def test_hashing(read_asset, config_injector):
config_injector({
'elasticsearch': {
'host': 'localhost',
'port': 9200,
'index': 'szurubooru_test',
},
})
image_hash.purge()
image_hash.add_image('test', read_asset('jpeg.jpg'))
paths = image_hash.get_all_paths()
results_exact = image_hash.search_by_image(read_asset('jpeg.jpg'))
results_similar = image_hash.search_by_image(
read_asset('jpeg-similar.jpg'))
assert len(paths) == 1
assert len(results_exact) == 1
assert len(results_similar) == 1
assert results_exact[0].path == 'test'
assert results_exact[0].score == 63
assert results_exact[0].distance == 0
assert results_similar[0].path == 'test'
assert results_similar[0].score == 17
assert abs(results_similar[0].distance - 0.20599895341812172) < 1e-8
def _sync_post_content(post: model.Post) -> None:
regenerate_thumb = False
if hasattr(post, '__content'):
content = getattr(post, '__content')
files.save(get_post_content_path(post), content)
delattr(post, '__content')
regenerate_thumb = True
if post.post_id and post.type in (
model.Post.TYPE_IMAGE, model.Post.TYPE_ANIMATION):
image_hash.delete_image(post.post_id)
image_hash.add_image(post.post_id, content)
if hasattr(post, '__thumbnail'):
if getattr(post, '__thumbnail'):
files.save(
get_post_thumbnail_backup_path(post),
getattr(post, '__thumbnail'))
else:
files.delete(get_post_thumbnail_backup_path(post))
delattr(post, '__thumbnail')
regenerate_thumb = True
if regenerate_thumb:
generate_post_thumbnail(post)
def _sync_post_content(post: model.Post) -> None:
regenerate_thumb = False
if hasattr(post, '__content'):
content = getattr(post, '__content')
files.save(get_post_content_path(post), content)
delattr(post, '__content')
regenerate_thumb = True
if post.post_id and post.type in (
model.Post.TYPE_IMAGE, model.Post.TYPE_ANIMATION):
image_hash.delete_image(post.post_id)
image_hash.add_image(post.post_id, content)
if hasattr(post, '__thumbnail'):
if getattr(post, '__thumbnail'):
files.save(
get_post_thumbnail_backup_path(post),
getattr(post, '__thumbnail'))
else:
files.delete(get_post_thumbnail_backup_path(post))
delattr(post, '__thumbnail')
regenerate_thumb = True
if regenerate_thumb:
generate_post_thumbnail(post)
def test_hashing(read_asset, config_injector):
config_injector({
'elasticsearch': {
'host': 'localhost',
'port': 9200,
'index': 'szurubooru_test',
'user': '******',
'pass': None,
},
})
if not image_hash.get_session().ping():
pytest.xfail('Unable to connect to ElasticSearch, '
'perhaps it is not available for this test?')
image_hash.purge()
image_hash.add_image('test', read_asset('jpeg.jpg'))
paths = image_hash.get_all_paths()
results_exact = image_hash.search_by_image(read_asset('jpeg.jpg'))
results_similar = image_hash.search_by_image(
read_asset('jpeg-similar.jpg'))
assert len(paths) == 1
assert len(results_exact) == 1
assert len(results_similar) == 1
assert results_exact[0].path == 'test'
assert results_exact[0].score == 63
assert results_exact[0].distance == 0
assert results_similar[0].path == 'test'
assert results_similar[0].score == 17
assert abs(results_similar[0].distance - 0.20599895341812172) < 1e-8
def populate_reverse_search() -> None:
excluded_post_ids = image_hash.get_all_paths()
post_ids_to_hash = (db.session.query(model.Post.post_id).filter(
(model.Post.type == model.Post.TYPE_IMAGE)
| (model.Post.type == model.Post.TYPE_ANIMATION)).filter(
~model.Post.post_id.in_(excluded_post_ids)).order_by(
model.Post.post_id.asc()).all())
for post_ids_chunk in util.chunks(post_ids_to_hash, 100):
posts_chunk = (db.session.query(model.Post).filter(
model.Post.post_id.in_(post_ids_chunk)).all())
for post in posts_chunk:
content_path = get_post_content_path(post)
if files.has(content_path):
image_hash.add_image(post.post_id, files.get(content_path))
def populate_reverse_search():
excluded_post_ids = image_hash.get_all_paths()
post_ids_to_hash = (db.session
.query(db.Post.post_id)
.filter(
(db.Post.type == db.Post.TYPE_IMAGE) |
(db.Post.type == db.Post.TYPE_ANIMATION))
.filter(~db.Post.post_id.in_(excluded_post_ids))
.order_by(db.Post.post_id.asc())
.all())
for post_ids_chunk in util.chunks(post_ids_to_hash, 100):
posts_chunk = (db.session
.query(db.Post)
.filter(db.Post.post_id.in_(post_ids_chunk))
.all())
for post in posts_chunk:
content_path = get_post_content_path(post)
if files.has(content_path):
image_hash.add_image(post.post_id, files.get(content_path))