lamindb 0.77.2__py3-none-any.whl → 1.0rc1__py3-none-any.whl

lamindb/_transform.py

@@ -1,26 +1,31 @@
 from __future__ import annotations
 
+import warnings
 from typing import TYPE_CHECKING
 
 from lamin_utils import logger
 from lamindb_setup.core._docs import doc_args
-from lnschema_core.models import Run, Transform
+
+from lamindb.models import Run, Transform
 
 from ._parents import _view_parents
 from ._run import delete_run_artifacts
+from .core._settings import settings
 from .core.exceptions import InconsistentKey
 from .core.versioning import message_update_key_in_version_family, process_revises
 
 if TYPE_CHECKING:
-    from lnschema_core.types import TransformType
+    from lamindb.base.types import TransformType
 
 
 def __init__(transform: Transform, *args, **kwargs):
     if len(args) == len(transform._meta.concrete_fields):
         super(Transform, transform).__init__(*args, **kwargs)
         return None
-    name: str | None = kwargs.pop("name") if "name" in kwargs else None
     key: str | None = kwargs.pop("key") if "key" in kwargs else None
+    description: str | None = (
+        kwargs.pop("description") if "description" in kwargs else None
+    )
     revises: Transform | None = kwargs.pop("revises") if "revises" in kwargs else None
     version: str | None = kwargs.pop("version") if "version" in kwargs else None
     type: TransformType | None = kwargs.pop("type") if "type" in kwargs else "pipeline"
@@ -28,47 +33,77 @@ def __init__(transform: Transform, *args, **kwargs):
     reference_type: str | None = (
         kwargs.pop("reference_type") if "reference_type" in kwargs else None
     )
-    if "is_new_version_of" in kwargs:
-        logger.warning("`is_new_version_of` will be removed soon, please use `revises`")
-        revises = kwargs.pop("is_new_version_of")
+    using_key = (
+        kwargs.pop("using_key") if "using_key" in kwargs else settings._using_key
+    )
+    if "name" in kwargs:
+        if key is None:
+            key = kwargs.pop("name")
+            warnings.warn(
+                f"`name` will be removed soon, please pass '{key}' to `key` instead",
+                FutureWarning,
+                stacklevel=2,
+            )
+        else:
+            # description wasn't exist, so no check necessary
+            description = kwargs.pop("name")
+            warnings.warn(
+                f"`name` will be removed soon, please pass '{description}' to `description` instead",
+                FutureWarning,
+                stacklevel=2,
+            )
     # below is internal use that we'll hopefully be able to eliminate
     uid: str | None = kwargs.pop("uid") if "uid" in kwargs else None
     if not len(kwargs) == 0:
         raise ValueError(
-            "Only name, key, version, type, revises, reference, "
+            "Only key, description, version, type, revises, reference, "
             f"reference_type can be passed, but you passed: {kwargs}"
         )
     if revises is None:
         # need to check uid before checking key
         if uid is not None:
             revises = (
-                Transform.filter(uid__startswith=uid[:-4], is_latest=True)
+                Transform.objects.using(using_key)
+                .filter(uid__startswith=uid[:-4], is_latest=True)
                 .order_by("-created_at")
                 .first()
             )
         elif key is not None:
-            revises = (
-                Transform.filter(key=key, is_latest=True)
+            candidate_for_revises = (
+                Transform.objects.using(using_key)
+                .filter(key=key, is_latest=True)
                 .order_by("-created_at")
                 .first()
             )
+            if candidate_for_revises is not None:
+                revises = candidate_for_revises
+                if candidate_for_revises.source_code is None:
+                    # no source code was yet saved, return the same transform
+                    logger.important(
+                        "no source code was yet saved, returning existing transform with same key"
+                    )
+                    uid = revises.uid
     if revises is not None and uid is not None and uid == revises.uid:
         from ._record import init_self_from_db, update_attributes
 
