pixeltable 0.4.0rc3__py3-none-any.whl → 0.4.1__py3-none-any.whl

pixeltable/plan.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 import dataclasses
 import enum
 from textwrap import dedent
-from typing import Any, Iterable, Literal, NamedTuple, Optional, Sequence
+from typing import Any, Iterable, Literal, Optional, Sequence
 from uuid import UUID
 
 import sqlalchemy as sql
@@ -12,7 +12,6 @@ import pixeltable as pxt
 from pixeltable import catalog, exceptions as excs, exec, exprs
 from pixeltable.catalog import Column, TableVersionHandle
 from pixeltable.exec.sql_node import OrderByClause, OrderByItem, combine_order_by_clauses, print_order_by_clause
-from pixeltable.utils.sample import sample_key
 
 
 def _is_agg_fn_call(e: exprs.Expr) -> bool:
@@ -159,16 +158,6 @@ class SampleClause:
         return format(threshold_int, '08x') + 'ffffffffffffffffffffffff'
 
 
-class SamplingClauses(NamedTuple):
-    """Clauses provided when rewriting a SampleClause"""
-
-    where: exprs.Expr
-    group_by_clause: Optional[list[exprs.Expr]]
-    order_by_clause: Optional[list[tuple[exprs.Expr, bool]]]
-    limit: Optional[exprs.Expr]
-    sample_clause: Optional[SampleClause]
-
-
 class Analyzer:
     """
     Performs semantic analysis of a query and stores the analysis state.
@@ -180,6 +169,8 @@ class Analyzer:
     group_by_clause: Optional[list[exprs.Expr]]  # None for non-aggregate queries; [] for agg query w/o grouping
     grouping_exprs: list[exprs.Expr]  # [] for non-aggregate queries or agg query w/o grouping
     order_by_clause: OrderByClause
+    stratify_exprs: list[exprs.Expr]  # [] if no stratiifcation is required
+    sample_clause: Optional[SampleClause]  # None if no sampling clause is present
 
     sql_elements: exprs.SqlElementCache
 
@@ -200,6 +191,7 @@ class Analyzer:
         where_clause: Optional[exprs.Expr] = None,
         group_by_clause: Optional[list[exprs.Expr]] = None,
         order_by_clause: Optional[list[tuple[exprs.Expr, bool]]] = None,
+        sample_clause: Optional[SampleClause] = None,
     ):
         if order_by_clause is None:
             order_by_clause = []
@@ -213,6 +205,11 @@ class Analyzer:
         self.group_by_clause = (
             [e.resolve_computed_cols() for e in group_by_clause] if group_by_clause is not None else None
         )
+        self.sample_clause = sample_clause
+        if self.sample_clause is not None and self.sample_clause.is_stratified:
+            self.stratify_exprs = [e.resolve_computed_cols() for e in sample_clause.stratify_exprs]
+        else:
+            self.stratify_exprs = []
         self.order_by_clause = [OrderByItem(e.resolve_computed_cols(), asc) for e, asc in order_by_clause]
 
         self.sql_where_clause = None
@@ -228,8 +225,11 @@ class Analyzer:
                 self.all_exprs.append(join_clause.join_predicate)
         if self.group_by_clause is not None:
             self.all_exprs.extend(self.group_by_clause)
+        self.all_exprs.extend(self.stratify_exprs)
         self.all_exprs.extend(e for e, _ in self.order_by_clause)
         if self.filter is not None:
+            if sample_clause is not None:
+                raise excs.Error(f'Filter {self.filter} not expressible in SQL')
             self.all_exprs.append(self.filter)
 
         self.agg_order_by = []
@@ -691,25 +691,13 @@ class Planner:
         # 2. for component views: iterator args
         iterator_args = [target.iterator_args] if target.iterator_args is not None else []
 
-        # If this contains a sample specification, modify / create where, group_by, order_by, and limit clauses
         from_clause = FromClause(tbls=[view.base])
-        where, group_by_clause, order_by_clause, limit, sample_clause = cls.create_sample_clauses(
-            from_clause, target.sample_clause, target.predicate, None, [], None
-        )
-
-        # if we're propagating an insert, we only want to see those base rows that were created for the current version
         base_analyzer = Analyzer(
-            from_clause,
-            iterator_args,
-            where_clause=where,
-            group_by_clause=group_by_clause,
-            order_by_clause=order_by_clause,
+            from_clause, iterator_args, where_clause=target.predicate, sample_clause=target.sample_clause
         )
         row_builder = exprs.RowBuilder(base_analyzer.all_exprs, stored_cols, [])
 
-        if target.sample_clause is not None and base_analyzer.filter is not None:
-            raise excs.Error(f'Filter {base_analyzer.filter} not expressible in SQL')
-
+        # if we're propagating an insert, we only want to see those base rows that were created for the current version
         # execution plan:
         # 1. materialize exprs computed from the base that are needed for stored view columns
         # 2. if it's an iterator view, expand the base rows into component rows
@@ -723,19 +711,13 @@ class Planner:
 
         # Create a new analyzer reflecting exactly what is required from the base table
         base_analyzer = Analyzer(
-            from_clause,
-            base_output_exprs,
-            where_clause=where,
-            group_by_clause=group_by_clause,
-            order_by_clause=order_by_clause,
+            from_clause, base_output_exprs, where_clause=target.predicate, sample_clause=target.sample_clause
         )
         base_eval_ctx = row_builder.create_eval_ctx(base_analyzer.all_exprs)
         plan = cls._create_query_plan(
             row_builder=row_builder,
             analyzer=base_analyzer,
             eval_ctx=base_eval_ctx,
-            limit=limit,
-            sample_clause=sample_clause,
             with_pk=True,
             exact_version_only=view.get_bases() if propagates_insert else [],
         )
@@ -818,62 +800,6 @@ class Planner:
         prefetch_node = exec.CachePrefetchNode(tbl_id, file_col_info, input_node)
         return prefetch_node
 
