lamindb_setup 1.14.1__py3-none-any.whl → 1.15.0__py3-none-any.whl

lamindb_setup/__init__.py

@@ -35,7 +35,7 @@ Migration management
 
 """
 
-__version__ = "1.14.1"  # denote a release candidate for 0.1.0 with 0.1rc1
+__version__ = "1.15.0"  # denote a release candidate for 0.1.0 with 0.1rc1
 
 import os
 import warnings
@@ -47,7 +47,7 @@ warnings.filterwarnings("ignore", category=DeprecationWarning, module="postgrest
 
 from packaging import version as packaging_version
 
-from . import core, errors, types
+from . import core, errors, io, types
 from ._check_setup import _check_instance_setup
 from ._connect_instance import connect
 from ._delete import delete

lamindb_setup/_connect_instance.py

@@ -204,18 +204,12 @@ def reset_django_module_variables():
     app_names = {app.name for app in apps.get_app_configs()}
     # always copy before iterations over sys.modules
     # see https://docs.python.org/3/library/sys.html#sys.modules
+    # this whole thing runs about 50ms in a big env
     for name, module in sys.modules.copy().items():
         if (
             module is not None
             and (not name.startswith("__") or name == "__main__")
             and name not in sys.builtin_module_names
-            and not (
-                hasattr(module, "__file__")
-                and module.__file__
-                and any(
-                    path in module.__file__ for path in ["/lib/python", "\\lib\\python"]
-                )
-            )
         ):
             try:
                 for k, v in vars(module).items():

lamindb_setup/_init_instance.py

@@ -7,8 +7,6 @@ from typing import TYPE_CHECKING, Literal
 from uuid import UUID
 
 import click
-from django.core.exceptions import FieldError
-from django.db.utils import IntegrityError, OperationalError, ProgrammingError
 from lamin_utils import logger
 
 from ._disconnect import disconnect
@@ -71,6 +69,10 @@ def register_storage_in_instance(ssettings: StorageSettings) -> Storage:
 
 
 def register_user(usettings: UserSettings, update_user: bool = True) -> None:
+    # we have to import this here dynamically because otherwise
+    # the except below will fail on re-connect due to reset
+    from django.core.exceptions import FieldError
+    from django.db.utils import IntegrityError, OperationalError, ProgrammingError
     from lamindb.models import User
 
     if not update_user and User.objects.filter(uid=usettings.uid).exists():

lamindb_setup/_setup_user.py

@@ -7,6 +7,7 @@ from lamin_utils import logger
 
 from ._check_setup import _check_instance_setup
 from ._init_instance import register_user
+from .core._aws_options import reset_aws_options_cache
 from .core._settings import settings
 from .core._settings_load import load_user_settings
 from .core._settings_save import save_user_settings
@@ -147,6 +148,8 @@ def login(
         register_user(user_settings)
 
     settings._user_settings = None
+    # aws s3 credentials are scoped to the user
+    reset_aws_options_cache()
     return user_settings
 
 
@@ -163,6 +166,8 @@ def logout():
     if current_user_settings_file().exists():
         current_user_settings_file().unlink()
         settings._user_settings = None
+        # aws s3 credentials are scoped to the user
+        reset_aws_options_cache()
         logger.success("logged out")
     else:
         logger.important("already logged out")

lamindb_setup/core/__init__.py

@@ -1,8 +1,23 @@
 """Core setup library.
 
+General
+-------
+
 .. autoclass:: SetupSettings
+
+User
+----
+
 .. autoclass:: UserSettings
+
+Instance
+--------
+
 .. autoclass:: InstanceSettings
+
+Storage
+-------
+
 .. autoclass:: StorageSettings
 
 """

lamindb_setup/core/_aws_options.py

@@ -245,3 +245,11 @@ def get_aws_options_manager() -> AWSOptionsManager:
         _aws_options_manager = AWSOptionsManager()
 
     return _aws_options_manager
+
+
+def reset_aws_options_cache():
+    global _aws_options_manager
+
+    if _aws_options_manager is not None:
+        _aws_options_manager._credentials_cache = {}
+        _aws_options_manager._parameters_cache = {}