+        if revises.key != key:
+            logger.warning("ignoring inconsistent key")
         init_self_from_db(transform, revises)
-        update_attributes(transform, {"name": name})
+        update_attributes(transform, {"description": description})
         return None
     if revises is not None and key is not None and revises.key != key:
         note = message_update_key_in_version_family(
             suid=revises.stem_uid,
             existing_key=revises.key,
             new_key=key,
-            registry="Artifact",
+            registry="Transform",
         )
         raise InconsistentKey(
-            f"`key` is {key}, but `revises.key` is '{revises.key}'\n\nEither do *not* pass `key`.\n\n{note}"
+            f"`key` is '{key}', but `revises.key` is '{revises.key}'\n\nEither do *not* pass `key`.\n\n{note}"
         )
-    new_uid, version, name, revises = process_revises(revises, version, name, Transform)
+    new_uid, version, key, description, revises = process_revises(
+        revises, version, key, description, Transform
+    )
     # this is only because the user-facing constructor allows passing a uid
     # most others don't
     if uid is None:
@@ -78,7 +113,7 @@ def __init__(transform: Transform, *args, **kwargs):
         has_consciously_provided_uid = True
     super(Transform, transform).__init__(
         uid=uid,
-        name=name,
+        description=description,
         key=key,
         type=type,
         version=version,
@@ -90,13 +125,6 @@ def __init__(transform: Transform, *args, **kwargs):
 
 
 def delete(self) -> None:
-    _source_code_artifact = None
-    if self._source_code_artifact is not None:
-        _source_code_artifact = self._source_code_artifact
-        self._source_code_artifact = None
-        self.save()
-    if _source_code_artifact is not None:
-        _source_code_artifact.delete(permanent=True)
     # query all runs and delete their artifacts
     runs = Run.filter(transform=self)
     for run in runs:
@@ -116,7 +144,7 @@ def latest_run(self) -> Run:
 def view_lineage(self, with_successors: bool = False, distance: int = 5):
     return _view_parents(
         record=self,
-        field="name",
+        field="key",
         with_children=with_successors,
         distance=distance,
         attr_name="predecessors",

lamindb/_ulabel.py

@@ -1,15 +1,11 @@
 from __future__ import annotations
 
-from typing import TYPE_CHECKING
-
 import lamindb_setup as ln_setup
-from lamindb_setup.core._docs import doc_args
-from lnschema_core import ULabel
+from lamin_utils import logger
 
-from ._utils import attach_func_to_class_method
+from lamindb.models import ULabel
 
-if TYPE_CHECKING:
-    from lnschema_core.types import ListLike
+from ._utils import attach_func_to_class_method
 
 
 def __init__(self, *args, **kwargs):
@@ -19,7 +15,9 @@ def __init__(self, *args, **kwargs):
     # now we proceed with the user-facing constructor
     if len(args) > 0:
         raise ValueError("Only one non-keyword arg allowed")
-    name: str | None = kwargs.pop("name") if "name" in kwargs else None
+    name: str = kwargs.pop("name") if "name" in kwargs else None
+    type: str | None = kwargs.pop("type") if "type" in kwargs else None
+    is_type: str | None = kwargs.pop("is_type") if "is_type" in kwargs else None
     description: str | None = (
         kwargs.pop("description") if "description" in kwargs else None
     )
@@ -31,8 +29,19 @@ def __init__(self, *args, **kwargs):
         raise ValueError(
             "Only name, description, reference, reference_type are valid keyword arguments"
         )
+    if is_type:
+        if name.endswith("s"):
+            logger.warning(
+                "`name` ends with 's', in case you're naming with plural, consider the singular for a type name"
+            )
+        if name[0].islower():
+            logger.warning(
+                "`name` starts with lowercase, in case you're naming a type, consider starting with uppercase"
+            )
     super(ULabel, self).__init__(
         name=name,
+        type=type,
+        is_type=is_type,
         description=description,
         reference=reference,
         reference_type=reference_type,

lamindb/_view.py

@@ -5,13 +5,12 @@ import importlib
 import inspect
 from typing import TYPE_CHECKING
 
-from IPython.display import HTML, display
 from lamin_utils import colors, logger
 from lamindb_setup import settings
 from lamindb_setup._init_instance import get_schema_module_name
-from lnschema_core import Feature, Record
 
 from lamindb.core import FeatureValue, ParamValue
+from lamindb.models import Feature, Record
 
 from ._feature import convert_pandas_dtype_to_lamin_dtype
 
@@ -24,6 +23,8 @@ is_run_from_ipython = getattr(builtins, "__IPYTHON__", False)
 def display_df_with_descriptions(
     df: pd.DataFrame, descriptions: dict[str, str] | None = None
 ):
+    from IPython.display import HTML, display
+
     if descriptions is None:
         display(df)
         return None
@@ -96,7 +97,7 @@ def display_df_with_descriptions(
 def view(
     df: pd.DataFrame | None = None,
     limit: int = 7,
-    schema: str | None = None,
+    modules: str | None = None,
     registries: list[str] | None = None,
 ) -> None:
     """View metadata.
@@ -104,8 +105,8 @@ def view(
     Args:
         df: A DataFrame to display.
         limit: Display the latest `n` records
-        schema: Schema module to view. Default's to
-            `None` and displays all schema modules.
+        modules: schema module to view. Default's to
+            `None` and displays all registry modules.
         registries: List of Record names. Defaults to
             `None` and lists all registries.
 