-    @classmethod
-    def create_sample_clauses(
-        cls,
-        from_clause: FromClause,
-        sample_clause: SampleClause,
-        where_clause: Optional[exprs.Expr],
-        group_by_clause: Optional[list[exprs.Expr]],
-        order_by_clause: Optional[list[tuple[exprs.Expr, bool]]],
-        limit: Optional[exprs.Expr],
-    ) -> SamplingClauses:
-        """tuple[
-            exprs.Expr,
-            Optional[list[exprs.Expr]],
-            Optional[list[tuple[exprs.Expr, bool]]],
-            Optional[exprs.Expr],
-            Optional[SampleClause],
-        ]:"""
-        """Construct clauses required for sampling under various conditions.
-        If there is no sampling, then return the original clauses.
-        If the sample is stratified, then return only the group by clause. The rest of the
-        mechanism for stratified sampling is provided by the SampleSqlNode.
-        If the sample is non-stratified, then rewrite the query to accommodate the supplied where clause,
-        and provide the other clauses required for sampling
-        """
-
-        # If no sample clause, return the original clauses
-        if sample_clause is None:
-            return SamplingClauses(where_clause, group_by_clause, order_by_clause, limit, None)
-
-        # If the sample clause is stratified, create a group by clause
-        if sample_clause.is_stratified:
-            group_by = sample_clause.stratify_exprs
-            # Note that limit is not possible here
-            return SamplingClauses(where_clause, group_by, order_by_clause, None, sample_clause)
-
-        else:
-            # If non-stratified sampling, construct a where clause, order_by, and limit clauses
-            # Construct an expression for sorting rows and limiting row counts
-            s_key = sample_key(
-                exprs.Literal(sample_clause.seed), *cls.rowid_columns(from_clause._first_tbl.tbl_version)
-            )
-
-            # Construct a suitable where clause
-            where = where_clause
-            if sample_clause.fraction is not None:
-                fraction_md5_hex = exprs.Expr.from_object(
-                    sample_clause.fraction_to_md5_hex(float(sample_clause.fraction))
-                )
-                f_where = s_key < fraction_md5_hex
-                where = where & f_where if where is not None else f_where
-
-            order_by: list[tuple[exprs.Expr, bool]] = [(s_key, True)]
-            limit = exprs.Literal(sample_clause.n)
-            # Note that group_by is not possible here
-            return SamplingClauses(where, None, order_by, limit, None)
-
     @classmethod
     def create_query_plan(
         cls,
@@ -898,21 +824,15 @@ class Planner:
         if exact_version_only is None:
             exact_version_only = []
 
-        # Modify clauses to include sample clause
-        where, group_by_clause, order_by_clause, limit, sample = cls.create_sample_clauses(
-            from_clause, sample_clause, where_clause, group_by_clause, order_by_clause, limit
-        )
-
         analyzer = Analyzer(
             from_clause,
             select_list,
-            where_clause=where,
+            where_clause=where_clause,
             group_by_clause=group_by_clause,
             order_by_clause=order_by_clause,
+            sample_clause=sample_clause,
         )
         row_builder = exprs.RowBuilder(analyzer.all_exprs, [], [])
-        if sample_clause is not None and analyzer.filter is not None:
-            raise excs.Error(f'Filter {analyzer.filter} not expressible in SQL')
 
         analyzer.finalize(row_builder)
         # select_list: we need to materialize everything that's been collected
@@ -923,7 +843,6 @@ class Planner:
             analyzer=analyzer,
             eval_ctx=eval_ctx,
             limit=limit,
-            sample_clause=sample,
             with_pk=True,
             exact_version_only=exact_version_only,
         )
@@ -939,7 +858,6 @@ class Planner:
         analyzer: Analyzer,
         eval_ctx: exprs.RowBuilder.EvalCtx,
         limit: Optional[exprs.Expr] = None,
-        sample_clause: Optional[SampleClause] = None,
         with_pk: bool = False,
         exact_version_only: Optional[list[catalog.TableVersionHandle]] = None,
     ) -> exec.ExecNode:
@@ -966,6 +884,7 @@ class Planner:
         # - join clause subexprs
         # - subexprs of Where clause conjuncts that can't be run in SQL
         # - all grouping exprs
+        # - all stratify exprs
         candidates = list(
             exprs.Expr.list_subexprs(
                 analyzer.select_list,
@@ -980,10 +899,12 @@ class Planner:
             candidates.extend(
                 exprs.Expr.subexprs(analyzer.filter, filter=sql_elements.contains, traverse_matches=False)
             )
-        if analyzer.group_by_clause is not None:
-            candidates.extend(
-                exprs.Expr.list_subexprs(analyzer.group_by_clause, filter=sql_elements.contains, traverse_matches=False)
-            )
+        candidates.extend(
+            exprs.Expr.list_subexprs(analyzer.grouping_exprs, filter=sql_elements.contains, traverse_matches=False)
+        )
+        candidates.extend(
+            exprs.Expr.list_subexprs(analyzer.stratify_exprs, filter=sql_elements.contains, traverse_matches=False)
+        )
         # not isinstance(...): we don't want to materialize Literals via a Select
         sql_exprs = exprs.ExprSet(e for e in candidates if not isinstance(e, exprs.Literal))
 
@@ -1028,6 +949,15 @@ class Planner:
             # we need to order the input for window functions
             plan.set_order_by(analyzer.get_window_fn_ob_clause())
 
