datachain 0.24.0__py3-none-any.whl → 0.24.2__py3-none-any.whl

datachain/data_storage/warehouse.py

@@ -218,7 +218,7 @@ class AbstractWarehouse(ABC, Serializable):
         limit = query._limit
         paginated_query = query.limit(page_size)
 
-        offset = 0
+        offset = query._offset or 0
         num_yielded = 0
 
         # Ensure we're using a thread-local connection
@@ -234,13 +234,13 @@ class AbstractWarehouse(ABC, Serializable):
                 # Cursor results are not thread-safe, so we convert them to a list
                 results = list(wh.dataset_rows_select(paginated_query.offset(offset)))
 
-                processed = False
+                processed = 0
                 for row in results:
-                    processed = True
+                    processed += 1
                     yield row
                     num_yielded += 1
 
-                if not processed:
+                if processed < page_size:
                     break  # no more results
                 offset += page_size
 
@@ -343,6 +343,8 @@ class AbstractWarehouse(ABC, Serializable):
         if (id_col := get_query_id_column(query)) is None:
             raise RuntimeError("sys__id column not found in query")
 
+        query = query._clone().offset(None).limit(None).order_by(None)
+
         if is_batched:
             for batch in ids:
                 yield list(self.dataset_rows_select(query.where(id_col.in_(batch))))

datachain/delta.py

@@ -6,6 +6,7 @@ from typing import TYPE_CHECKING, Callable, Optional, TypeVar, Union
 import datachain
 from datachain.dataset import DatasetDependency
 from datachain.error import DatasetNotFoundError
+from datachain.project import Project
 
 if TYPE_CHECKING:
     from typing_extensions import Concatenate, ParamSpec
@@ -50,15 +51,24 @@ def _append_steps(dc: "DataChain", other: "DataChain"):
 
 def _get_delta_chain(
     source_ds_name: str,
+    source_ds_project: Project,
     source_ds_version: str,
     source_ds_latest_version: str,
     on: Union[str, Sequence[str]],
     compare: Optional[Union[str, Sequence[str]]] = None,
 ) -> "DataChain":
     """Get delta chain for processing changes between versions."""
-    source_dc = datachain.read_dataset(source_ds_name, version=source_ds_version)
+    source_dc = datachain.read_dataset(
+        source_ds_name,
+        namespace=source_ds_project.namespace.name,
+        project=source_ds_project.name,
+        version=source_ds_version,
+    )
     source_dc_latest = datachain.read_dataset(
-        source_ds_name, version=source_ds_latest_version
+        source_ds_name,
+        namespace=source_ds_project.namespace.name,
+        project=source_ds_project.name,
+        version=source_ds_latest_version,
     )
 
     # Calculate diff between source versions
@@ -67,12 +77,15 @@ def _get_delta_chain(
 
 def _get_retry_chain(
     name: str,
+    project: Project,
     latest_version: str,
     source_ds_name: str,
-    source_ds_latest_version: str,
+    source_ds_project: Project,
+    source_ds_version: str,
     on: Union[str, Sequence[str]],
     right_on: Optional[Union[str, Sequence[str]]],
     delta_retry: Optional[Union[bool, str]],
+    diff_chain: "DataChain",
 ) -> Optional["DataChain"]:
     """Get retry chain for processing error records and missing records."""
     # Import here to avoid circular import
@@ -81,35 +94,49 @@ def _get_retry_chain(
     retry_chain = None
 
     # Read the latest version of the result dataset for retry logic
-    result_dataset = datachain.read_dataset(name, version=latest_version)
-    source_dc_latest = datachain.read_dataset(
-        source_ds_name, version=source_ds_latest_version
+    result_dataset = datachain.read_dataset(
+        name,
+        namespace=project.namespace.name,
+        project=project.name,
+        version=latest_version,
+    )
+    source_dc = datachain.read_dataset(
+        source_ds_name,
+        namespace=source_ds_project.namespace.name,
+        project=source_ds_project.name,
+        version=source_ds_version,
     )
 