lamindb_setup/io.py

@@ -0,0 +1,194 @@
+from __future__ import annotations
+
+import json
+import warnings
+from importlib import import_module
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+import pandas as pd
+from django.db import models, transaction
+from rich.progress import Progress
+
+if TYPE_CHECKING:
+    from collections.abc import Sequence
+    from typing import Literal
+
+
+def _get_registries(module_name: str) -> list[str]:
+    """Get registry class names from a module."""
+    schema_module = import_module(module_name)
+    exclude = {"SQLRecord", "BaseSQLRecord"}
+
+    if module_name == "lamindb":
+        module_filter = lambda cls, name: cls.__module__.startswith(
+            f"{module_name}.models."
+        ) and name in dir(schema_module)
+    else:
+        module_filter = (
+            lambda cls, name: cls.__module__ == f"{module_name}.models"
+            and name in dir(schema_module)
+        )
+
+    return [
+        name
+        for name in dir(schema_module.models)
+        if (
+            name[0].isupper()
+            and isinstance(cls := getattr(schema_module.models, name, None), type)
+            and issubclass(cls, models.Model)
+            and module_filter(cls, name)
+            and name not in exclude
+        )
+    ]
+
+
+def _export_registry_to_parquet(registry: type[models.Model], directory: Path) -> None:
+    """Export a single registry table to parquet."""
+    import lamindb_setup as ln_setup
+
+    table_name = registry._meta.db_table
+    with warnings.catch_warnings():
+        warnings.filterwarnings("ignore", message="Skipped unsupported reflection")
+        df = pd.read_sql_table(table_name, ln_setup.settings.instance.db)
+    df.to_parquet(directory / f"{table_name}.parquet", compression=None)
+
+
+def export_db(
+    module_names: Sequence[str] | None = None,
+    *,
+    output_dir: str | Path = "./lamindb_export/",
+) -> None:
+    """Export registry tables and many-to-many link tables to parquet files.
+
+    Ensure that you connect to postgres instances using `use_root_db_user=True`.
+
+    Args:
+        module_names: Module names to export (e.g., ["lamindb", "bionty", "wetlab"]).
+            Defaults to "lamindb" if not provided.
+        output_dir: Directory path for exported parquet files.
+    """
+    directory = Path(output_dir)
+    directory.mkdir(parents=True, exist_ok=True)
+
+    module_names = module_names or ["lamindb"]
+    modules = {name: _get_registries(name) for name in module_names}
+    total_models = sum(len(models) for models in modules.values())
+
+    with Progress() as progress:
+        task = progress.add_task("Exporting", total=total_models)
+        for module_name, model_names in modules.items():
+            schema_module = import_module(module_name)
+            for model_name in model_names:
+                progress.update(task, description=f"[cyan]{module_name}.{model_name}")
+                registry = getattr(schema_module, model_name)
+                _export_registry_to_parquet(registry, directory)
+                for field in registry._meta.many_to_many:
+                    link_orm = getattr(registry, field.name).through
+                    _export_registry_to_parquet(link_orm, directory)
+                progress.advance(task)
+
+
+def _import_registry(
+    registry: type[models.Model],
+    directory: Path,
+    if_exists: Literal["fail", "replace", "append"] = "replace",
+) -> None:
+    """Import a single registry table from parquet.
+
+    Uses raw SQL export instead of django to later circumvent FK constraints.
+    """
+    table_name = registry._meta.db_table
+    parquet_file = directory / f"{table_name}.parquet"
+
+    if not parquet_file.exists():
+        print(f"Skipped {table_name} (file not found)")
+        return
+
+    df = pd.read_parquet(parquet_file)
+
+    old_foreign_key_columns = [col for col in df.columns if col.endswith("_old")]
+    if old_foreign_key_columns:
+        df = df.drop(columns=old_foreign_key_columns)
+
+    for col in df.columns:
+        if df[col].dtype == "object":
+            df[col] = df[col].apply(
+                lambda x: json.dumps(x) if isinstance(x, (dict, list)) else x
+            )
+
+    from django.db import connection
+
+    df.to_sql(table_name, connection.connection, if_exists=if_exists, index=False)
+
+
+def import_db(
+    module_names: Sequence[str] | None = None,
+    *,
+    input_dir: str | Path = "./lamindb_export/",
+    if_exists: Literal["fail", "replace", "append"] = "replace",
+) -> None:
+    """Import registry and link tables from parquet files.
+
+    Temporarily disables FK constraints to allow insertion in arbitrary order.
+    Requires superuser/RDS admin privileges for postgres databases.
+
+    Args:
+        input_dir: Directory containing parquet files to import.
+        module_names: Module names to import (e.g., ["lamindb", "bionty", "wetlab"]).
+        if_exists: How to behave if table exists: 'fail', 'replace', or 'append'.
+    """
+    from django.db import connection
+
+    import lamindb_setup as ln_setup
+
+    directory = Path(input_dir)
+
+    if not directory.exists():
+        raise ValueError(f"Directory does not exist: {directory}")
+
+    if module_names is None:
+        parquet_files = list(directory.glob("*.parquet"))
+        detected_modules = {
+            f.name.split("_")[0] for f in parquet_files if "_" in f.name
+        }
+        module_names = sorted(detected_modules)
+
+    modules = {name: _get_registries(name) for name in module_names}
+    total_models = sum(len(models) for models in modules.values())
+
+    # Disable FK constraints to allow insertion in arbitrary order
+    if ln_setup.settings.instance.dialect == "sqlite":
+        with connection.cursor() as cursor:
+            if ln_setup.settings.instance.dialect == "postgresql":
+                cursor.execute("SET session_replication_role = 'replica'")
+            elif ln_setup.settings.instance.dialect == "sqlite":
+                cursor.execute("PRAGMA foreign_keys = OFF")
+
+    with transaction.atomic():
+        if ln_setup.settings.instance.dialect == "postgresql":
+            with connection.cursor() as cursor:
+                cursor.execute("SET CONSTRAINTS ALL DEFERRED")
+
+        with Progress() as progress:
+            task = progress.add_task("Importing", total=total_models)
+            for module_name, model_names in modules.items():
+                schema_module = import_module(module_name)
+                for model_name in model_names:
+                    progress.update(
+                        task, description=f"[cyan]{module_name}.{model_name}"
+                    )
+                    registry = getattr(schema_module, model_name)
+                    _import_registry(registry, directory, if_exists=if_exists)
+                    for field in registry._meta.many_to_many:
+                        link_orm = getattr(registry, field.name).through
+                        _import_registry(link_orm, directory, if_exists=if_exists)
+                    progress.advance(task)
+
+    # Re-enable FK constraints again
+    if ln_setup.settings.instance.dialect == "sqlite":
+        with connection.cursor() as cursor:
+            if ln_setup.settings.instance.dialect == "postgresql":
+                cursor.execute("SET session_replication_role = 'origin'")
+            elif ln_setup.settings.instance.dialect == "sqlite":
+                cursor.execute("PRAGMA foreign_keys = ON")