+        if analyzer.sample_clause is not None:
+            plan = exec.SqlSampleNode(
+                row_builder,
+                input=plan,
+                select_list=tbl_scan_exprs,
+                sample_clause=analyzer.sample_clause,
+                stratify_exprs=analyzer.stratify_exprs,
+            )
+
         plan = cls._insert_prefetch_node(tbl.tbl_version.id, row_builder, plan)
 
         if analyzer.group_by_clause is not None:
@@ -1050,26 +980,12 @@ class Planner:
                 sql_elements.contains_all(analyzer.select_list)
                 and sql_elements.contains_all(analyzer.grouping_exprs)
                 and isinstance(plan, exec.SqlNode)
-                and plan.to_cte(keep_pk=(sample_clause is not None)) is not None
+                and plan.to_cte() is not None
             ):
-                if sample_clause is not None:
-                    plan = exec.SqlSampleNode(
-                        row_builder,
-                        input=plan,
-                        select_list=analyzer.select_list,
-                        stratify_exprs=analyzer.group_by_clause,
-                        sample_clause=sample_clause,
-                    )
-                else:
-                    plan = exec.SqlAggregationNode(
-                        row_builder,
-                        input=plan,
-                        select_list=analyzer.select_list,
-                        group_by_items=analyzer.group_by_clause,
-                    )
+                plan = exec.SqlAggregationNode(
+                    row_builder, input=plan, select_list=analyzer.select_list, group_by_items=analyzer.group_by_clause
+                )
             else:
-                if sample_clause is not None:
-                    raise excs.Error('Sample clause not supported with Python aggregation')
                 input_sql_node = plan.get_node(exec.SqlNode)
                 assert combined_ordering is not None
                 input_sql_node.set_order_by(combined_ordering)

pixeltable/share/packager.py

@@ -127,7 +127,7 @@ class TablePackager:
         # We use snappy compression for the Parquet tables; the entire bundle will be bzip2-compressed later, so
         # faster compression should provide good performance while still reducing temporary storage utilization.
         parquet_writer = pq.ParquetWriter(parquet_file, parquet_schema, compression='SNAPPY')
-        filter_tv = self.table._tbl_version.get()
+        filter_tv = self.table._tbl_version_path.tbl_version.get()
         row_iter = tv.store_tbl.dump_rows(tv.version, filter_tv.store_tbl, filter_tv.version)
         for pa_table in self.__to_pa_tables(row_iter, sql_types, media_cols, parquet_schema):
             parquet_writer.write_table(pa_table)
@@ -238,7 +238,7 @@ class TablePackager:
         - Documents are replaced by a thumbnail as a base64-encoded webp
         """
         # First 8 columns
-        preview_cols = dict(itertools.islice(self.table._schema.items(), 0, 8))
+        preview_cols = dict(itertools.islice(self.table._get_schema().items(), 0, 8))
         select_list = [self.table[col_name] for col_name in preview_cols]
         # First 5 rows
         rows = list(self.table.select(*select_list).head(n=5))

pixeltable/type_system.py

@@ -395,6 +395,36 @@ class ColumnType:
                 raise excs.Error(f'Standard Python type `{name}` cannot be used here; use `{suggestion}` instead')
         raise excs.Error(f'Unknown type: {t}')
 
+    @classmethod
+    def from_json_schema(cls, schema: dict[str, Any]) -> Optional[ColumnType]:
+        # We first express the JSON schema as a Python type, and then convert it to a Pixeltable type.
+        # TODO: Is there a meaningful fallback if one of these operations fails? (Maybe another use case for a pxt Any
+        #     type?)
+        py_type = cls.__json_schema_to_py_type(schema)
+        return cls.from_python_type(py_type) if py_type is not None else None
+
+    @classmethod
+    def __json_schema_to_py_type(cls, schema: dict[str, Any]) -> Union[type, _GenericAlias, None]:
+        if 'type' in schema:
+            if schema['type'] == 'null':
+                return type(None)
+            if schema['type'] == 'string':
+                return str
+            if schema['type'] == 'integer':
+                return int
+            if schema['type'] == 'number':
+                return float
+            if schema['type'] == 'boolean':
+                return bool
+            if schema['type'] in ('array', 'object'):
+                return list
+        elif 'anyOf' in schema:
+            subscripts = tuple(cls.__json_schema_to_py_type(subschema) for subschema in schema['anyOf'])
+            if all(subscript is not None for subscript in subscripts):
+                return Union[subscripts]
+
+        return None
+
     def validate_literal(self, val: Any) -> None:
         """Raise TypeError if val is not a valid literal for this type"""
         if val is None:

{pixeltable-0.4.0rc3.dist-info → pixeltable-0.4.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pixeltable
-Version: 0.4.0rc3
+Version: 0.4.1
 Summary: AI Data Infrastructure: Declarative, Multimodal, and Incremental
 License: Apache-2.0
 Keywords: data-science,machine-learning,database,ai,computer-vision,chatbot,ml,artificial-intelligence,feature-engineering,multimodal,mlops,feature-store,vector-database,llm,genai