     # Handle error records if delta_retry is a string (column name)
     if isinstance(delta_retry, str):
         error_records = result_dataset.filter(C(delta_retry) != "")
-        error_source_records = source_dc_latest.merge(
+        error_source_records = source_dc.merge(
             error_records, on=on, right_on=right_on, inner=True
-        ).select(*list(source_dc_latest.signals_schema.values))
+        ).select(*list(source_dc.signals_schema.values))
         retry_chain = error_source_records
 
     # Handle missing records if delta_retry is True
     elif delta_retry is True:
-        missing_records = source_dc_latest.subtract(
-            result_dataset, on=on, right_on=right_on
-        )
+        missing_records = source_dc.subtract(result_dataset, on=on, right_on=right_on)
         retry_chain = missing_records
 
-    return retry_chain
+    # Subtract also diff chain since some items might be picked
+    # up by `delta=True` itself (e.g. records got modified AND are missing in the
+    # result dataset atm)
+    return retry_chain.subtract(diff_chain, on=on) if retry_chain else None
 
 
 def _get_source_info(
     name: str,
+    project: Project,
     latest_version: str,
     catalog,
 ) -> tuple[
-    Optional[str], Optional[str], Optional[str], Optional[list[DatasetDependency]]
+    Optional[str],
+    Optional[Project],
+    Optional[str],
+    Optional[str],
+    Optional[list[DatasetDependency]],
 ]:
     """Get source dataset information and dependencies.
 
@@ -118,23 +145,34 @@ def _get_source_info(
         Returns (None, None, None, None) if source dataset was removed.
     """
     dependencies = catalog.get_dataset_dependencies(
-        name, latest_version, indirect=False
+        name, latest_version, project=project, indirect=False
     )
 
     dep = dependencies[0]
     if not dep:
         # Starting dataset was removed, back off to normal dataset creation
-        return None, None, None, None
+        return None, None, None, None, None
 
+    source_ds_project = catalog.metastore.get_project(dep.project, dep.namespace)
     source_ds_name = dep.name
     source_ds_version = dep.version
-    source_ds_latest_version = catalog.get_dataset(source_ds_name).latest_version
-
-    return source_ds_name, source_ds_version, source_ds_latest_version, dependencies
+    source_ds_latest_version = catalog.get_dataset(
+        source_ds_name, project=source_ds_project
+    ).latest_version
+
+    return (
+        source_ds_name,
+        source_ds_project,
+        source_ds_version,
+        source_ds_latest_version,
+        dependencies,
+    )
 
 
 def delta_retry_update(
     dc: "DataChain",
+    namespace_name: str,
+    project_name: str,
     name: str,
     on: Union[str, Sequence[str]],
     right_on: Optional[Union[str, Sequence[str]]] = None,
@@ -173,11 +211,12 @@ def delta_retry_update(
     """
 
     catalog = dc.session.catalog
+    project = catalog.metastore.get_project(project_name, namespace_name)
     dc._query.apply_listing_pre_step()
 
     # Check if dataset exists
     try:
-        dataset = catalog.get_dataset(name)
+        dataset = catalog.get_dataset(name, project=project)
         latest_version = dataset.latest_version
     except DatasetNotFoundError:
         # First creation of result dataset
@@ -189,19 +228,29 @@ def delta_retry_update(
     retry_chain = None
     processing_chain = None
 
-    source_ds_name, source_ds_version, source_ds_latest_version, dependencies = (
-        _get_source_info(name, latest_version, catalog)
-    )
+    (
+        source_ds_name,
+        source_ds_project,
+        source_ds_version,
+        source_ds_latest_version,
+        dependencies,
+    ) = _get_source_info(name, project, latest_version, catalog)
 
     # If source_ds_name is None, starting dataset was removed
     if source_ds_name is None:
         return None, None, True
 