{lamindb_setup-1.14.1.dist-info → lamindb_setup-1.15.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: lamindb_setup
-Version: 1.14.1
+Version: 1.15.0
 Summary: Setup & configure LaminDB.
 Author-email: Lamin Labs <open-source@lamin.ai>
 Requires-Python: >=3.10

{lamindb_setup-1.14.1.dist-info → lamindb_setup-1.15.0.dist-info}/RECORD

@@ -1,27 +1,26 @@
-lamindb_setup/__init__.py,sha256=a30otKx_cPd7N6xUqPN0FBI59xCEsc5f_1KuETyUHm8,3211
+lamindb_setup/__init__.py,sha256=jFfiLt0yIvOHoum_g8DmraECRdU-EOd-iSmAKTZyG3c,3215
 lamindb_setup/_cache.py,sha256=pGvDNVHGx4HWr_6w5ajqEJOdysmaGc6F221qFnXkT-k,2747
 lamindb_setup/_check.py,sha256=28PcG8Kp6OpjSLSi1r2boL2Ryeh6xkaCL87HFbjs6GA,129
 lamindb_setup/_check_setup.py,sha256=ToKMxsUq8dQBQh8baOrNVlSb1iC8h4zTg5dV8wMu0W4,6760
-lamindb_setup/_connect_instance.py,sha256=SmDgCmC44SI8A8iwPosRQ2H5edfIPywblnIy12npv-I,18012
+lamindb_setup/_connect_instance.py,sha256=Yp24MMAnypKE14fMzOYCE8oqW0EJTiRHrBuugubfkBM,17816
 lamindb_setup/_delete.py,sha256=KS3r-xGFuDmAbzPUy-9JR-YnPShYdaHjDRQrAmXQ0qM,5863
 lamindb_setup/_disconnect.py,sha256=FT8EpCm5XXDdhDH7QtAnkO3KPatq2HqT9VXGNjgJDbk,1232
 lamindb_setup/_django.py,sha256=uIQflpkp8l3axyPaKURlk3kacgpElVP5KOKmFxYSMGk,1454
 lamindb_setup/_entry_points.py,sha256=sKwXPX9xjOotoAjvgkU5LBwjjHLWVkh0ZGdiSsrch9k,522
-lamindb_setup/_exportdb.py,sha256=QLjoH4dEwqa01A12naKaDPglCCzl2_VLKWFfJRE_uSg,2113
-lamindb_setup/_importdb.py,sha256=fKv9ev5OOj_-bmzC8XZ1GxOcjIjI486yrHSHDWQrJeI,1874
-lamindb_setup/_init_instance.py,sha256=8ejD6zjV0eF7KR-DvnmDAVJb9Ty0hjaPtIkFbyLDvA0,14806
+lamindb_setup/_init_instance.py,sha256=zNXmZPUHYda1CfLGtsvo4gNhHprK9QVPfffUIfBlTxE,14938
 lamindb_setup/_migrate.py,sha256=SN8uphuQX-8XShH5odLyzV8-eyXATDxB5hWoxwxmgBU,11264
 lamindb_setup/_register_instance.py,sha256=RdUZxZWHLdbvdNZWpF8e0UWROb_T0cStWbzc5yUw34I,1047
 lamindb_setup/_schema.py,sha256=b3uzhhWpV5mQtDwhMINc2MabGCnGLESy51ito3yl6Wc,679
 lamindb_setup/_schema_metadata.py,sha256=af1Es7qFKGPRdNmk48384HiB2r-cDTdBPu0wB9qrga4,15526
 lamindb_setup/_set_managed_storage.py,sha256=y5YQASsWNYVWUYeLgh3N2YBETYP7mBtbpxe3X_Vgb5I,2699
