\d\d (\d\d:){2}\d\d\.\d{6}.*$') +NEW_ARCHIVE_MATCHER = re.compile(r'\d{4}-(\d{2}-){4}\d\d_.*$') ARCHIVE_SUFFIXES = {'.txt', '.html', '.json', '.png', '.jpg', '.jpeg'} def is_archive_file(path: pathlib.Path) -> bool: """ - Archive files are expected to start with the following: %Y-%m-%d %H:%M:%S.%f + Archive files are expected to start with the following: %Y-%m-%d-%H-%M-%S they must have one of the following suffixes: .txt, .html, .json, .png, .jpg, .jpeg """ - return path.is_file() and path.suffix.lower() in ARCHIVE_SUFFIXES and bool(ARCHIVE_MATCHER.match(path.name)) + if not path.is_file(): + return False + if path.suffix.lower() not in ARCHIVE_SUFFIXES: + return False + if bool(ARCHIVE_MATCHER.match(path.name)) or bool(NEW_ARCHIVE_MATCHER.match(path.name)): + return True + return False def _refresh_archives(): @@ -366,18 +386,18 @@ def _refresh_archives(): singlefile_paths = set() for domain_directory in filter(lambda i: i.is_dir(), archive_directory.iterdir()): logger.debug(f'Refreshing directory: {domain_directory}') - archives_files = filter(is_archive_file, domain_directory.iterdir()) + archives_files = filter(is_archive_file, walk(domain_directory)) archive_groups = group_archive_files(archives_files) archive_count = 0 for chunk in chunks(archive_groups, 20): - archive_count += 1 with get_db_session(commit=True) as session: for dt, files in chunk: + archive_count += 1 singlefile_paths.add(str(files[0])) upsert_archive(dt, files, session) if archive_count: - logger.info(f'Inserted/updated {archive_count} archives') + logger.info(f'Inserted/updated {archive_count} archives in {domain_directory}') singlefile_paths = list(singlefile_paths) with get_db_curs(commit=True) as curs: diff --git a/modules/archive/test/test_lib.py b/modules/archive/test/test_lib.py index e868be84..3c7c9bc2 100644 --- a/modules/archive/test/test_lib.py +++ b/modules/archive/test/test_lib.py @@ -5,13 +5,13 @@ from datetime import datetime import mock from modules.archive.lib import get_or_create_domain, _refresh_archives, \ - get_new_archive_files, delete_archive, do_archive, get_domains + get_new_archive_files, delete_archive, do_archive, get_domains, get_archive_directory, group_archive_files from modules.archive.models import Archive, Domain -from wrolpi.common import get_media_directory +from wrolpi.dates import local_timezone from wrolpi.db import get_db_session from wrolpi.media_path import MediaPath from wrolpi.root_api import CustomJSONEncoder -from wrolpi.test.common import TestAPI, wrap_test_db, wrap_media_directory +from wrolpi.test.common import TestAPI, wrap_test_db def make_fake_request_archive(readability=True, screenshot=True, title=True): @@ -34,23 +34,6 @@ class TestArchive(TestAPI): super().setUp() (pathlib.Path(self.tmp_dir.name) / 'archive').mkdir(exist_ok=True) - @mock.patch('modules.archive.lib.now', lambda: datetime(2001, 1, 1)) - def test_get_new_archive_files(self): - s, r, t, j, c = map(str, get_new_archive_files('https://example.com/two')) - assert str(s).endswith('archive/example.com/2001-01-01 00:00:00.000000.html') - assert str(r).endswith('archive/example.com/2001-01-01 00:00:00.000000-readability.html') - assert str(t).endswith('archive/example.com/2001-01-01 00:00:00.000000-readability.txt') - assert str(j).endswith('archive/example.com/2001-01-01 00:00:00.000000-readability.json') - assert str(c).endswith('archive/example.com/2001-01-01 00:00:00.000000.png') - - s, r, t, j, c = get_new_archive_files('https://www.example.com/one') - # Leading www. is removed. - assert str(s).endswith('archive/example.com/2001-01-01 00:00:00.000000.html') - assert str(r).endswith('archive/example.com/2001-01-01 00:00:00.000000-readability.html') - assert str(t).endswith('archive/example.com/2001-01-01 00:00:00.000000-readability.txt') - assert str(j).endswith('archive/example.com/2001-01-01 00:00:00.000000-readability.json') - assert str(c).endswith('archive/example.com/2001-01-01 00:00:00.000000.png') - @wrap_test_db def test_no_screenshot(self): with mock.patch('modules.archive.lib.request_archive', make_fake_request_archive(screenshot=False)): @@ -103,42 +86,6 @@ class TestArchive(TestAPI): except ValueError as e: self.assertIn('relative', str(e), f'Relative path error was not raised') - @wrap_test_db - def test_refresh_archives(self): - with wrap_media_directory(): - archive_directory = get_media_directory() / 'archive' - archive_directory.mkdir() - - # Make some test files to refresh. - example = archive_directory / 'example.com' - example.mkdir() - (example / '2021-10-05 16:20:10.346823.html').touch() - (example / '2021-10-05 16:20:10.346823.png').touch() - (example / '2021-10-05 16:20:10.346823-readability.txt').touch() - (example / '2021-10-05 16:20:10.346823-readability.json').touch() - (example / '2021-10-05 16:20:10.346823-readability.html').touch() - with (example / '2021-10-05 16:20:10.346823-readability.json').open('wt') as fh: - json.dump({'url': 'foo'}, fh) - - # These should log an error because they are missing the singlefile path. - (example / '2021-10-05 16:20:10.346824-readability.html').touch() - (example / '2021-10-05 16:20:10.346825-readability.json').touch() - # These are not archives and should be ignored. - (example / '2021-10-05 16:20:10.346826-something.html').touch() - (example / 'random file').touch() - (example / '2021-10-05 16:20:10.346827-something.html').mkdir() - (example / '2021-10-05 16:20:10 invalid date').touch() - - # This single archive is found. - _refresh_archives() - with get_db_session() as session: - self.assertEqual(session.query(Archive).count(), 1) - - # Running the refresh does not result in a new archive. - _refresh_archives() - with get_db_session() as session: - self.assertEqual(session.query(Archive).count(), 1) - def test_archive_refresh_deleted_archive(test_session, archive_directory, archive_factory): """ @@ -191,9 +138,7 @@ def test_archive_refresh_deleted_archive(test_session, archive_directory, archiv def test_refresh_archives_fills_contents(test_session, archive_factory, test_client): - """ - Refreshing archives fills in any missing contents. - """ + """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') archive3 = archive_factory('example.org') @@ -328,3 +273,153 @@ def test_get_title_from_html(test_session, fake_now): with mock.patch('modules.archive.lib.request_archive', fake_request_archive): archive = do_archive('example.com') assert archive.title is None + + +def test_get_new_archive_files(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)) + s, r, t, j, c = map(str, get_new_archive_files('https://example.com/two', None)) + assert str(s).endswith('archive/example.com/2001-01-01-00-00-00_NA.html') + assert str(r).endswith('archive/example.com/2001-01-01-00-00-00_NA-readability.html') + assert str(t).endswith('archive/example.com/2001-01-01-00-00-00_NA-readability.txt') + assert str(j).endswith('archive/example.com/2001-01-01-00-00-00_NA-readability.json') + assert str(c).endswith('archive/example.com/2001-01-01-00-00-00_NA.png') + + s, r, t, j, c = get_new_archive_files('https://www.example.com/one', 'Title') + # Leading www. is removed. + assert str(s).endswith('archive/example.com/2001-01-01-00-00-00_Title.html') + assert str(r).endswith('archive/example.com/2001-01-01-00-00-00_Title-readability.html') + assert str(t).endswith('archive/example.com/2001-01-01-00-00-00_Title-readability.txt') + assert str(j).endswith('archive/example.com/2001-01-01-00-00-00_Title-readability.json') + assert str(c).endswith('archive/example.com/2001-01-01-00-00-00_Title.png') + + +def test_title_in_filename(test_session, fake_now, test_directory): + """ + The Archive files have the title in the path. + """ + fake_now(datetime(2000, 1, 1)) + + with mock.patch('modules.archive.lib.request_archive', make_fake_request_archive()): + archive1 = do_archive('example.com') + + assert archive1.title == 'ジにてこちら' + + assert str(archive1.singlefile_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_ジにてこちら.html' + assert str(archive1.readability_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_ジにてこちら-readability.html' + assert str(archive1.readability_txt_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_ジにてこちら-readability.txt' + assert str(archive1.readability_json_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_ジにてこちら-readability.json' + assert str(archive1.screenshot_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_ジにてこちら.png' + + def fake_request_archive(_): + singlefile = '\ntest single-file\nジにてこちら\n' # no title in HTML + r = dict( + # No Title from Readability. + content=f'test readability content', + textContent='test readability textContent', + ) + s = b'screenshot data' + return singlefile, r, s + + with mock.patch('modules.archive.lib.request_archive', fake_request_archive): + archive2 = do_archive('example.com') + + assert str(archive2.singlefile_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_NA.html' + assert str(archive2.readability_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_NA-readability.html' + assert str(archive2.readability_txt_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_NA-readability.txt' + assert str(archive2.readability_json_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_NA-readability.json' + assert str(archive2.screenshot_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_NA.png' + + def fake_request_archive(_): + singlefile = '\ntest single-file\nジにてこちら\ndangerous ;\\//_title</html></html>' + r = dict( + # No Title from Readability. + content=f'<html>test readability content</html>', + textContent='<html>test readability textContent</html>', + ) + s = b'screenshot data' + return singlefile, r, s + + with mock.patch('modules.archive.lib.request_archive', fake_request_archive): + archive3 = do_archive('example.com') + + assert str(archive3.singlefile_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_dangerous ;_title.html' + assert str(archive3.readability_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_dangerous ;_title-readability.html' + assert str(archive3.readability_txt_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_dangerous ;_title-readability.txt' + assert str(archive3.readability_json_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_dangerous ;_title-readability.json' + assert str(archive3.screenshot_path.path.relative_to(test_directory)) == \ + 'archive/2000-01-01-00-00-00_dangerous ;_title.png' + + +def test_refresh_archives(test_session, test_directory): + """Archives can be found and put in the database. A single Archive will have multiple files.""" + archive_directory = get_archive_directory() + + # Make some test files to refresh. + example = archive_directory / 'example.com' + example.mkdir() + (example / '2021-10-05 16:20:10.346823.html').touch() + (example / '2021-10-05 16:20:10.346823.png').touch() + (example / '2021-10-05 16:20:10.346823-readability.txt').touch() + (example / '2021-10-05 16:20:10.346823-readability.json').touch() + (example / '2021-10-05 16:20:10.346823-readability.html').touch() + with (example / '2021-10-05 16:20:10.346823-readability.json').open('wt') as fh: + json.dump({'url': 'foo'}, fh) + + # These should log an error because they are missing the singlefile path. + (example / '2021-10-05 16:20:10-readability.html').touch() + (example / '2021-10-05 16:20:11-readability.json').touch() + (example / 'random file').touch() + (example / '2021-10-05 16:20:10.346827-something.html').mkdir() + (example / '2021-10-05 16:20:10 no extension').touch() + + # This single archive is found. + _refresh_archives() + with get_db_session() as session: + assert session.query(Archive).count() == 1 + + # Running the refresh does not result in a new archive. + _refresh_archives() + with get_db_session() as session: + assert session.query(Archive).count() == 1 + + # Archives file format was changed, lets check the new formats are found. + (example / '2021-10-05-16-20-10_The Title.html').touch() + (example / '2021-10-05-16-20-10_The Title-readability.json').write_text(json.dumps({'url': 'bar'})) + _refresh_archives() + with get_db_session() as session: + assert session.query(Archive).count() == 2 + + +def test_group_archive_files(test_directory): + files = [ + pathlib.Path('2021-10-05 16:20:10.346823.html'), + pathlib.Path('2021-10-05 16:20:10.346823-readability.json'), + pathlib.Path('2000-01-01-00-00-00_Title.html'), + pathlib.Path('2000-01-01-00-00-00_Title-readability.json'), + ] + assert list(group_archive_files(files)) == [ + (local_timezone(datetime(2000, 1, 1, 0, 0, 0)), + ( + pathlib.Path('2000-01-01-00-00-00_Title.html'), None, None, + pathlib.Path('2000-01-01-00-00-00_Title-readability.json'), None, + )), + (local_timezone(datetime(2021, 10, 5, 16, 20, 10)), + ( + pathlib.Path('2021-10-05 16:20:10.346823.html'), None, None, + pathlib.Path('2021-10-05 16:20:10.346823-readability.json'), None)), + ] diff --git a/wrolpi/common.py b/wrolpi/common.py index 9056e473..320e2796 100644 --- a/wrolpi/common.py +++ b/wrolpi/common.py @@ -646,3 +646,12 @@ def walk(path: Path) -> Generator[Path, None, None]: yield path if path.is_dir(): yield from walk(path) + + +# These characters are invalid in Windows or Linux. +INVALID_FILE_CHARS = re.compile(r'[/<>:\|"\\\?\*]') + + +def escape_file_name(name: str) -> str: + """Replace any invalid characters in a file name with "_".""" + return INVALID_FILE_CHARS.sub('', name) diff --git a/wrolpi/test/test_common.py b/wrolpi/test/test_common.py index afc3455e..5b37a51c 100644 --- a/wrolpi/test/test_common.py +++ b/wrolpi/test/test_common.py @@ -7,7 +7,8 @@ from decimal import Decimal import pytest -from wrolpi.common import combine_dicts, insert_parameter, date_range, api_param_limiter, chdir, zig_zag +from wrolpi.common import combine_dicts, insert_parameter, date_range, api_param_limiter, chdir, zig_zag, \ + escape_file_name from wrolpi.dates import set_timezone, now from wrolpi.errors import InvalidTimezone from wrolpi.test.common import build_test_directories, wrap_media_directory @@ -315,3 +316,23 @@ def test_zig_zag(low, high, expected): result = next(zagger) assert result == i assert low <= result < high + + +@pytest.mark.parametrize( + 'name,expected', [ + ('', ''), + ('foo', 'foo'), + ('foo\\', 'foo'), + ('foo/', 'foo'), + ('foo<', 'foo'), + ('foo>', 'foo'), + ('foo:', 'foo'), + ('foo|', 'foo'), + ('foo"', 'foo'), + ('foo?', 'foo'), + ('foo*', 'foo'), + ('foo&', 'foo&'), + ] +) +def test_escape_file_name(name, expected): + assert escape_file_name(name) == expected