+    assert source_ds_project
     assert source_ds_version
     assert source_ds_latest_version
 
     diff_chain = _get_delta_chain(
-        source_ds_name, source_ds_version, source_ds_latest_version, on, compare
+        source_ds_name,
+        source_ds_project,
+        source_ds_version,
+        source_ds_latest_version,
+        on,
+        compare,
     )
 
     # Filter out removed dep
@@ -215,12 +264,15 @@ def delta_retry_update(
     if delta_retry:
         retry_chain = _get_retry_chain(
             name,
+            project,
             latest_version,
             source_ds_name,
-            source_ds_latest_version,
+            source_ds_project,
+            source_ds_version,
             on,
             right_on,
             delta_retry,
+            diff_chain,
         )
 
     # Combine delta and retry chains
@@ -236,7 +288,12 @@ def delta_retry_update(
     if processing_chain is None or (processing_chain and processing_chain.empty):
         return None, None, False
 
-    latest_dataset = datachain.read_dataset(name, version=latest_version)
+    latest_dataset = datachain.read_dataset(
+        name,
+        namespace=project.namespace.name,
+        project=project.name,
+        version=latest_version,
+    )
     compared_chain = latest_dataset.diff(
         processing_chain,
         on=right_on or on,

datachain/lib/dc/datachain.py

@@ -598,6 +598,8 @@ class DataChain:
 
             result_ds, dependencies, has_changes = delta_retry_update(
                 self,
+                namespace_name,
+                project_name,
                 name,
                 on=self._delta_on,
                 right_on=self._delta_result_on,

datachain/query/dataset.py

@@ -11,6 +11,7 @@ from collections.abc import Generator, Iterable, Iterator, Sequence
 from copy import copy
 from functools import wraps
 from secrets import token_hex
+from types import GeneratorType
 from typing import (
     TYPE_CHECKING,
     Any,
@@ -557,8 +558,8 @@ class UDFStep(Step, ABC):
         """
         assert self.partition_by is not None
 
-        if isinstance(self.partition_by, Sequence):
-            list_partition_by = self.partition_by
+        if isinstance(self.partition_by, (list, tuple, GeneratorType)):
+            list_partition_by = list(self.partition_by)
         else:
             list_partition_by = [self.partition_by]
 
@@ -575,7 +576,10 @@ class UDFStep(Step, ABC):
             f.dense_rank().over(order_by=partition_by).label(PARTITION_COLUMN_ID),
         ]
         self.catalog.warehouse.db.execute(
-            tbl.insert().from_select(cols, query.with_only_columns(*cols))
+            tbl.insert().from_select(
+                cols,
+                query.offset(None).limit(None).with_only_columns(*cols),
+            )
         )
 
         return tbl
@@ -601,13 +605,10 @@ class UDFStep(Step, ABC):
         if self.partition_by is not None:
             partition_tbl = self.create_partitions_table(query)
             temp_tables.append(partition_tbl.name)
-
-            subq = query.subquery()
-            query = (
-                sqlalchemy.select(*subq.c)
-                .outerjoin(partition_tbl, partition_tbl.c.sys__id == subq.c.sys__id)
-                .add_columns(*partition_columns())
-            )
+            query = query.outerjoin(
+                partition_tbl,
+                partition_tbl.c.sys__id == query.selected_columns.sys__id,
+            ).add_columns(*partition_columns())
 
         query, tables = self.process_input_query(query)
         temp_tables.extend(t.name for t in tables)

{datachain-0.24.0.dist-info → datachain-0.24.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: datachain
-Version: 0.24.0
+Version: 0.24.2
 Summary: Wrangle unstructured AI data at scale
 Author-email: Dmitry Petrov <support@dvc.org>
 License-Expression: Apache-2.0