{pixeltable-0.4.0rc3.dist-info → pixeltable-0.4.1.dist-info}/RECORD

@@ -1,48 +1,48 @@
-pixeltable/__init__.py,sha256=-uXHuiXH98kAlCupUTPbkBh4ToZgxcYUOk7-c9hqCC8,1439
-pixeltable/__version__.py,sha256=l1gVhW4YjeNouBt4ti5Rxsd1ETc8Ls5nKRC8aMceXrM,120
-pixeltable/catalog/__init__.py,sha256=rQmjveID4bk6NI4Ql09lGsZ0K0HVE2l1yqKAveipHzc,558
-pixeltable/catalog/catalog.py,sha256=ExHm2bbAs_4-T6zq3ow2fQ5a9Zja6G3WzHoo3Qk0X40,59399
-pixeltable/catalog/column.py,sha256=v3SE6CQs_ZXTDkBeiJ7gu9sxBJD-6EbkjsB1TZvLSJI,11280
+pixeltable/__init__.py,sha256=WHZxJNz7vYcx_z3ebEg-RF22lJmlkfeqF2gv6mnrM1A,1449
+pixeltable/__version__.py,sha256=BP1_olpMAIqE5wUeIGRoLW4mcG51_gER8ganU5vYe08,112
+pixeltable/catalog/__init__.py,sha256=5gYj-MoahjDAaU4gQNGao_JODRL_1WigVe79F828FcA,602
+pixeltable/catalog/catalog.py,sha256=hauRCj1aRb1C6C3AOBCjDdnCpHa46mEBlG7tVHHBxz8,71340
+pixeltable/catalog/column.py,sha256=SBJO7kpnHjJupMywrD3kDUyhvoyi8GgAcDgky3Bf-f8,11230
 pixeltable/catalog/dir.py,sha256=HFemOf67Nfw13EOpQsR_UgzP2L1w4LDfw2009DrSK0Y,2063
-pixeltable/catalog/globals.py,sha256=7fNUs67D18PZ1ZajcGz7KQOV6tuXEcYLSrePzgmDkRw,4036
-pixeltable/catalog/insertable_table.py,sha256=tl3tvMVPzi9A_SRARV6660NQnt8w09hNbOYZwMsSJw8,9195
+pixeltable/catalog/globals.py,sha256=L09X-TA5iIcFQ-sQ7a372RfQ0bJpFs7C0sVHNXRG5Cw,4261
+pixeltable/catalog/insertable_table.py,sha256=f4Lwxlbgqi48BTWzWVEtA9i_SZ5jw47-Psw9WbKf6jw,9338
 pixeltable/catalog/named_function.py,sha256=vZ-j7P4HugWh9OmUzBMwyRYvO3tQn9jWyJz_1stPavU,1210
 pixeltable/catalog/path.py,sha256=gk8TIlO_7Jpji5mAN0dUNvHmvU0uneTHeB_qCTWnszQ,2529
-pixeltable/catalog/schema_object.py,sha256=B9Am6W1bWnAMFZ4qFTfYw6g8F0-q_2aezBB1bUHL_RY,1812
-pixeltable/catalog/table.py,sha256=JUiQPQP33486KiPVio8aBTibNGUYwulDgU1UfGDaprU,68609
-pixeltable/catalog/table_version.py,sha256=1dq2uq-I2E4d2SkBMA5MQFqkutPQCi-7HlNo24Trrjw,66775
-pixeltable/catalog/table_version_handle.py,sha256=E0iPIp2PUbE7r3WZzerBJ7fgKX3vaTWPnn6cOwKLF-A,2581
-pixeltable/catalog/table_version_path.py,sha256=2JEprCdDzCfr9dPxlMhsQ80jYqJ0-8uBNgNXt3D-bao,9042
-pixeltable/catalog/view.py,sha256=1UN-k0Ck0IWyy2Ae5b520L08GAfQ17L4FaZtS2bQ-TA,14786
-pixeltable/config.py,sha256=gnRI4G9GE7mQJDcMcn8JsEzYk8oKVfHB-BwoLRWnRDo,3971
-pixeltable/dataframe.py,sha256=L1lNcfHtxxRw_b8kgkmvcuYXQyniAzwLnI4ww79P5B4,60393
-pixeltable/env.py,sha256=BtBxkKgU3wi6_0rfor3k3Kx6SF660JH24sWprbBzB8w,36132
-pixeltable/exceptions.py,sha256=eI2f4oXqv678_fDOxeYOTPE_EF1CMX3pwCqtJvN8i7s,926
+pixeltable/catalog/schema_object.py,sha256=jqOhZcI9QbT_EseDRQsVrp4pZ7jKB1wy4Sa-6aAvUCI,2004
+pixeltable/catalog/table.py,sha256=IXgh7VSyGI-tvNl0G1_zQP7qJ9QyZucefXWX6IQ8rG4,69993
+pixeltable/catalog/table_version.py,sha256=mZSBzMI6RTIfCjySd2OZjz7Eqc6h9g_BLbZ4PPRxfIM,66158
+pixeltable/catalog/table_version_handle.py,sha256=aEIM9gX1MK95dip7OYmM9iGidE41SRi4lvoOqcL6kNo,2600
+pixeltable/catalog/table_version_path.py,sha256=uRATYbAN0OXEkW1GZEC4hZulzvh8IJRwaN66ifUTJfw,9787
+pixeltable/catalog/view.py,sha256=czgvZf9Z-VYXaVadSktrVtd9BdgLABaQWEjOcWJySck,14645
+pixeltable/config.py,sha256=HIAyV3UGlcuQnHofKsid7wup518q_WfN6G-KB4yu_3g,4280
+pixeltable/dataframe.py,sha256=dTpjM72Cj2cYMo0_HH2BJ6P3s9rQAF4suuL413BDyJw,60673
+pixeltable/env.py,sha256=I8rvQQjvHB68rYCoSurR4pjn5qm6EvSwRL1CblUVPWs,35625
+pixeltable/exceptions.py,sha256=Gm8d3TL2iiv6Pj2DLd29wp_j41qNBhxXL9iTQnL4Nk4,1116
 pixeltable/exec/__init__.py,sha256=hQvj4ra4ubxu94qyuCBTHKsuYGzundkTTluOTIb5Bx8,524
 pixeltable/exec/aggregation_node.py,sha256=HqzISO1nv7_DFyqjZLRkjtbDJl9fIEto1i6Kh5ru8vA,4498
 pixeltable/exec/cache_prefetch_node.py,sha256=GOa70eJDFY3FQV3VvJOrUVI8LFvro-r-V6sh3w-eJAc,12130
