datachain 0.3.13__tar.gz → 0.3.15__tar.gz

{datachain-0.3.13/src/datachain.egg-info → datachain-0.3.15}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: datachain
-Version: 0.3.13
+Version: 0.3.15
 Summary: Wrangle unstructured AI data at scale
 Author-email: Dmitry Petrov <support@dvc.org>
 License: Apache-2.0

{datachain-0.3.13 → datachain-0.3.15}/examples/computer_vision/iptc_exif_xmp_lib.py

@@ -1,4 +1,10 @@
-# pip install defusedxml
+"""
+To install the required dependencies:
+
+  pip install datachain[examples]
+
+"""
+
 import json
 
 from PIL import (

{datachain-0.3.13 → datachain-0.3.15}/examples/computer_vision/llava2_image_desc_lib.py

@@ -1,4 +1,10 @@
-# pip install accelerate torch
+"""
+To install the required dependencies:
+
+  pip install datachain[examples]
+
+"""
+
 import torch
 from transformers import (
     AutoProcessor,

{datachain-0.3.13 → datachain-0.3.15}/examples/get_started/json-csv-reader.py

@@ -1,5 +1,3 @@
-# pip install datamodel-code-generator jmespath
-
 from typing import Optional
 
 from pydantic import BaseModel

{datachain-0.3.13 → datachain-0.3.15}/examples/get_started/torch-loader.py

@@ -1,4 +1,9 @@
-# pip install Pillow torchvision
+"""
+To install the required dependencies:
+
+  pip install datachain[torch]
+
+"""
 
 import os
 from posixpath import basename

{datachain-0.3.13 → datachain-0.3.15}/examples/get_started/udfs/stateful.py

@@ -1,7 +1,7 @@
 """
-To install dependencies:
+To install the required dependencies:
 
-  pip install open_clip_torch
+  pip install datachain[examples]
 
 """
 

{datachain-0.3.13 → datachain-0.3.15}/noxfile.py

@@ -40,6 +40,7 @@ def tests(session: nox.Session) -> None:
         "--cov-report=xml",
         "--durations=10",
         "--numprocesses=logical",
+        "--dist=loadgroup",
         *session.posargs,
         env={"COVERAGE_FILE": f".coverage.{session.python}"},
     )

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/asyn.py

@@ -1,14 +1,8 @@
 import asyncio
-from collections.abc import Awaitable, Coroutine, Iterable
+from collections.abc import AsyncIterable, Awaitable, Coroutine, Iterable, Iterator
 from concurrent.futures import ThreadPoolExecutor
 from heapq import heappop, heappush
-from typing import (
-    Any,
-    Callable,
-    Generic,
-    Optional,
-    TypeVar,
-)
+from typing import Any, Callable, Generic, Optional, TypeVar
 
 from fsspec.asyn import get_loop
 
@@ -16,6 +10,7 @@ ASYNC_WORKERS = 20
 
 InputT = TypeVar("InputT", contravariant=True)  # noqa: PLC0105
 ResultT = TypeVar("ResultT", covariant=True)  # noqa: PLC0105
+T = TypeVar("T")
 
 
 class AsyncMapper(Generic[InputT, ResultT]):
@@ -226,7 +221,7 @@ class OrderedMapper(AsyncMapper[InputT, ResultT]):
         self._push_result(self._next_yield, None)
 
 
-def iter_over_async(ait, loop):
+def iter_over_async(ait: AsyncIterable[T], loop) -> Iterator[T]:
     """Wrap an asynchronous iterator into a synchronous one"""
     ait = ait.__aiter__()
 

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/catalog/catalog.py

@@ -621,10 +621,6 @@ class Catalog:
                 code_ast.body[-1:] = new_expressions
         return code_ast
 