{datachain-0.24.0.dist-info → datachain-0.24.2.dist-info}/RECORD

@@ -4,7 +4,7 @@ datachain/asyn.py,sha256=RH_jFwJcTXxhEFomaI9yL6S3Onau6NZ6FSKfKFGtrJE,9689
 datachain/cache.py,sha256=ESVRaCJXEThMIfGEFVHx6wJPOZA7FYk9V6WxjyuqUBY,3626
 datachain/config.py,sha256=g8qbNV0vW2VEKpX-dGZ9pAn0DAz6G2ZFcr7SAV3PoSM,4272
 datachain/dataset.py,sha256=wDrukmkDnYP0X8bAGY-7O1NDE3DWCFqrH8VVDpXM9Ok,25263
-datachain/delta.py,sha256=4RqLLc9dJLF8x9GG9IDgi86DwuPerZQ4HAUnNBeACw8,8446
+datachain/delta.py,sha256=fTEhCedseUsHuH_Ek52NXFhFPyFD_6MioEH5sCilNgo,9897
 datachain/error.py,sha256=OWwWMkzZYJrkcoEDGhJHMf7SfKvxcsOLRF94mjPf29I,1609
 datachain/job.py,sha256=x5PB6d5sqx00hePNNkirESlOVAvnmkEM5ygUgQmAhsk,1262
 datachain/listing.py,sha256=T4bCgdCRuFW7bsPUG2PSl5om2nfJL6fzB84m7mCO8cA,7136
@@ -53,7 +53,7 @@ datachain/data_storage/metastore.py,sha256=9mWYOKK3AoHeKPGFm-WBfPrmnYHhwYeXx5MOu
 datachain/data_storage/schema.py,sha256=o3JbURKXRg3IJyIVA4QjHHkn6byRuz7avbydU2FlvNY,9897
 datachain/data_storage/serializer.py,sha256=6G2YtOFqqDzJf1KbvZraKGXl2XHZyVml2krunWUum5o,927
 datachain/data_storage/sqlite.py,sha256=tT_soVi6l_pFSKaDktA1t4qW_vmPvXnvYSf4TZTKZYk,30067
-datachain/data_storage/warehouse.py,sha256=_7btARw-kd-Nx19S0qW6JqdF3VYyypQXFzsXq68SWKI,32327
+datachain/data_storage/warehouse.py,sha256=2Bp2fXfcm-acwYjDWqVzGjoIQSAR4L56GPNtPcaT2gU,32418
 datachain/diff/__init__.py,sha256=-OFZzgOplqO84iWgGY7kfe60NXaWR9JRIh9T-uJboAM,9668
 datachain/fs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datachain/fs/reference.py,sha256=A8McpXF0CqbXPqanXuvpKu50YLB3a2ZXA3YAPxtBXSM,914
@@ -103,7 +103,7 @@ datachain/lib/convert/values_to_tuples.py,sha256=j5yZMrVUH6W7b-7yUvdCTGI7JCUAYUO
 datachain/lib/dc/__init__.py,sha256=HD0NYrdy44u6kkpvgGjJcvGz-UGTHui2azghcT8ZUg0,838
 datachain/lib/dc/csv.py,sha256=q6a9BpapGwP6nwy6c5cklxQumep2fUp9l2LAjtTJr6s,4411
 datachain/lib/dc/database.py,sha256=g5M6NjYR1T0vKte-abV-3Ejnm-HqxTIMir5cRi_SziE,6051
-datachain/lib/dc/datachain.py,sha256=dFI7JX5-41HLgA-TUR99dtR1lvk2vokaMC3mbIW1XT4,85814
+datachain/lib/dc/datachain.py,sha256=dOPtNOYx6ocFr61YHTDrKGoMTDWDY0AZt8MLh79EJkc,85876
 datachain/lib/dc/datasets.py,sha256=U4xqAfs6FdW8HIJjeayQaIg1dunaIsVXYGqfq_sDSv0,13274
 datachain/lib/dc/hf.py,sha256=PJl2wiLjdRsMz0SYbLT-6H8b-D5i2WjeH7li8HHOk_0,2145