 pixeltable/exec/component_iteration_node.py,sha256=FZszWHrzsjHxCbUTwXtJIlgQqgYtvKZB6QWiDGkfIbs,4757
 pixeltable/exec/data_row_batch.py,sha256=EAB15JRhXbWIe91x1J5N5lFiMXzjB8NGTFjZsBDSbf8,3393
 pixeltable/exec/exec_context.py,sha256=jKeLStfkjwCKKAooC-7a7qZUnZU5O0_JQhanhVerV9c,984
-pixeltable/exec/exec_node.py,sha256=WIN1sBEBNS7TlBk5QpHsDUCZUUdcvs6Os_Bxq2HoWdo,4077
+pixeltable/exec/exec_node.py,sha256=MsuCO7nCpmqfuNxTNKsz36sJVDrR-o-0-3S2FcXLwvM,4237
 pixeltable/exec/expr_eval/__init__.py,sha256=sQThSEByK_DLfB-_-18RFhpARx49cSXYEkpCDyi0vQI,61
 pixeltable/exec/expr_eval/evaluators.py,sha256=-6s_y29Wh8p35SVKkXtnA0NkzcHVw1Z8PgHGiFrMsqs,17135
-pixeltable/exec/expr_eval/expr_eval_node.py,sha256=ABkC2yAcMSCtQ7AvNAn6rfj3AMscVzezttKSs1ExNhw,18972
+pixeltable/exec/expr_eval/expr_eval_node.py,sha256=klPhvsug91GiPIHkwcTj1ympxsoj8nbNHzkzkC-NR88,18953
 pixeltable/exec/expr_eval/globals.py,sha256=fFrj2O53TgHDfVF8dgnyn1fPLi4ZHQuylewf5aHMwYk,7752
 pixeltable/exec/expr_eval/row_buffer.py,sha256=YY0thdlMNNReEOTyPp36xKPeMeXSZ0VrI9bJsXgo7sU,2744
 pixeltable/exec/expr_eval/schedulers.py,sha256=tAvCQKa1q0x7y7cdnGcTGbeku8QcoKH1GkgSm8ktOnM,17000
 pixeltable/exec/in_memory_data_node.py,sha256=vmxD2Jwn15Wjkf_3wufr35SPjb60H_I4zpUKaO1Zo_s,3592
 pixeltable/exec/row_update_node.py,sha256=zU0eSyn81-vRrjAMOadRqU8luTshnPUtIbS7npyLBKY,2798
-pixeltable/exec/sql_node.py,sha256=MM9wD0AMXh11qJayhMgieCJ36hww-ym1p7kfS2y9-wM,26540
+pixeltable/exec/sql_node.py,sha256=cMoBGPOFVmvL3UFjCKxhU3huJu_ko0PRr55-XhbF1i0,27572
 pixeltable/exprs/__init__.py,sha256=AxSMjKNavCT9F6vBaNR-nwX2iupAI5hbMb5hEj65Tfk,1096
 pixeltable/exprs/arithmetic_expr.py,sha256=sZPao0qdFWbrDx0eiAVxw1wGHJXZ5ZoCpQaScysBldE,7333
 pixeltable/exprs/array_slice.py,sha256=8Zv0E2RghdJi1Mbk0kKtOz2ccGQuXwLLb6R9v1jk7hA,2180
-pixeltable/exprs/column_property_ref.py,sha256=28UcsrMIoXmMy6DAgJ0GP6Vw4sYwXcspVySmKSvejaE,4178
-pixeltable/exprs/column_ref.py,sha256=LljuQiNYh6dQa_EV57ylPfNkQD-lfHBFTGDY8WlnnfY,14942
+pixeltable/exprs/column_property_ref.py,sha256=jIin85Ubjvx2L_c_VRROVA8MVrMVGO5P4OPwuZ4wwTY,4268
+pixeltable/exprs/column_ref.py,sha256=PaKMLMs1t8Ld8zdTHJujQxtu_V5bgYh3rvPEAXg13jc,14956
 pixeltable/exprs/comparison.py,sha256=lgaRx000ZaNH10A4hrtsi5XoZKE-CNEONGMi7jxJfcM,5133
 pixeltable/exprs/compound_predicate.py,sha256=vJVRVueAmaKnjiHCLWyh8wHgktzzK0DVqbOIQJgTjF8,3801
 pixeltable/exprs/data_row.py,sha256=8p0eWunvg_2ZH-e7Gtz_woxBAoizaCV2QtN87Is8_ys,11484
-pixeltable/exprs/expr.py,sha256=YFPL81s9N1_Obl15zpk_yYUvKqVoBT8fpAXhkYqgHLs,36000
+pixeltable/exprs/expr.py,sha256=uLEMuJzHwPW3hBIcsASnhjcPyGBkY8_8sFCqgRetKP0,36013
 pixeltable/exprs/expr_dict.py,sha256=2ZeZ0eACx3VrRNEOjipuT5WxOIzjXQ_DSip8NTH0KRo,1584
 pixeltable/exprs/expr_set.py,sha256=OlRTbHAAYH2fOEs1HE-8DIu7Z247xVfoT_9Y58GZoOQ,2559
 pixeltable/exprs/function_call.py,sha256=_PxrEACVyiihdQdmTiiSv5WkZfOXSQFcGO18wPueM_Y,21989
@@ -66,18 +66,19 @@ pixeltable/ext/__init__.py,sha256=UgDXWzGWiQIrwOuEvWTePLBcR2kecllPAE7gp-42Awg,45
 pixeltable/ext/functions/__init__.py,sha256=Ox3kUHn5IslVEmEKsjrHfkHDrUkmLl9RCO2YkrPJkgc,193
 pixeltable/ext/functions/whisperx.py,sha256=qda6kFQSvZTY2asfrYPwHb1cvSa03LbhJ-Wf9b7qPhw,2355
 pixeltable/ext/functions/yolox.py,sha256=dX22nMb-0n2hZi7WhZ1Y4LIpFk5loyeXXuSUcc2Fgrg,3724
-pixeltable/func/__init__.py,sha256=2BtyvXvFrLF0S5Qnmtg9S7dDSEVoBh5GDX6XFC25d7o,509
+pixeltable/func/__init__.py,sha256=SQPtGr_9dZNyXzxaZQcP3oVLKnbbs4UqV6sg8XUQHxQ,535
 pixeltable/func/aggregate_function.py,sha256=5_MgqHAlMaacX2sPIHv_auTvYXtqR5MIZy_WqYQSdho,13264
 pixeltable/func/callable_function.py,sha256=g_pA-g631YcFGLix9PpHYfgjOeS2qF0Csm1VxX8fah0,9278
