Add, replace, cache and delete artifacts¶

import pytest
import lamindb as ln

ln.setup.login("testuser1")
ln.setup.init(storage="s3://lamindb-ci/test-add-replace-cache")

✓ logged in with email testuser1@lamin.ai (uid: DzTjkKse)

→ go to: https://lamin.ai/testuser1/test-add-replace-cache

! updating cloud SQLite 's3://lamindb-ci/test-add-replace-cache/23d258b416df505da3a53878e9797d26.lndb' of instance 'testuser1/test-add-replace-cache'

! locked instance (to unlock and push changes to the cloud SQLite file, call: lamin disconnect)

Save with auto-managed (`key=None`)¶

AUTO_KEY_PREFIX = ln.core.storage.paths.AUTO_KEY_PREFIX
root = ln.settings.storage.root

artifact = ln.Artifact("./test-files/iris.csv", description="iris.csv")

! no run & transform got linked, call `ln.track()` & re-run

artifact.save()

Artifact(uid='goOwpQtkOsMzqdf10000', is_latest=True, description='iris.csv', suffix='.csv', size=224, hash='iwc1TmF1TW_l5weDvscSHw', _hash_type='md5', visibility=1, _key_is_virtual=True, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:30 UTC)

key_path = root / f"{AUTO_KEY_PREFIX}{artifact.uid}{artifact.suffix}"
assert key_path.exists()

cache_csv_path = artifact.cache()
print(cache_csv_path)
assert cache_csv_path.suffix == ".csv"

! run input wasn't tracked, call `ln.track()` and re-run

/home/runner/.cache/lamindb/lamindb-ci/test-add-replace-cache/.lamindb/goOwpQtkOsMzqdf10000.csv

artifact.replace("./test-files/iris.data")
artifact.save()

! no run & transform got linked, call `ln.track()` & re-run

Artifact(uid='goOwpQtkOsMzqdf10000', is_latest=True, description='iris.csv', suffix='.data', size=182, hash='42Br6no9CjB6s5ZbmO-bmw', _hash_type='md5', visibility=1, _key_is_virtual=True, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:30 UTC)

old_key_path = key_path
new_key_path = root / f"{AUTO_KEY_PREFIX}{artifact.uid}{artifact.suffix}"

The suffix changed:

print(old_key_path)
print(new_key_path)
assert not old_key_path.exists()
assert new_key_path.exists()

s3://lamindb-ci/test-add-replace-cache/.lamindb/goOwpQtkOsMzqdf10000.csv
s3://lamindb-ci/test-add-replace-cache/.lamindb/goOwpQtkOsMzqdf10000.data

cache_data_path = artifact.cache()
print(cache_data_path)
assert cache_data_path.suffix == ".data"
assert cache_data_path.stat().st_mtime >= cache_csv_path.stat().st_mtime

! run input wasn't tracked, call `ln.track()` and re-run

/home/runner/.cache/lamindb/lamindb-ci/test-add-replace-cache/.lamindb/goOwpQtkOsMzqdf10000.data

artifact.delete(permanent=True)

Save with manually passed real `key`¶

ln.settings.creation._artifact_use_virtual_keys = False

artifact = ln.Artifact("./test-files/iris.csv", key="iris.csv")

! no run & transform got linked, call `ln.track()` & re-run

artifact.save()

Artifact(uid='pRHhbMF21kJIPcps0000', is_latest=True, key='iris.csv', suffix='.csv', size=224, hash='iwc1TmF1TW_l5weDvscSHw', _hash_type='md5', visibility=1, _key_is_virtual=False, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:30 UTC)

key_path = root / "iris.csv"

assert key_path.exists()

artifact.replace("./test-files/new_iris.csv")

! no run & transform got linked, call `ln.track()` & re-run

artifact.save()

Artifact(uid='pRHhbMF21kJIPcps0000', is_latest=True, key='iris.csv', suffix='.csv', size=229, hash='lp2-ycXcKcaliUTnR_TqHA', _hash_type='md5', visibility=1, _key_is_virtual=False, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:30 UTC)

Check paths: no changes here, as the suffix didn’t change.

old_key_path = key_path
new_key_path = root / "new_iris.csv"

old_key_path

S3Path('s3://lamindb-ci/test-add-replace-cache/iris.csv')

new_key_path

S3Path('s3://lamindb-ci/test-add-replace-cache/new_iris.csv')

assert old_key_path.exists()
assert not new_key_path.exists()

artifact.replace("./test-files/iris.data")

! no run & transform got linked, call `ln.track()` & re-run

! replacing the file will replace key 'iris.csv' with 'iris.data' and delete 'iris.csv' upon `save()`

artifact.save()

