pixeltable 0.4.0rc1__py3-none-any.whl → 0.4.0rc3__py3-none-any.whl

pixeltable/share/packager.py

@@ -1,4 +1,7 @@
+import base64
 import datetime
+import io
+import itertools
 import json
 import logging
 import tarfile
@@ -10,15 +13,18 @@ from typing import Any, Iterator, Optional
 from uuid import UUID
 
 import more_itertools
+import numpy as np
+import PIL.Image
 import pyarrow as pa
 import pyarrow.parquet as pq
 import sqlalchemy as sql
 
 import pixeltable as pxt
-from pixeltable import catalog, exceptions as excs, metadata
+from pixeltable import catalog, exceptions as excs, metadata, type_system as ts
 from pixeltable.env import Env
 from pixeltable.metadata import schema
 from pixeltable.utils import sha256sum
+from pixeltable.utils.formatter import Formatter
 from pixeltable.utils.media_store import MediaStore
 
 _logger = logging.getLogger('pixeltable')
@@ -46,6 +52,10 @@ class TablePackager:
     media_files: dict[Path, str]  # Mapping from local media file paths to their tarball names
     md: dict[str, Any]
 
+    bundle_path: Path
+    preview_header: dict[str, str]
+    preview: list[list[Any]]
+
     def __init__(self, table: catalog.Table, additional_md: Optional[dict[str, Any]] = None) -> None:
         self.table = table
         self.tmp_dir = Path(Env.get().create_tmp_path())
@@ -67,7 +77,8 @@ class TablePackager:
         Export the table to a tarball containing Parquet tables and media files.
         """
         assert not self.tmp_dir.exists()  # Packaging can only be done once per TablePackager instance
-        _logger.info(f"Packaging table '{self.table._path()}' and its ancestors in: {self.tmp_dir}")
+
+        _logger.info(f'Packaging table {self.table._path()!r} and its ancestors in: {self.tmp_dir}')
         self.tmp_dir.mkdir()
         with open(self.tmp_dir / 'metadata.json', 'w', encoding='utf8') as fp:
             json.dump(self.md, fp)
@@ -75,12 +86,20 @@ class TablePackager:
         self.tables_dir.mkdir()
         with catalog.Catalog.get().begin_xact(for_write=False):
             for tv in self.table._tbl_version_path.get_tbl_versions():
-                _logger.info(f"Exporting table '{tv.get().versioned_name}'.")
+                _logger.info(f'Exporting table {tv.get().versioned_name!r}.')
                 self.__export_table(tv.get())
+
         _logger.info('Building archive.')
-        bundle_path = self.__build_tarball()
-        _logger.info(f'Packaging complete: {bundle_path}')
-        return bundle_path
+        self.bundle_path = self.__build_tarball()
+
+        _logger.info('Extracting preview data.')
+        self.md['count'] = self.table.count()
+        preview_header, preview = self.__extract_preview_data()
+        self.md['preview_header'] = preview_header
+        self.md['preview'] = preview
+
+        _logger.info(f'Packaging complete: {self.bundle_path}')
+        return self.bundle_path
 
     def __export_table(self, tv: catalog.TableVersion) -> None:
         """
@@ -207,6 +226,96 @@ class TablePackager:
                 tf.add(src_file, arcname=f'media/{dest_name}')
         return bundle_path
 