 pixeltable/func/expr_template_function.py,sha256=wEidKrOBTZkA3U1PAtG6-6RlDFiiRJszIG4zNOuPcNY,5940
 pixeltable/func/function.py,sha256=w1U3j8XNeE4ZZ-rKuG13aTa8YGFkWAXjalh4j29_-e4,23136
 pixeltable/func/function_registry.py,sha256=7AQ1bdF2DJbTRn9xx6s5cC_VHtCBXGt_GyJJEjJHcMw,12308
 pixeltable/func/globals.py,sha256=5Wo4GPxYgHRRk5nvV0h_lAthKSalxKvj5n1p-uMPR0U,1501
-pixeltable/func/query_template_function.py,sha256=25nvgadWgk16BU7X5uI7PvKfkoluJSuOqsAcY5AN-Yw,7915
+pixeltable/func/mcp.py,sha256=P9M2w8cm7ad-XmAcf2ZThfWmD8W46De1spwX98bZL4Y,2861
+pixeltable/func/query_template_function.py,sha256=0wCcU06bv9KKJRlOsoLg8-vVNSLIYpKDCdYkUV7WShY,8040
 pixeltable/func/signature.py,sha256=0PI6xdhLgwy9-GMkzkm7GlsBnsNMiS9aoNI9LWXwvN0,13700
-pixeltable/func/tools.py,sha256=bmHnnd9lXkQ8sYHp5RSMF56NimSTE3uhG2xbIxs4Np4,5726
-pixeltable/func/udf.py,sha256=qQfaX1O3ZhUvSgiNnitW7nRKnZFJ5yu_Fj9ioqQgjqg,13219
-pixeltable/functions/__init__.py,sha256=x_ZodeEtayD2XDjlDZE4L_QTzajF6riftIc5P4ZjEiY,578
+pixeltable/func/tools.py,sha256=hKmQFvfpBvtLcItPRpqAmqt_tDg6latwyfv5FXBofBc,6074
+pixeltable/func/udf.py,sha256=6tKpMt37t3BmXwRyA5fFAd6OM4D5EPEd2KuAr7DQhr0,13231
+pixeltable/functions/__init__.py,sha256=Akk6Nk-rpz2D_V4kJTfyP56xnNbCz3EtxVAuwLoiysA,588
 pixeltable/functions/anthropic.py,sha256=G2E0sH5vP933eZZxhz1tOByy5cg6N2XPvhSqIBzqufo,8782
 pixeltable/functions/audio.py,sha256=7bsm4igQEW7RYSrSevwqaUOqyEnvBbPbJ8c-VknDl1E,657
 pixeltable/functions/bedrock.py,sha256=lTCFHjYunF3minBGWcjXR90yJ8resFjXr4niyKhfxms,4217
@@ -86,37 +87,38 @@ pixeltable/functions/deepseek.py,sha256=IAo2e_DhkM0A5NrskxuPQUGYzIYAl4do_mdO1Qc3
 pixeltable/functions/fireworks.py,sha256=q7eWlYfiWbA0d9r3CB_NAe1fw3q-Z7qsw2gyGJNgWLQ,4786
 pixeltable/functions/gemini.py,sha256=ZsbySkoMdOgZEUfFUccDbIdrbLb6DGRxzD88fHW-cRI,8317
 pixeltable/functions/globals.py,sha256=ZXBV2LPXT2-yQYHHE7q8N1WdAr0WxiIO1ax0qwxhmK8,5118
-pixeltable/functions/huggingface.py,sha256=KM1OH0Jt6XWF2jfpHb6rGhi1mV-AQNYAsHAyQfzW4qw,20560
+pixeltable/functions/groq.py,sha256=FpR_LJpfZfzyhEvoBMMbQpQ-VQSRzBsS9U21qaINwww,3593
+pixeltable/functions/huggingface.py,sha256=cJyf86qMcvivkGGNduNHAvh_idI-e4wJm0Zje1KJ2vQ,20611
 pixeltable/functions/image.py,sha256=IKXljMma-uU88efptC3F4aywau7DYcD-Nqd3YpmRNRw,13971
 pixeltable/functions/json.py,sha256=d7-AvwytUQtQYF_JnWJkptT_Yq0NgMpWfVk-m3U6qTY,807
 pixeltable/functions/llama_cpp.py,sha256=1QB4vQ7J4Za1mL93bRIBXgokNtpzzYr_QU6KF27i9xo,3919
 pixeltable/functions/math.py,sha256=eZEFjXxNHDHjcCsOMhzfNbJthTsmtNxtSFV8AEeRIfM,4979
-pixeltable/functions/mistralai.py,sha256=PTXQegC2LO5Pw0zXBO_SVV7I2c5qBvqVVgfm_mK1ir0,5845
+pixeltable/functions/mistralai.py,sha256=Fk52mfWUfxVy-yCxhH6wrGS7nLLSiOOrWxbTkkiQ-O8,5542
 pixeltable/functions/ollama.py,sha256=4-6h9Foq_7Ut7JtEHGkeg1KbuKaFywSuMrKiw0xAyCA,4231
-pixeltable/functions/openai.py,sha256=kzi8HApD971O54Xl82G0jI7n0E2ui476kV1wAKObx88,27768
-pixeltable/functions/replicate.py,sha256=SLMPNi44QMa16TVTImZRkNMXXyRZ0mmpnK6P5uXQE0k,2467
+pixeltable/functions/openai.py,sha256=SxhYrL3vgIfjzwCPnjR6yoaNr7BbFwpGy7Su1FSY7G4,27713
+pixeltable/functions/replicate.py,sha256=sPvRGr0j0kCDc6Vz3mPUioFflApijukvZWJJUO2bqIQ,2429
 pixeltable/functions/string.py,sha256=LdBNOna5PUSPmM5VlJ_qhmwzyFhumW0k6Dvx2rXSZtc,25356
 pixeltable/functions/timestamp.py,sha256=0zp4urJagCcNLfJE0ltTCft-J9qs2C716TmRngKYaa0,9171
 pixeltable/functions/together.py,sha256=A8J19BXywyWQ6a2_n05-8uIG5jquOBGqPmW3mb-NrIc,8842
-pixeltable/functions/util.py,sha256=lVya13gcao8T34OGX7zy1cglQPNwaBbSBw57bVPyHNs,745
+pixeltable/functions/util.py,sha256=uQNkyBSkTVMe1wbUI2Q0nz-mM3qPVTF86yK8c9OFIcE,954