-lamindb_setup/_setup_user.py,sha256=DapdzT3u0f5LN5W9W9A6PWw-n8ejcJciQtHN9b5lidA,5889
+lamindb_setup/_setup_user.py,sha256=ojq7UP2Aia8GTCr6m8fylFx9VSuvGu0HmvIJ8RzymE0,6108
 lamindb_setup/_silence_loggers.py,sha256=AKF_YcHvX32eGXdsYK8MJlxEaZ-Uo2f6QDRzjKFCtws,1568
 lamindb_setup/errors.py,sha256=lccF3X3M2mcbHVG_0HxfuJRFFpUE-42paccIxFOfefQ,1958
+lamindb_setup/io.py,sha256=7wU3g2AH1h18wQerltA7p8VP0n47ZOsy32XlRcBbBAc,7321
 lamindb_setup/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 lamindb_setup/types.py,sha256=fuQxZJnrGYe7a_Ju9n1RqO-HhkOAr1l1xjpAg9dmBu8,605
-lamindb_setup/core/__init__.py,sha256=z3Gy27SfnbO1gUMi-fVrL0yeLzUQ4K14vHOpw10hENo,536
-lamindb_setup/core/_aws_options.py,sha256=NtDLfR2BIz3MiR4rGrBu4uW70MFy2p3hjxCnN1sGDB8,9414
+lamindb_setup/core/__init__.py,sha256=gMqbkPeD-T6o_KykonvNiUJJQYz5SCk2yGgFNWFUBNc,603
+lamindb_setup/core/_aws_options.py,sha256=6SAs3dhKxNadQqt7Ce7uHSus0DsU_LcNZoCQWvhgBWA,9626
 lamindb_setup/core/_aws_storage.py,sha256=QEtV-riQrwfivcwqHnXBbkJ-9YyNEXL4fLoCmOHZ1BI,2003
 lamindb_setup/core/_clone.py,sha256=2NlXV04yykqg_k7z59C_kD1F1Hi4H-55H-JtNjhenQ0,3691
 lamindb_setup/core/_deprecated.py,sha256=M3vpM4fZPOncxY2qsXQAPeaEph28xWdv7tYaueaUyAA,2554
@@ -45,7 +44,7 @@ lamindb_setup/core/exceptions.py,sha256=qjMzqy_uzPA7mCOdnoWnS_fdA6OWbdZGftz-YYpl
 lamindb_setup/core/hashing.py,sha256=Y8Uc5uSGTfU6L2R_gb5w8DdHhGRog7RnkK-e9FEMjPY,3680
 lamindb_setup/core/types.py,sha256=T7NwspfRHgIIpYsXDcApks8jkOlGeGRW-YbVLB7jNIo,67
 lamindb_setup/core/upath.py,sha256=bi3k8AYeiGB_NtVTO9e9gHsfs2AFB4fXiVHcbNpnlpI,35780
-lamindb_setup-1.14.1.dist-info/LICENSE,sha256=UOZ1F5fFDe3XXvG4oNnkL1-Ecun7zpHzRxjp-XsMeAo,11324
-lamindb_setup-1.14.1.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
-lamindb_setup-1.14.1.dist-info/METADATA,sha256=hdWYVFwWZWwNHnvRIi0A7v2in1mln0UVbqxr88FlsQA,1798
-lamindb_setup-1.14.1.dist-info/RECORD,,
+lamindb_setup-1.15.0.dist-info/LICENSE,sha256=UOZ1F5fFDe3XXvG4oNnkL1-Ecun7zpHzRxjp-XsMeAo,11324
+lamindb_setup-1.15.0.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+lamindb_setup-1.15.0.dist-info/METADATA,sha256=aVm2bsYK6bWwBDEcp1IJR3kp8xrRnJb5CYK_KT09HCA,1798
+lamindb_setup-1.15.0.dist-info/RECORD,,