+    def __extract_preview_data(self) -> tuple[dict[str, str], list[list[Any]]]:
+        """
+        Extract a preview of the table data for display in the UI.
+
+        In order to bound the size of the output data, all "unbounded" data types are resized:
+        - Strings are abbreviated as per Formatter.abbreviate()
+        - Arrays and JSON are shortened and formatted as strings
+        - Images are resized to thumbnail size as a base64-encoded webp
+        - Videos are replaced by their first frame and resized as above
+        - 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))
+        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))
+
+        preview_header = {col_name: str(col_type._type) for col_name, col_type in preview_cols.items()}
+        preview = [
+            [self.__encode_preview_data(val, col_type)]
+            for row in rows
+            for val, col_type in zip(row.values(), preview_cols.values())
+        ]
+
+        return preview_header, preview
+
+    def __encode_preview_data(self, val: Any, col_type: ts.ColumnType) -> Any:
+        if val is None:
+            return None
+
+        match col_type._type:
+            case ts.ColumnType.Type.STRING:
+                assert isinstance(val, str)
+                return Formatter.abbreviate(val)
+
+            case ts.ColumnType.Type.INT | ts.ColumnType.Type.FLOAT | ts.ColumnType.Type.BOOL:
+                return val
+
+            case ts.ColumnType.Type.TIMESTAMP | ts.ColumnType.Type.DATE:
+                return str(val)
+
+            case ts.ColumnType.Type.ARRAY:
+                assert isinstance(val, np.ndarray)
+                return Formatter.format_array(val)
+
+            case ts.ColumnType.Type.JSON:
+                # We need to escape the JSON string server-side for security reasons.
+                # Therefore we don't escape it here, in order to avoid double-escaping.
+                return Formatter.format_json(val, escape_strings=False)
+
+            case ts.ColumnType.Type.IMAGE:
+                # Rescale the image to minimize data transfer size
+                assert isinstance(val, PIL.Image.Image)
+                return self.__encode_image(val)
+
+            case ts.ColumnType.Type.VIDEO:
+                assert isinstance(val, str)
+                return self.__encode_video(val)
+
+            case ts.ColumnType.Type.AUDIO:
+                return None
+
+            case ts.ColumnType.Type.DOCUMENT:
+                assert isinstance(val, str)
+                return self.__encode_document(val)
+
+            case _:
+                raise AssertionError(f'Unrecognized column type: {col_type._type}')
+
+    def __encode_image(self, img: PIL.Image.Image) -> str:
+        # Heuristic for thumbnail sizing:
+        # Standardize on a width of 240 pixels (to most efficiently utilize the columnar display).
+        # But, if the aspect ratio is below 2:3, bound the height at 360 pixels (to avoid unboundedly tall thumbnails
+        #     in the case of highly oblong images).
+        if img.height > img.width * 1.5:
+            scaled_img = img.resize((img.width * 360 // img.height, 360))
+        else:
+            scaled_img = img.resize((240, img.height * 240 // img.width))
+        with io.BytesIO() as buffer:
+            scaled_img.save(buffer, 'webp')
+            return base64.b64encode(buffer.getvalue()).decode()
+
+    def __encode_video(self, video_path: str) -> Optional[str]:
+        thumb = Formatter.extract_first_video_frame(video_path)
+        return self.__encode_image(thumb) if thumb is not None else None
+
+    def __encode_document(self, doc_path: str) -> Optional[str]:
+        thumb = Formatter.make_document_thumbnail(doc_path)
+        return self.__encode_image(thumb) if thumb is not None else None
+
 
 class TableRestorer:
     """

pixeltable/utils/formatter.py

@@ -63,10 +63,10 @@ class Formatter:
         """
         Escapes special characters in `val`, and abbreviates `val` if its length exceeds `_STRING_MAX_LEN`.
         """
-        return cls.__escape(cls.__abbreviate(val, cls.__STRING_MAX_LEN))
+        return cls.__escape(cls.abbreviate(val))
 
     @classmethod
-    def __abbreviate(cls, val: str, max_len: int) -> str:
+    def abbreviate(cls, val: str, max_len: int = __STRING_MAX_LEN) -> str:
         if len(val) > max_len:
             edgeitems = (max_len - len(cls.__STRING_SEP)) // 2
             return f'{val[:edgeitems]}{cls.__STRING_SEP}{val[-edgeitems:]}'
@@ -94,41 +94,45 @@ class Formatter:
         )
 
     @classmethod
-    def format_json(cls, val: Any) -> str:
+    def format_json(cls, val: Any, escape_strings: bool = True) -> str:
         if isinstance(val, str):
             # JSON-like formatting will be applied to strings that appear nested within a list or dict
             # (quote the string; escape any quotes inside the string; shorter abbreviations).
             # However, if the string appears in top-level position (i.e., the entire JSON value is a
             # string), then we format it like an ordinary string.
-            return cls.format_string(val)
+            return cls.format_string(val) if escape_strings else cls.abbreviate(val)
         # In all other cases, dump the JSON struct recursively.
-        return cls.__format_json_rec(val)
+        return cls.__format_json_rec(val, escape_strings)
 