Artifact(uid='pRHhbMF21kJIPcps0000', is_latest=True, key='iris.data', suffix='.data', size=182, hash='42Br6no9CjB6s5ZbmO-bmw', _hash_type='md5', visibility=1, _key_is_virtual=False, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:30 UTC)

new_key_path = root / "iris.data"

old_key_path

S3Path('s3://lamindb-ci/test-add-replace-cache/iris.csv')

new_key_path

S3Path('s3://lamindb-ci/test-add-replace-cache/iris.data')

assert not old_key_path.exists()
assert new_key_path.exists()

artifact.delete(permanent=True, storage=True)

Save from memory¶

import pandas as pd

iris = pd.read_csv("./test-files/iris.csv")

artifact = ln.Artifact.from_df(iris, description="iris_store", key="iris.parquet")

! no run & transform got linked, call `ln.track()` & re-run

artifact.save()

Artifact(uid='nlHBhG4b4vnyCJFD0000', is_latest=True, description='iris_store', key='iris.parquet', suffix='.parquet', type='dataset', size=4024, hash='SiBYfdBZ0X7v6j8jmk_spw', _hash_type='md5', _accessor='DataFrame', visibility=1, _key_is_virtual=False, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:31 UTC)

key_path = root / "iris.parquet"

assert key_path.exists()

artifact.replace(data=iris[:-1])

! no run & transform got linked, call `ln.track()` & re-run

assert artifact.key == "iris.parquet"

artifact.save()

Artifact(uid='nlHBhG4b4vnyCJFD0000', is_latest=True, description='iris_store', key='iris.parquet', suffix='.parquet', type='dataset', size=4004, hash='TsfJvZq5WiSYmM13BHugBg', _hash_type='md5', _accessor='DataFrame', visibility=1, _key_is_virtual=False, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:31 UTC)

assert key_path.exists()

artifact.replace("./test-files/new_iris.csv")

! no run & transform got linked, call `ln.track()` & re-run

! replacing the file will replace key 'iris.parquet' with 'iris.csv' and delete 'iris.parquet' upon `save()`

artifact.save()

Artifact(uid='nlHBhG4b4vnyCJFD0000', is_latest=True, description='iris_store', key='iris.csv', suffix='.csv', type='dataset', size=229, hash='lp2-ycXcKcaliUTnR_TqHA', _hash_type='md5', _accessor='DataFrame', visibility=1, _key_is_virtual=False, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:31 UTC)

old_key_path = key_path
new_key_path = root / "iris.csv"

old_key_path

S3Path('s3://lamindb-ci/test-add-replace-cache/iris.parquet')

new_key_path

S3Path('s3://lamindb-ci/test-add-replace-cache/iris.csv')

assert not old_key_path.exists()
assert new_key_path.exists()

# we use the path in the next section
path_in_storage = artifact.path
artifact.delete(permanent=True, storage=False)

→ a file/folder remains here: s3://lamindb-ci/test-add-replace-cache/iris.csv

Save with manually passed virtual `key`¶

ln.settings.creation._artifact_use_virtual_keys = True

artifact = ln.Artifact("./test-files/iris.csv", key="iris.csv")

! no run & transform got linked, call `ln.track()` & re-run

artifact.save()

Artifact(uid='wcar9n9GvuTvzHlB0000', is_latest=True, key='iris.csv', suffix='.csv', size=224, hash='iwc1TmF1TW_l5weDvscSHw', _hash_type='md5', visibility=1, _key_is_virtual=True, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:31 UTC)

with pytest.raises(ValueError):
    artifact.replace(path_in_storage)

! no run & transform got linked, call `ln.track()` & re-run

# return an existing artifact if the hash is the same
assert artifact == artifact.replace("./test-files/iris.csv")

! no run & transform got linked, call `ln.track()` & re-run

→ returning existing artifact with same hash: Artifact(uid='wcar9n9GvuTvzHlB0000', is_latest=True, key='iris.csv', suffix='.csv', size=224, hash='iwc1TmF1TW_l5weDvscSHw', _hash_type='md5', visibility=1, _key_is_virtual=True, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:31 UTC)

fpath = artifact.path
assert fpath.suffix == ".csv" and fpath.stem == artifact.uid

artifact.replace("./test-files/iris.data")

! no run & transform got linked, call `ln.track()` & re-run

artifact.save()

Artifact(uid='wcar9n9GvuTvzHlB0000', is_latest=True, key='iris.data', suffix='.data', size=182, hash='42Br6no9CjB6s5ZbmO-bmw', _hash_type='md5', visibility=1, _key_is_virtual=True, storage_id=1, created_by_id=1, created_at=2024-11-13 13:00:31 UTC)

assert artifact.key == "iris.data"

assert not fpath.exists()

fpath = artifact.path
assert fpath.suffix == ".data" and fpath.stem == artifact.uid

artifact.delete(permanent=True, storage=True)

path_in_storage.unlink()

ln.setup.delete("test-add-replace-cache", force=True)

→ deleted storage record on hub 137768267d4556a491fc02b58f1b630b

→ deleted instance record on hub 23d258b416df505da3a53878e9797d26

previous

Track artifacts, in-memory objects & folders [S3 storage]

next

Test AnnDataAccessor