@@ -127,13 +128,13 @@ def view(
     else:
         show = logger.print
 
-    if schema is not None:
-        schema_names = [schema]
+    if modules is not None:
+        module_names = [modules]
     else:
-        schema_names = ["core"] + list(settings.instance.schema)
+        module_names = ["core"] + list(settings.instance.modules)
 
-    for schema_name in schema_names:
-        schema_module = importlib.import_module(get_schema_module_name(schema_name))
+    for module_name in module_names:
+        schema_module = importlib.import_module(get_schema_module_name(module_name))
         # the below is necessary because a schema module might not have been
         # explicitly accessed
         importlib.reload(schema_module)
@@ -145,7 +146,7 @@ def view(
             and issubclass(registry, Record)
             and registry is not Record
         }
-        if schema_name == "core":
+        if module_name == "core":
             all_registries.update({FeatureValue, ParamValue})
         if registries is not None:
             filtered_registries = {
@@ -155,9 +156,9 @@ def view(
             }
         else:
             filtered_registries = all_registries
-        if len(schema_names) > 1:
-            section = f"* module: {colors.green(colors.bold(schema_name))} *"
-            section_no_color = f"* module: {schema_name} *"
+        if len(module_names) > 1:
+            section = f"* module: {colors.green(colors.bold(module_name))} *"
+            section_no_color = f"* module: {module_name} *"
             logger.print("*" * len(section_no_color))
             logger.print(section)
             logger.print("*" * len(section_no_color))

lamindb/base/__init__.py

@@ -0,0 +1,24 @@
+"""Base library.
+
+Is available also when no instance is connected.
+
+Modules:
+
+.. autosummary::
+   :toctree: .
+
+   types
+
+Utils:
+
+.. autosummary::
+   :toctree: .
+
+   doc_args
+   deprecated
+
+"""
+
+from lamindb_setup.core import deprecated, doc_args
+
+from . import types