     @classmethod
-    def __format_json_rec(cls, val: Any) -> str:
+    def __format_json_rec(cls, val: Any, escape_strings: bool) -> str:
         if isinstance(val, str):
-            return cls.__escape(json.dumps(cls.__abbreviate(val, cls.__NESTED_STRING_MAX_LEN)))
+            formatted = json.dumps(cls.abbreviate(val, cls.__NESTED_STRING_MAX_LEN))
+            return cls.__escape(formatted) if escape_strings else formatted
         if isinstance(val, float):
             return cls.format_float(val)
         if isinstance(val, np.ndarray):
             return cls.format_array(val)
         if isinstance(val, list):
             if len(val) < cls.__LIST_THRESHOLD:
-                components = [cls.__format_json_rec(x) for x in val]
+                components = [cls.__format_json_rec(x, escape_strings) for x in val]
             else:
-                components = [cls.__format_json_rec(x) for x in val[: cls.__LIST_EDGEITEMS]]
+                components = [cls.__format_json_rec(x, escape_strings) for x in val[: cls.__LIST_EDGEITEMS]]
                 components.append('...')
-                components.extend(cls.__format_json_rec(x) for x in val[-cls.__LIST_EDGEITEMS :])
+                components.extend(cls.__format_json_rec(x, escape_strings) for x in val[-cls.__LIST_EDGEITEMS :])
             return '[' + ', '.join(components) + ']'
         if isinstance(val, dict):
-            kv_pairs = (f'{cls.__format_json_rec(k)}: {cls.__format_json_rec(v)}' for k, v in val.items())
+            kv_pairs = (
+                f'{cls.__format_json_rec(k, escape_strings)}: {cls.__format_json_rec(v, escape_strings)}'
+                for k, v in val.items()
+            )
             return '{' + ', '.join(kv_pairs) + '}'
 
         # Everything else
         try:
             return json.dumps(val)
         except TypeError:  # Not JSON serializable
-            return str(val)
+            return cls.__escape(str(val))
 
     def format_img(self, img: Image.Image) -> str:
         """
@@ -152,22 +156,19 @@ class Formatter:
             """
 
     def format_video(self, file_path: str) -> str:
-        thumb_tag = ''
         # Attempt to extract the first frame of the video to use as a thumbnail,
         # so that the notebook can be exported as HTML and viewed in contexts where
         # the video itself is not accessible.
         # TODO(aaron-siegel): If the video is backed by a concrete external URL,
         # should we link to that instead?
-        with av.open(file_path) as container:
-            try:
-                thumb = next(container.decode(video=0)).to_image()
-                assert isinstance(thumb, Image.Image)
-                with io.BytesIO() as buffer:
-                    thumb.save(buffer, 'jpeg')
-                    thumb_base64 = base64.b64encode(buffer.getvalue()).decode()
-                    thumb_tag = f'poster="data:image/jpeg;base64,{thumb_base64}"'
-            except Exception:
-                pass
+        thumb = self.extract_first_video_frame(file_path)
+        if thumb is None:
+            thumb_tag = ''
+        else:
+            with io.BytesIO() as buffer:
+                thumb.save(buffer, 'jpeg')
+                thumb_base64 = base64.b64encode(buffer.getvalue()).decode()
+                thumb_tag = f'poster="data:image/jpeg;base64,{thumb_base64}"'
         if self.__num_rows > 1:
             width = 320
         elif self.__num_cols > 1:
@@ -182,6 +183,16 @@ class Formatter:
         </div>
         """
 
+    @classmethod
+    def extract_first_video_frame(cls, file_path: str) -> Optional[Image.Image]:
+        with av.open(file_path) as container:
+            try:
+                img = next(container.decode(video=0)).to_image()
+                assert isinstance(img, Image.Image)
+                return img
+            except Exception:
+                return None
+
     def format_audio(self, file_path: str) -> str:
         return f"""
         <div class="pxt_audio">
@@ -191,29 +202,18 @@ class Formatter:
         </div>
         """
 
-    def format_document(self, file_path: str) -> str:
-        max_width = max_height = 320
+    def format_document(self, file_path: str, max_width: int = 320, max_height: int = 320) -> str:
         # by default, file path will be shown as a link
         inner_element = file_path
         inner_element = html.escape(inner_element)
-        # try generating a thumbnail for different types and use that if successful
-        if file_path.lower().endswith('.pdf'):
-            try:
-                import fitz  # type: ignore[import-untyped]
 