 pixeltable/functions/video.py,sha256=jS4YhMofD448YhGtI6ZXBAkeGw_AYYQTN0AbgHh_hok,6933
 pixeltable/functions/vision.py,sha256=_a0wY3akkVhWnnxlq__1VzSLylytlNadpNOOPOwSfLk,15393
 pixeltable/functions/whisper.py,sha256=c9E6trhc2UcShVaGaEBCUEpArke1ql3MV5We0qSgmuU,2960
-pixeltable/globals.py,sha256=cM3uKvDJWIs1I9A8rzXaZsKCDSaLBh7KVb1lakQqGzk,32377
+pixeltable/globals.py,sha256=cQUzDiYzDftRhX1jwi1hi6OYx0zlDAMr04nShq75Knk,32226
 pixeltable/index/__init__.py,sha256=97aFuxiP_oz1ldn5iq8IWApkOV8XG6ZIBW5-9rkS0vM,122
 pixeltable/index/base.py,sha256=200s7v3Zy810bRlbSAYzxxaEjVssl6r8esTHiSvWRwQ,1704
 pixeltable/index/btree.py,sha256=8B06D67ay0DFUtEBC5q4bLjxMq7ILpKyyoLAiSaamzA,2503
 pixeltable/index/embedding_index.py,sha256=B_k_3UJmSv7t2ljUg8GC_D4t1jc03PVsTAvxqiTmHBA,11754
 pixeltable/io/__init__.py,sha256=Yjq13pBCBoaZv-OkIY2XSusVOC5b6cB5C6NbgJq5H1g,620
 pixeltable/io/datarows.py,sha256=p1UGxQOTjqI6kgQNAa3aj8TkylcNDtaGBTorOg_Pk84,6088
-pixeltable/io/external_store.py,sha256=0ayZybxKRc600n-W7-3kEo0n_6LjAkIz-B4kLoPyukc,15554
+pixeltable/io/external_store.py,sha256=ZGZBMlfeg2euSMub0RK9Yv_h0gnItmEZPRg2IUDmJaE,15566
 pixeltable/io/fiftyone.py,sha256=v0r28bIk2I0TRP5DfVHtBIUa4DpIJDK5sgExxOmHZ_w,6882
 pixeltable/io/globals.py,sha256=Z8ww-Pcm59ql1tvame8z0Mu1thIy5BPbW-TswGRXt4s,11368
 pixeltable/io/hf_datasets.py,sha256=gWyBH_0iFvxcrrxMY9_W399ZRcNDCmWFOAMmb1apnY0,5246
-pixeltable/io/label_studio.py,sha256=rEJCN4hmONRMXVGkj1Cq_E-wowcvg6d7KToXxheOh2w,31242
+pixeltable/io/label_studio.py,sha256=y_0yEQWKgoLoHUQll6RTbm_Ou7LU882VZoEGb4WWnFs,31260
 pixeltable/io/pandas.py,sha256=AbOeRDlA4MvUvianSKixsU-x-64nasPWw4HCHD6emz4,8981
 pixeltable/io/parquet.py,sha256=-cxyy9wMRzGFDJWhUIjACfQMyAmajyoFcTXSkB8qESE,7818
-pixeltable/io/table_data_conduit.py,sha256=oag9BDAjz6CZCCvZWuk3FnBNWb_m01bUzvbM0BW8tPI,21997
+pixeltable/io/table_data_conduit.py,sha256=8SEcOPTgPiKHqlDp0rvGcPOF4v8jRX5TwHwfi5MHYt4,22003
 pixeltable/io/utils.py,sha256=YMfhpqMitWz1PhXJGkCNOgNtEM1AZ55S0zXVhljC5kY,4260
 pixeltable/iterators/__init__.py,sha256=bU4EmbX85J1URmRw6G71f2I77b1ctqngEOwDmRB3T0w,455
 pixeltable/iterators/audio.py,sha256=wSVGdL5GeO3uY_lU-pNlY49E5dExIaJWY6oaXm-MnSU,9150
@@ -125,7 +127,7 @@ pixeltable/iterators/document.py,sha256=wJYSnzusJFaxipv5y0uQw-surN9fFz0Aq-s7w_l_
 pixeltable/iterators/image.py,sha256=nWm-03CxNvHRdTr8U6PvWEnEiquqIQNG5rB-3Y44Mm4,3440
 pixeltable/iterators/string.py,sha256=URj5edWp-CsorjN_8nnfWGvtIFs_Zh4VPm6htlJbFkU,1257
 pixeltable/iterators/video.py,sha256=L5S1YPmT_zM11vW9fK6d5nQpUvHVewQWmfDmy4BD45E,9134
-pixeltable/metadata/__init__.py,sha256=fZ3HaB-L8lHJdrtmKqfIJJlvYG0CqzzxKFRrh9IWoa0,3154
+pixeltable/metadata/__init__.py,sha256=3_uviy74LAghHBKU-Ptg3BFmIrH6AoUr6nwvAnfWjWQ,3154
 pixeltable/metadata/converters/convert_10.py,sha256=myYIo1DyccnsQUxDKG6mafnU5ge_EhZpHg_pesKBoK4,708
 pixeltable/metadata/converters/convert_12.py,sha256=Ci-qyZW1gqci-8wnjeOB5afdq7KTuN-hVSV9OqSPx8g,162
 pixeltable/metadata/converters/convert_13.py,sha256=B-_EkL0pSl1mAiv6DymeUAyBQUcYcV1qDdNz3Q359kc,1369
@@ -152,16 +154,17 @@ pixeltable/metadata/converters/convert_33.py,sha256=ZZV3FTyyouBM1eNymXxfHV-Oqmgu
 pixeltable/metadata/converters/convert_34.py,sha256=1hi7m49CMzHRD25rrePS-SMCsZ-4opzDhY0JqU8Jzw4,690
 pixeltable/metadata/converters/convert_35.py,sha256=c88qft0RFQbdFIE_PZRHMjeku1r5HCLN7wrvndQSXdI,266
 pixeltable/metadata/converters/convert_36.py,sha256=g1rhZhAYfZpAwUgE3D1aipIR4RNvikhbKcrnBJzm0wM,1215
+pixeltable/metadata/converters/convert_37.py,sha256=IVZGtKFaaYMGBs39V_H_okWvpxxadTUWqxoln0cNeQI,392