lamindb/base/fields.py

@@ -0,0 +1,281 @@
+from django.db import models
+
+
+class CharField(models.CharField):
+    """Custom `CharField` with default values for `blank`, `default`, and `max_length`.
+
+    Django default values for `CharField` are `blank=False`, `default=""`, undefined `max_length`.
+    """
+
+    def __init__(self, max_length: int = 255, **kwargs):
+        kwargs["max_length"] = max_length  # Set max_length in kwargs
+        kwargs.setdefault("blank", True)
+        kwargs.setdefault("default", None)
+        super().__init__(**kwargs)  # Pass all arguments as kwargs
+
+
+class TextField(models.TextField):
+    """Custom `TextField` with default values for `blank` and `default`.
+
+    Django default values for `TextField` are `blank=False`, `default=''`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        kwargs.setdefault("default", None)
+        super().__init__(*args, **kwargs)
+
+
+class ForeignKey(models.ForeignKey):
+    """Custom `ForeignKey` with default values for `blank`.
+
+    Django default value for `ForeignKey` `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class BooleanField(models.BooleanField):
+    """Custom `BooleanField` with default values for `blank` and `default`.
+
+    Django default values for `BooleanField` are `blank=False`, `default=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        kwargs.setdefault("default", None)
+        super().__init__(*args, **kwargs)
+
+
+class DateField(models.DateField):
+    """Custom `DateField` with default values for `blank`.
+
+    Django default values for `DateField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class DateTimeField(models.DateTimeField):
+    """Custom `DateTimeField` with default values for `blank`.
+
+    Django default values for `DateTimeField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class BigIntegerField(models.BigIntegerField):
+    """Custom `BigIntegerField` with default values for `blank`.
+
+    Django default values for `BigIntegerField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        kwargs.setdefault("default", None)
+        super().__init__(*args, **kwargs)
+
+
+class IntegerField(models.IntegerField):
+    """Custom `IntegerField` with default values for `blank`.
+
+    Django default values for `IntegerField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class OneToOneField(models.OneToOneField):
+    """Custom `OneToOneField` with default values for `blank`.
+
+    Django default values for `OneToOneField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class FloatField(models.FloatField):
+    """Custom `FloatField` with default values for `blank`.
+
+    Django default values for `FloatField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class DecimalField(models.DecimalField):
+    """Custom `DecimalField` with default values for `blank`.
+
+    Django default values for `DecimalField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class JSONField(models.JSONField):
+    """Custom `JSONField` with default values for `blank`.
+
+    Django default values for `JSONField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class DurationField(models.DurationField):
+    """Custom `DurationField` with default values for `blank`.
+
+    Django default values for `DurationField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class URLField(models.URLField):
+    """Custom `URLField` with default values for `blank`.
+
+    Django default values for `URLField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class EmailField(models.EmailField):
+    """Custom `EmailField` with default values for `blank`.
+
+    Django default values for `EmailField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class TimeField(models.TimeField):
+    """Custom `TimeField` with default values for `blank`.
+
+    Django default values for `TimeField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class SlugField(models.SlugField):
+    """Custom `SlugField` with default values for `blank`.
+
+    Django default values for `SlugField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class UUIDField(models.UUIDField):
+    """Custom `UUIDField` with default values for `blank`.
+
+    Django default values for `UUIDField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class PositiveIntegerField(models.PositiveIntegerField):
+    """Custom `PositiveIntegerField` with default values for `blank`.
+
+    Django default values for `PositiveIntegerField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class PositiveSmallIntegerField(models.PositiveSmallIntegerField):
+    """Custom `PositiveSmallIntegerField` with default values for `blank`.
+
+    Django default values for `PositiveSmallIntegerField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class SmallIntegerField(models.SmallIntegerField):
+    """Custom `SmallIntegerField` with default values for `blank`.
+
+    Django default values for `SmallIntegerField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class BinaryField(models.BinaryField):
+    """Custom `BinaryField` with default values for `blank`.
+
+    Django default values for `BinaryField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class GenericIPAddressField(models.GenericIPAddressField):
+    """Custom `GenericIPAddressField` with default values for `blank`.
+
+    Django default values for `GenericIPAddressField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class FileField(models.FileField):
+    """Custom `FileField` with default values for `blank`.
+
+    Django default values for `FileField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)
+
+
+class ImageField(models.ImageField):
+    """Custom `ImageField` with default values for `blank`.
+
+    Django default values for `ImageField` are `blank=False`.
+    """
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault("blank", True)
+        super().__init__(*args, **kwargs)

lamindb/base/ids.py

@@ -0,0 +1,103 @@
+"""IDs.
+
+Base generators:
+
+.. autosummary::
+   :toctree: .
+
+   base26
+   base62
+   base64
+
+8 base62 characters:
+
+======= ===========
+n       p_collision
+======= ===========
+100k    2e-05
+1M      2e-03
+======= ===========
+
+12 base62 characters:
+
+======= ===========
+n       p_collision
+======= ===========
+100M    2e-06
+1B      2e-04
+======= ===========
+
+20 base62 characters (62**20=7e+35) roughly matches UUID (2*122=5e+36):
+
+======= ===========
+n       p_collision
+======= ===========
+3e15    1e-6
+======= ===========
+
+"""
+
+import secrets
+import string
+
+
+def base64(n_char: int) -> str:
+    """Random Base64 string."""
+    alphabet = string.digits + string.ascii_letters.swapcase() + "_" + "-"
+    id = "".join(secrets.choice(alphabet) for i in range(n_char))
+    return id
+
+
+def base62(n_char: int) -> str:
+    """Random Base62 string."""
+    alphabet = string.digits + string.ascii_letters.swapcase()
+    id = "".join(secrets.choice(alphabet) for i in range(n_char))
+    return id
+
+
+# the following cannot be serialized by Django
+# class Base62:
+#     def __init__(self, n_char: int):
+#         self.n_char = n_char
+
+#     def __call__(self):
+#         return base62(self.n_char)
+
+
+def base26(n_char: int):
+    """ASCII lowercase."""
+    alphabet = string.ascii_lowercase
+    id = "".join(secrets.choice(alphabet) for i in range(n_char))
+    return id
+
+
+def base62_4() -> str:
+    return base62(4)
+
+
+def base62_8() -> str:
+    return base62(8)
+
+
+def base62_12() -> str:
+    return base62(12)
+
+
+def base62_14() -> str:
+    return base62(14)
+
+
+def base62_16() -> str:
+    return base62(16)
+
+
+def base62_18() -> str:
+    return base62(18)
+
+
+def base62_20() -> str:
+    return base62(20)
+
+
+def base62_24() -> str:
+    return base62(24)