-                doc = fitz.open(file_path)
-                p = doc.get_page_pixmap(0)
-                while p.width > max_width or p.height > max_height:
-                    # shrink(1) will halve each dimension
-                    p.shrink(1)
-                data = p.tobytes(output='jpeg')
-                thumb_base64 = base64.b64encode(data).decode()
-                img_src = f'data:image/jpeg;base64,{thumb_base64}'
-                inner_element = f"""
-                    <img style="object-fit: contain; border: 1px solid black;" src="{img_src}" />
-                """
-            except Exception:
-                logging.warning(f'Failed to produce PDF thumbnail {file_path}. Make sure you have PyMuPDF installed.')
+        thumb = self.make_document_thumbnail(file_path, max_width, max_height)
+        if thumb is not None:
+            with io.BytesIO() as buffer:
+                thumb.save(buffer, 'webp')
+                thumb_base64 = base64.b64encode(buffer.getvalue()).decode()
+                thumb_tag = f'data:image/webp;base64,{thumb_base64}'
+            inner_element = f'<img style="object-fit: contain; border: 1px solid black;" src="{thumb_tag}" />'
 
         return f"""
         <div class="pxt_document" style="width:{max_width}px;">
@@ -223,6 +223,28 @@ class Formatter:
         </div>
         """
 
+    @classmethod
+    def make_document_thumbnail(
+        cls, file_path: str, max_width: int = 320, max_height: int = 320
+    ) -> Optional[Image.Image]:
+        """
+        Returns a thumbnail image of a document.
+        """
+        if file_path.lower().endswith('.pdf'):
+            try:
+                import fitz  # type: ignore[import-untyped]
+
+                doc = fitz.open(file_path)
+                pixmap = doc.get_page_pixmap(0)
+                while pixmap.width > max_width or pixmap.height > max_height:
+                    # shrink(1) will halve each dimension
+                    pixmap.shrink(1)
+                return pixmap.pil_image()
+            except Exception:
+                logging.warning(f'Failed to produce PDF thumbnail {file_path}. Make sure you have PyMuPDF installed.')
+
+        return None
+
     @classmethod
     def __create_source_tag(cls, http_address: str, file_path: str) -> str:
         src_url = get_file_uri(http_address, file_path)

pixeltable/utils/sample.py

@@ -0,0 +1,25 @@
+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)

{pixeltable-0.4.0rc1.dist-info → pixeltable-0.4.0rc3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pixeltable
-Version: 0.4.0rc1
+Version: 0.4.0rc3
 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
@@ -36,6 +36,7 @@ Requires-Dist: numpy (>=1.25)
 Requires-Dist: pandas (>=2.0,<3.0)
 Requires-Dist: pgvector (>=0.2.1)
 Requires-Dist: pillow (>=9.3.0)
+Requires-Dist: pillow-heif (>=0.15.0)
 Requires-Dist: pixeltable-pgserver (==0.3.1)
 Requires-Dist: psutil (>=5.9.5)
 Requires-Dist: psycopg[binary] (>=3.1.18)