 datachain/lib/dc/json.py,sha256=dNijfJ-H92vU3soyR7X1IiDrWhm6yZIGG3bSnZkPdAE,2733
@@ -125,7 +125,7 @@ datachain/model/ultralytics/pose.py,sha256=pBlmt63Qe68FKmexHimUGlNbNOoOlMHXG4fzX
 datachain/model/ultralytics/segment.py,sha256=63bDCj43E6iZ0hFI5J6uQfksdCmjEp6sEm1XzVaE8pw,2986
 datachain/query/__init__.py,sha256=7DhEIjAA8uZJfejruAVMZVcGFmvUpffuZJwgRqNwe-c,263
 datachain/query/batch.py,sha256=-goxLpE0EUvaDHu66rstj53UnfHpYfBUGux8GSpJ93k,4306
-datachain/query/dataset.py,sha256=C60VM0pScsrWcMqLNdX-tU0HE1SnEE9lRN3TU8CfTu4,61223
+datachain/query/dataset.py,sha256=mKee4PkQHYPT96utPjM1DocURU4TghAR7AHtYkzdqwY,61292
 datachain/query/dispatch.py,sha256=A0nPxn6mEN5d9dDo6S8m16Ji_9IvJLXrgF2kqXdi4fs,15546
 datachain/query/metrics.py,sha256=DOK5HdNVaRugYPjl8qnBONvTkwjMloLqAr7Mi3TjCO0,858
 datachain/query/params.py,sha256=O_j89mjYRLOwWNhYZl-z7mi-rkdP7WyFmaDufsdTryE,863
@@ -157,9 +157,9 @@ datachain/sql/sqlite/vector.py,sha256=ncW4eu2FlJhrP_CIpsvtkUabZlQdl2D5Lgwy_cbfqR
 datachain/toolkit/__init__.py,sha256=eQ58Q5Yf_Fgv1ZG0IO5dpB4jmP90rk8YxUWmPc1M2Bo,68
 datachain/toolkit/split.py,sha256=ktGWzY4kyzjWyR86dhvzw-Zhl0lVk_LOX3NciTac6qo,2914
 datachain/torch/__init__.py,sha256=gIS74PoEPy4TB3X6vx9nLO0Y3sLJzsA8ckn8pRWihJM,579
-datachain-0.24.0.dist-info/licenses/LICENSE,sha256=8DnqK5yoPI_E50bEg_zsHKZHY2HqPy4rYN338BHQaRA,11344
-datachain-0.24.0.dist-info/METADATA,sha256=QWSVON3r5d5d18gRMs9G5DNV4z-kBBY47dMYUEFR0b0,13281
-datachain-0.24.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-datachain-0.24.0.dist-info/entry_points.txt,sha256=0GMJS6B_KWq0m3VT98vQI2YZodAMkn4uReZ_okga9R4,49
-datachain-0.24.0.dist-info/top_level.txt,sha256=lZPpdU_2jJABLNIg2kvEOBi8PtsYikbN1OdMLHk8bTg,10
-datachain-0.24.0.dist-info/RECORD,,
+datachain-0.24.2.dist-info/licenses/LICENSE,sha256=8DnqK5yoPI_E50bEg_zsHKZHY2HqPy4rYN338BHQaRA,11344
+datachain-0.24.2.dist-info/METADATA,sha256=66Dz51BXpod8ZZG-pcmodALbszVZjNJXykMIrYRwXdA,13281
+datachain-0.24.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+datachain-0.24.2.dist-info/entry_points.txt,sha256=0GMJS6B_KWq0m3VT98vQI2YZodAMkn4uReZ_okga9R4,49
+datachain-0.24.2.dist-info/top_level.txt,sha256=lZPpdU_2jJABLNIg2kvEOBi8PtsYikbN1OdMLHk8bTg,10
+datachain-0.24.2.dist-info/RECORD,,