 pixeltable/metadata/converters/util.py,sha256=95pfg9amEOmhho32PIbNYnqagVIN9adIcLXxB6zSYDY,7527
-pixeltable/metadata/notes.py,sha256=7HKk4yQLh8zaTVgv4pGXvpn3PoF5x7OFO8YSog5TLHI,1408
-pixeltable/metadata/schema.py,sha256=5r-loBIyYiZdkbszrLdn0XktOgDh96UVp_GScmm-14g,11489
-pixeltable/plan.py,sha256=bL6B32J63P1QtGjLhwwcfbh2SrSEeuyb1Zh6txdKQ6Q,52136
+pixeltable/metadata/notes.py,sha256=OVHn9iW_y1nyGaMfu0qG3IBkuCdjgNdI6lJnR7LeRIg,1441
+pixeltable/metadata/schema.py,sha256=l8ZGcASiTrtaMW9jPDMtZmoT2BamGIDhujNUVgdag5E,11720
+pixeltable/plan.py,sha256=lM_wZZgxIfiIgN3-IIuyCd2URn-p29T1DwAe6vZ284U,48540
 pixeltable/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 pixeltable/share/__init__.py,sha256=AtR4nS6YkfkFRkXA-zZXFTK5pSQjHry8MnxdVLUk5SA,68
-pixeltable/share/packager.py,sha256=udwsorQDWX0N9eHcPHlT5xyAeUxGvT7DBnsFuq8PALM,32149
+pixeltable/share/packager.py,sha256=wqOuuUXOor6YLzR9oZSYuJmdbkfBetZB3NkndsqbX3I,32172
 pixeltable/share/publish.py,sha256=U6PzOUYiZaPu-sVNjh2nN8qzY2-uMsYeTwQCCuGk7Jg,6537
 pixeltable/store.py,sha256=2tfMgJNVVbGShgAKghGAIFMZ-ViJHE7N9kWd2ugbi2A,24691
-pixeltable/type_system.py,sha256=DSrof2NgKhBzvt7pbDNrGlZ3rkkDJ7MQsQ9rqk9N9pA,53988
+pixeltable/type_system.py,sha256=P-ykQDPKbMZHtkozdrcBBpuX60Zc1nzwjwzSwlxIzPg,55319
 pixeltable/utils/__init__.py,sha256=Pwgu-Sg1XkxzdCZ4ZhWP77UgLP3tnQsyCKaUJLF4ajo,1741
 pixeltable/utils/arrow.py,sha256=74wIy58rDYZJBVQ1g85NqzFyiQBvEQhnJ0Gi82iZ0dw,6421
 pixeltable/utils/coco.py,sha256=Y1DWVYguZD4VhKyf7JruYfHWvhkJLq39fzbiSm5cdyY,7304
@@ -179,11 +182,10 @@ pixeltable/utils/iceberg.py,sha256=L_s9G9NMIGMQdRHtNkks6ntTVW4DKKAw97R9gRmtw5s,5
 pixeltable/utils/media_store.py,sha256=Dhsnj1ZPRSX0iyGOu4JU4pC3fvSBd7sQpruVHqzKm7A,3089
 pixeltable/utils/pytorch.py,sha256=564VHRdDHwD9h0v5lBHEDTJ8c6zx8wuzWYx8ZYjBxlI,3621
 pixeltable/utils/s3.py,sha256=pxip2MlCqd2Qon2dzJXzfxvwtZyc-BAsjAnLL4J_OXY,587
-pixeltable/utils/sample.py,sha256=Pj8oSZw0WsdASD1BpTtKiWP4cwef7KQqVAfIFKlJNxA,643
 pixeltable/utils/sql.py,sha256=Sa4Lh-VGe8GToU5W7DRiWf2lMl9B6saPqemiT0ZdHEc,806
 pixeltable/utils/transactional_directory.py,sha256=OFKmu90oP7KwBAljwjnzP_w8euGdAXob3y4Nx9SCNHA,1357
-pixeltable-0.4.0rc3.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-pixeltable-0.4.0rc3.dist-info/METADATA,sha256=iRhLy9JQXIsjwIOKQnvJBQCJuGfh_Xc56W3dcgWGs-A,20580
-pixeltable-0.4.0rc3.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-pixeltable-0.4.0rc3.dist-info/entry_points.txt,sha256=ToOd-pRgG7AitEBgYoBCRRB4-KVDQ0pj_9T4a1LgwA4,97
-pixeltable-0.4.0rc3.dist-info/RECORD,,
+pixeltable-0.4.1.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+pixeltable-0.4.1.dist-info/METADATA,sha256=k3yxQDjBX8NOeWjh2S25JaRMMnoksx_EKQtYmENuKpM,20577
+pixeltable-0.4.1.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+pixeltable-0.4.1.dist-info/entry_points.txt,sha256=ToOd-pRgG7AitEBgYoBCRRB4-KVDQ0pj_9T4a1LgwA4,97
+pixeltable-0.4.1.dist-info/RECORD,,

pixeltable/utils/sample.py

@@ -1,25 +0,0 @@
-import sqlalchemy as sql
-
-from pixeltable.func.udf import udf
-
-
-@udf
-def sample_key(seed: int, *key_fields: int) -> str:
-    """
-    Create a sample key from the given seed and key fields.
-
-    Args:
-        seed: The seed value.
-        rowids: The rowids to include in the sample key.
-
-    Returns:
-        A string key for each row
-    """
-    raise NotImplementedError('SampleKey creation is not implemented in python.')
-
-
-@sample_key.to_sql
-def _(seed: sql.ColumnElement, *key_fields: sql.ColumnElement) -> sql.ColumnElement:
-    from pixeltable.exec.sql_node import SqlSampleNode
-
-    return SqlSampleNode.key_sql_expr(seed, key_fields)