{pixeltable-0.4.0rc1.dist-info → pixeltable-0.4.0rc3.dist-info}/RECORD

@@ -1,7 +1,7 @@
 pixeltable/__init__.py,sha256=-uXHuiXH98kAlCupUTPbkBh4ToZgxcYUOk7-c9hqCC8,1439
-pixeltable/__version__.py,sha256=8n05k74lzCnT5buUnz2d9g9yaA6gdHgtEi-A_Z20pO4,120
+pixeltable/__version__.py,sha256=l1gVhW4YjeNouBt4ti5Rxsd1ETc8Ls5nKRC8aMceXrM,120
 pixeltable/catalog/__init__.py,sha256=rQmjveID4bk6NI4Ql09lGsZ0K0HVE2l1yqKAveipHzc,558
-pixeltable/catalog/catalog.py,sha256=mMivgHA6ru3HGZAiEfiQyP2yJEI7-LFhkGg_PlFSewc,59246
+pixeltable/catalog/catalog.py,sha256=ExHm2bbAs_4-T6zq3ow2fQ5a9Zja6G3WzHoo3Qk0X40,59399
 pixeltable/catalog/column.py,sha256=v3SE6CQs_ZXTDkBeiJ7gu9sxBJD-6EbkjsB1TZvLSJI,11280
 pixeltable/catalog/dir.py,sha256=HFemOf67Nfw13EOpQsR_UgzP2L1w4LDfw2009DrSK0Y,2063
 pixeltable/catalog/globals.py,sha256=7fNUs67D18PZ1ZajcGz7KQOV6tuXEcYLSrePzgmDkRw,4036
@@ -9,16 +9,16 @@ pixeltable/catalog/insertable_table.py,sha256=tl3tvMVPzi9A_SRARV6660NQnt8w09hNbO
 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=iE4PAulsxnlCjmSOMxVFD2mXNEch7owty4tbS-IfOHk,68084
-pixeltable/catalog/table_version.py,sha256=FqVuZdAbp7Bo0dt7zQGHtwTOUmS2qAiEtHrMQSOxqek,66289
+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=jlLyu71LMmz33WnfnXX3jANRJOtGM6SDifppApI0yJI,13585
+pixeltable/catalog/view.py,sha256=1UN-k0Ck0IWyy2Ae5b520L08GAfQ17L4FaZtS2bQ-TA,14786
 pixeltable/config.py,sha256=gnRI4G9GE7mQJDcMcn8JsEzYk8oKVfHB-BwoLRWnRDo,3971
-pixeltable/dataframe.py,sha256=U8QRC74phOtSnVMI9uxTZR_GB_GKUQLe0_Md7e3Rf80,51851
-pixeltable/env.py,sha256=1t5ZBrLCLSPaz0nD_sCQvvfB6tjBppVI9Io4iJtdmNE,36024
+pixeltable/dataframe.py,sha256=L1lNcfHtxxRw_b8kgkmvcuYXQyniAzwLnI4ww79P5B4,60393
+pixeltable/env.py,sha256=BtBxkKgU3wi6_0rfor3k3Kx6SF660JH24sWprbBzB8w,36132
 pixeltable/exceptions.py,sha256=eI2f4oXqv678_fDOxeYOTPE_EF1CMX3pwCqtJvN8i7s,926
-pixeltable/exec/__init__.py,sha256=m4AF2kzFtsTRUevupgWs8er0oJQisCR9ROW7ZJlL3dw,509
+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
@@ -26,14 +26,14 @@ pixeltable/exec/data_row_batch.py,sha256=EAB15JRhXbWIe91x1J5N5lFiMXzjB8NGTFjZsBD
 pixeltable/exec/exec_context.py,sha256=jKeLStfkjwCKKAooC-7a7qZUnZU5O0_JQhanhVerV9c,984
 pixeltable/exec/exec_node.py,sha256=WIN1sBEBNS7TlBk5QpHsDUCZUUdcvs6Os_Bxq2HoWdo,4077
 pixeltable/exec/expr_eval/__init__.py,sha256=sQThSEByK_DLfB-_-18RFhpARx49cSXYEkpCDyi0vQI,61
-pixeltable/exec/expr_eval/evaluators.py,sha256=AJkuleKhEz_8W6F3dcjTbWyKfv4IzpTh3cvQ3--OkRg,16865
+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/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=SBnmFvt1Iit5Sx1-iUqvtwyHNBdZPHlikhlrUtFnncs,20098