lamindb_setup/_exportdb.py

@@ -1,68 +0,0 @@
-from __future__ import annotations
-
-from importlib import import_module
-from pathlib import Path
-
-MODELS = {
-    "lamindb": {
-        "Collection": False,
-        "Artifact": False,
-        "Transform": False,
-        "Run": True,
-        "User": False,
-        "Storage": False,
-        "Feature": False,
-        "Schema": False,
-        "ULabel": False,
-    },
-    # "bionty": {
-    #     "Organism": False,
-    #     "Gene": False,
-    #     "Protein": False,
-    #     "CellMarker": False,
-    #     "Tissue": False,
-    #     "CellType": False,
-    #     "Disease": False,
-    #     "CellLine": False,
-    #     "Phenotype": False,
-    #     "Pathway": False,
-    #     "ExperimentalFactor": False,
-    #     "DevelopmentalStage": False,
-    #     "Ethnicity": False,
-    #     "Source": False,
-    # },
-    # "wetlab": {
-    #     "ExperimentType": False,
-    #     "Experiment": False,
-    #     "Well": False,
-    #     "TreatmentTarget": False,
-    #     "Treatment": False,
-    #     "Biosample": False,
-    #     "Techsample": False,
-    # },
-}
-
-
-def exportdb() -> None:
-    directory = Path("./lamindb_export/")
-    directory.mkdir(parents=True, exist_ok=True)
-    import pandas as pd
-
-    import lamindb_setup as ln_setup
-
-    def export_registry(registry, directory):
-        table_name = registry._meta.db_table
-        df = pd.read_sql_table(table_name, ln_setup.settings.instance.db)
-        df.to_parquet(directory / f"{table_name}.parquet", compression=None)
-
-    # export data to parquet files
-    print(f"\nexporting data to parquet files in: {directory}\n")
-    for module_name, models in MODELS.items():
-        for model_name in models.keys():
-            schema_module = import_module(f"lnschema_{module_name}")
-            registry = getattr(schema_module, model_name)
-            export_registry(registry, directory)
-            many_to_many_names = [field.name for field in registry._meta.many_to_many]
-            for many_to_many_name in many_to_many_names:
-                link_orm = getattr(registry, many_to_many_name).through
-                export_registry(link_orm, directory)

lamindb_setup/_importdb.py

@@ -1,50 +0,0 @@
-from __future__ import annotations
-
-from importlib import import_module
-from pathlib import Path
-
-from ._exportdb import MODELS
-
-
-def import_registry(registry, directory, connection):
-    import pandas as pd
-
-    table_name = registry._meta.db_table
-    df = pd.read_parquet(directory / f"{table_name}.parquet")
-    old_foreign_key_columns = [
-        column for column in df.columns if column.endswith("_old")
-    ]
-    for column in old_foreign_key_columns:
-        df.drop(column, axis=1, inplace=True)
-    df.to_sql(table_name, connection, if_exists="append", index=False)
-
-
-def importdb() -> None:
-    # import data from parquet files
-    directory = Path("./lamindb_export/")
-    if directory.exists():
-        response = input(
-            f"\n\nDo you want to import registries from here: {directory}? (y/n)\n"
-        )
-        if response != "y":
-            return None
-    from sqlalchemy import create_engine, text
-
-    import lamindb_setup as ln_setup
-
-    engine = create_engine(ln_setup.settings.instance.db, echo=False)
-    with engine.begin() as connection:
-        if ln_setup.settings.instance.dialect == "postgresql":
-            connection.execute(text("SET CONSTRAINTS ALL DEFERRED;"))
-        for module_name, models in MODELS.items():
-            for model_name in models.keys():
-                print(model_name)
-                schema_module = import_module(f"lnschema_{module_name}")
-                registry = getattr(schema_module, model_name)
-                import_registry(registry, directory, connection)
-                many_to_many_names = [
-                    field.name for field in registry._meta.many_to_many
-                ]
-                for many_to_many_name in many_to_many_names:
-                    link_orm = getattr(registry, many_to_many_name).through
-                    import_registry(link_orm, directory, connection)