+pixeltable/exec/sql_node.py,sha256=MM9wD0AMXh11qJayhMgieCJ36hww-ym1p7kfS2y9-wM,26540
 pixeltable/exprs/__init__.py,sha256=AxSMjKNavCT9F6vBaNR-nwX2iupAI5hbMb5hEj65Tfk,1096
 pixeltable/exprs/arithmetic_expr.py,sha256=sZPao0qdFWbrDx0eiAVxw1wGHJXZ5ZoCpQaScysBldE,7333
 pixeltable/exprs/array_slice.py,sha256=8Zv0E2RghdJi1Mbk0kKtOz2ccGQuXwLLb6R9v1jk7hA,2180
@@ -41,8 +41,8 @@ pixeltable/exprs/column_property_ref.py,sha256=28UcsrMIoXmMy6DAgJ0GP6Vw4sYwXcspV
 pixeltable/exprs/column_ref.py,sha256=LljuQiNYh6dQa_EV57ylPfNkQD-lfHBFTGDY8WlnnfY,14942
 pixeltable/exprs/comparison.py,sha256=lgaRx000ZaNH10A4hrtsi5XoZKE-CNEONGMi7jxJfcM,5133
 pixeltable/exprs/compound_predicate.py,sha256=vJVRVueAmaKnjiHCLWyh8wHgktzzK0DVqbOIQJgTjF8,3801
-pixeltable/exprs/data_row.py,sha256=y16wVbp_ISNmMUP18PtX5vFrc4zxCiV3f9jCoReKl6I,11396
-pixeltable/exprs/expr.py,sha256=JABPT8dGISMq_ciG67GPdX00nKTrJwQ1DQhxSRqkC4A,35602
+pixeltable/exprs/data_row.py,sha256=8p0eWunvg_2ZH-e7Gtz_woxBAoizaCV2QtN87Is8_ys,11484
+pixeltable/exprs/expr.py,sha256=YFPL81s9N1_Obl15zpk_yYUvKqVoBT8fpAXhkYqgHLs,36000
 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
@@ -52,7 +52,7 @@ pixeltable/exprs/inline_expr.py,sha256=XYVKKXZN9BtHN5qlvZna-mgdOlot6WcmPu5usRBYe
 pixeltable/exprs/is_null.py,sha256=NfA_485hfT69pWyY6u8BhykDUkz5k91AH93azGu6lCg,1087
 pixeltable/exprs/json_mapper.py,sha256=bJSB39sZgpN9KS0RReDnUhTCwg-4Y4cgXXaFNy3o3wU,7035
 pixeltable/exprs/json_path.py,sha256=sFuDjfz8_rlea4TKd68CS4pQTUiLDi68YwsgcQRHffI,7162
-pixeltable/exprs/literal.py,sha256=kLAqc05uRHTN1IGAmRUNRkglo9CYualF7zW_U7c2pwU,4846
+pixeltable/exprs/literal.py,sha256=OCJL_pw_WKqx3bXMEwL6yNaKVAKDtGRzSZUFwucRxZI,4860
 pixeltable/exprs/method_ref.py,sha256=NNhJTGo7luZLh8EJdFIZAax9LiiqqDCEK1AwPmHip0w,2642
 pixeltable/exprs/object_ref.py,sha256=idYFcT27jv0BjtJT3paL37xDrZZc35_3eCJyQOIqdZU,1999
 pixeltable/exprs/row_builder.py,sha256=LL8q_8AthTqH8duWqKwdaS8DlvUsC8chclzQRR4MJAE,21150
@@ -75,34 +75,34 @@ pixeltable/func/function_registry.py,sha256=7AQ1bdF2DJbTRn9xx6s5cC_VHtCBXGt_GyJJ
 pixeltable/func/globals.py,sha256=5Wo4GPxYgHRRk5nvV0h_lAthKSalxKvj5n1p-uMPR0U,1501
 pixeltable/func/query_template_function.py,sha256=25nvgadWgk16BU7X5uI7PvKfkoluJSuOqsAcY5AN-Yw,7915
 pixeltable/func/signature.py,sha256=0PI6xdhLgwy9-GMkzkm7GlsBnsNMiS9aoNI9LWXwvN0,13700
-pixeltable/func/tools.py,sha256=DIfkOEj9Bp797Ew014_4YJePoUW40fQ6mvbCeg0FBR0,5721
+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/functions/anthropic.py,sha256=oTiXMfy3MupBInGRFEVkp-pHu06gF1ezJjEM2ypyvXw,9323
+pixeltable/functions/anthropic.py,sha256=G2E0sH5vP933eZZxhz1tOByy5cg6N2XPvhSqIBzqufo,8782
 pixeltable/functions/audio.py,sha256=7bsm4igQEW7RYSrSevwqaUOqyEnvBbPbJ8c-VknDl1E,657
 pixeltable/functions/bedrock.py,sha256=lTCFHjYunF3minBGWcjXR90yJ8resFjXr4niyKhfxms,4217
 pixeltable/functions/date.py,sha256=WUwqyrOWB8A00cTNEd6Vd7anQZo40_-7EWhpfpI-P6c,5323
-pixeltable/functions/deepseek.py,sha256=KYIa-UJJUTOt9cCfmP6k6nM4MpKm1MBU8F-jWk3CycY,3827
-pixeltable/functions/fireworks.py,sha256=k0vUXxeeNYWfL_tdLgF9c-vOilr0g2tTeLkAL9SJ6ws,4972
-pixeltable/functions/gemini.py,sha256=zWPsvtq0mPFBC_-4N7NDuhTYZfAzRMmZa9S4GFjIpLg,8328
+pixeltable/functions/deepseek.py,sha256=IAo2e_DhkM0A5NrskxuPQUGYzIYAl4do_mdO1Qc3PeY,3338
+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/image.py,sha256=IKXljMma-uU88efptC3F4aywau7DYcD-Nqd3YpmRNRw,13971
 pixeltable/functions/json.py,sha256=d7-AvwytUQtQYF_JnWJkptT_Yq0NgMpWfVk-m3U6qTY,807
-pixeltable/functions/llama_cpp.py,sha256=uf7WSZIhKDa492snnQv5ojGVLNdBWvuw0Ou3Mch1c_I,3874
+pixeltable/functions/llama_cpp.py,sha256=1QB4vQ7J4Za1mL93bRIBXgokNtpzzYr_QU6KF27i9xo,3919
 pixeltable/functions/math.py,sha256=eZEFjXxNHDHjcCsOMhzfNbJthTsmtNxtSFV8AEeRIfM,4979
-pixeltable/functions/mistralai.py,sha256=yZge5T385RoiFGXEZ6OhwWHj0JnsZ8tN8Jb3VkfDmXc,6274
-pixeltable/functions/ollama.py,sha256=AmkP532HwWeTyWkTnHm_hIk0CFjzV5MwCCPnM9Kb7KM,4231
-pixeltable/functions/openai.py,sha256=aDh1L2mBbSlrM8c1Rbh2QsCnmBESItLqzZ-frdgb05k,29259
+pixeltable/functions/mistralai.py,sha256=PTXQegC2LO5Pw0zXBO_SVV7I2c5qBvqVVgfm_mK1ir0,5845
+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/string.py,sha256=LdBNOna5PUSPmM5VlJ_qhmwzyFhumW0k6Dvx2rXSZtc,25356
 pixeltable/functions/timestamp.py,sha256=0zp4urJagCcNLfJE0ltTCft-J9qs2C716TmRngKYaa0,9171
-pixeltable/functions/together.py,sha256=ufg0RehEoQEqBy9EHSKY4N3ZNT5O__cwDS0Ll467eLk,10014
+pixeltable/functions/together.py,sha256=A8J19BXywyWQ6a2_n05-8uIG5jquOBGqPmW3mb-NrIc,8842
 pixeltable/functions/util.py,sha256=lVya13gcao8T34OGX7zy1cglQPNwaBbSBw57bVPyHNs,745
 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=IBeT0bJ4VxfNcmPkjyplZMYyM3bZ9PPFOPIq25GiZE4,32062
+pixeltable/globals.py,sha256=cM3uKvDJWIs1I9A8rzXaZsKCDSaLBh7KVb1lakQqGzk,32377
 pixeltable/index/__init__.py,sha256=97aFuxiP_oz1ldn5iq8IWApkOV8XG6ZIBW5-9rkS0vM,122
 pixeltable/index/base.py,sha256=200s7v3Zy810bRlbSAYzxxaEjVssl6r8esTHiSvWRwQ,1704
 pixeltable/index/btree.py,sha256=8B06D67ay0DFUtEBC5q4bLjxMq7ILpKyyoLAiSaamzA,2503
@@ -125,7 +125,7 @@ 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=gr9PE_iZ1Cze8h2Pj03c1ItShqUq7IxCwVnkV18DZbE,2607
+pixeltable/metadata/__init__.py,sha256=fZ3HaB-L8lHJdrtmKqfIJJlvYG0CqzzxKFRrh9IWoa0,3154
 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
@@ -151,13 +151,14 @@ pixeltable/metadata/converters/convert_32.py,sha256=YENfuQ_mqhnZWrvHJffaGEqd1IwS
 pixeltable/metadata/converters/convert_33.py,sha256=ZZV3FTyyouBM1eNymXxfHV-Oqmgu2s0KNP6AG3zc5eM,574
 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/util.py,sha256=95pfg9amEOmhho32PIbNYnqagVIN9adIcLXxB6zSYDY,7527
-pixeltable/metadata/notes.py,sha256=mCRyBvHNMsQ1fmORGcVIuiPwy7ya7vqbF8jmiSaCMNg,1347
-pixeltable/metadata/schema.py,sha256=JSbZEwEzFFAV8__i5VtrkfhxgedlFI76rxk3_esNMF8,11382
-pixeltable/plan.py,sha256=0xEHwqIUsgmSlLbYfmkyHuQNiD6lOumvx4UiNIk-KUw,43639
+pixeltable/metadata/notes.py,sha256=7HKk4yQLh8zaTVgv4pGXvpn3PoF5x7OFO8YSog5TLHI,1408
+pixeltable/metadata/schema.py,sha256=5r-loBIyYiZdkbszrLdn0XktOgDh96UVp_GScmm-14g,11489
+pixeltable/plan.py,sha256=bL6B32J63P1QtGjLhwwcfbh2SrSEeuyb1Zh6txdKQ6Q,52136
 pixeltable/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 pixeltable/share/__init__.py,sha256=AtR4nS6YkfkFRkXA-zZXFTK5pSQjHry8MnxdVLUk5SA,68
-pixeltable/share/packager.py,sha256=SyM-mzKoHLhgahFmJPkPSbBd1553nqs5FZnVTn4WNsg,27689
+pixeltable/share/packager.py,sha256=udwsorQDWX0N9eHcPHlT5xyAeUxGvT7DBnsFuq8PALM,32149
 pixeltable/share/publish.py,sha256=U6PzOUYiZaPu-sVNjh2nN8qzY2-uMsYeTwQCCuGk7Jg,6537
 pixeltable/store.py,sha256=2tfMgJNVVbGShgAKghGAIFMZ-ViJHE7N9kWd2ugbi2A,24691
 pixeltable/type_system.py,sha256=DSrof2NgKhBzvt7pbDNrGlZ3rkkDJ7MQsQ9rqk9N9pA,53988
@@ -172,16 +173,17 @@ pixeltable/utils/description_helper.py,sha256=acibNm36wkZG7h6k8gjcypTD_PVV2SL7Yg
 pixeltable/utils/documents.py,sha256=x3UHU7eykibyA3eVkSrCK1CQoaid228vp96WUEESssU,3105
 pixeltable/utils/exception_handler.py,sha256=yrTAtUJEOhldps_k6aRLEf5yQ8gYGhl9c6ewYNC4Qfc,2476
 pixeltable/utils/filecache.py,sha256=8RZZiEkD4awZpR-mn7OhoZPc6_JlPUNSBnMU8BcEAv4,10864
-pixeltable/utils/formatter.py,sha256=_pYQOhBh2dZBeCTUKuWaIzm7JRWeMepMZwSd5KTv-tw,9220
+pixeltable/utils/formatter.py,sha256=tbMxE9rBw6wdKUnJhNZ8h9uAF8dZKcihQ2KesqAag9A,10096
 pixeltable/utils/http_server.py,sha256=B5iQ1s_VuwsVC7pUm1joGjLZqaluV8_RfFiU8V1FuG8,2453
 pixeltable/utils/iceberg.py,sha256=L_s9G9NMIGMQdRHtNkks6ntTVW4DKKAw97R9gRmtw5s,553
 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-0.4.0rc1.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-pixeltable-0.4.0rc1.dist-info/METADATA,sha256=uD_AiJ5nJXk4YoVxcjAulmt9qsJGRruGMZfjdX_264Y,20542
-pixeltable-0.4.0rc1.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-pixeltable-0.4.0rc1.dist-info/entry_points.txt,sha256=ToOd-pRgG7AitEBgYoBCRRB4-KVDQ0pj_9T4a1LgwA4,97
-pixeltable-0.4.0rc1.dist-info/RECORD,,
+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,,