-    def parse_url(self, uri: str, **config: Any) -> tuple[Client, str]:
-        config = config or self.client_config
-        return Client.parse_url(uri, self.cache, **config)
-
     def get_client(self, uri: StorageURI, **config: Any) -> Client:
         """
         Return the client corresponding to the given source `uri`.
@@ -651,17 +647,16 @@ class Catalog:
         partial_path: Optional[str]
 
         client_config = client_config or self.client_config
-        client, path = self.parse_url(source, **client_config)
+        uri, path = Client.parse_url(source)
+        client = Client.get_client(source, self.cache, **client_config)
         stem = os.path.basename(os.path.normpath(path))
         prefix = (
             posixpath.dirname(path)
             if glob.has_magic(stem) or client.fs.isfile(source)
             else path
         )
-        storage_dataset_name = Storage.dataset_name(
-            client.uri, posixpath.join(prefix, "")
-        )
-        source_metastore = self.metastore.clone(client.uri)
+        storage_dataset_name = Storage.dataset_name(uri, posixpath.join(prefix, ""))
+        source_metastore = self.metastore.clone(uri)
 
         columns = [
             Column("path", String),
@@ -675,15 +670,13 @@ class Catalog:
         ]
 
         if skip_indexing:
-            source_metastore.create_storage_if_not_registered(client.uri)
-            storage = source_metastore.get_storage(client.uri)
-            source_metastore.init_partial_id(client.uri)
-            partial_id = source_metastore.get_next_partial_id(client.uri)
+            source_metastore.create_storage_if_not_registered(uri)
+            storage = source_metastore.get_storage(uri)
+            source_metastore.init_partial_id(uri)
+            partial_id = source_metastore.get_next_partial_id(uri)
 
-            source_metastore = self.metastore.clone(
-                uri=client.uri, partial_id=partial_id
-            )
-            source_metastore.init(client.uri)
+            source_metastore = self.metastore.clone(uri=uri, partial_id=partial_id)
+            source_metastore.init(uri)
 
             source_warehouse = self.warehouse.clone()
             dataset = self.create_dataset(
@@ -701,20 +694,16 @@ class Catalog:
             in_progress,
             partial_id,
             partial_path,
-        ) = source_metastore.register_storage_for_indexing(
-            client.uri, force_update, prefix
-        )
+        ) = source_metastore.register_storage_for_indexing(uri, force_update, prefix)
         if in_progress:
             raise PendingIndexingError(f"Pending indexing operation: uri={storage.uri}")
 
         if not need_index:
             assert partial_id is not None
             assert partial_path is not None
-            source_metastore = self.metastore.clone(
-                uri=client.uri, partial_id=partial_id
-            )
+            source_metastore = self.metastore.clone(uri=uri, partial_id=partial_id)
             source_warehouse = self.warehouse.clone()
-            dataset = self.get_dataset(Storage.dataset_name(client.uri, partial_path))
+            dataset = self.get_dataset(Storage.dataset_name(uri, partial_path))
             lst = Listing(storage, source_metastore, source_warehouse, client, dataset)
             logger.debug(
                 "Using cached listing %s. Valid till: %s",
@@ -731,11 +720,11 @@ class Catalog:
 
             return lst, path
 
-        source_metastore.init_partial_id(client.uri)
-        partial_id = source_metastore.get_next_partial_id(client.uri)
+        source_metastore.init_partial_id(uri)
+        partial_id = source_metastore.get_next_partial_id(uri)
 
-        source_metastore.init(client.uri)
-        source_metastore = self.metastore.clone(uri=client.uri, partial_id=partial_id)
+        source_metastore.init(uri)
+        source_metastore = self.metastore.clone(uri=uri, partial_id=partial_id)
 
         source_warehouse = self.warehouse.clone()
 
@@ -1370,7 +1359,7 @@ class Catalog:
 
     def signed_url(self, source: str, path: str, client_config=None) -> str:
         client_config = client_config or self.client_config
-        client, _ = self.parse_url(source, **client_config)
+        client = Client.get_client(source, self.cache, **client_config)
         return client.url(path)
 
     def export_dataset_table(
@@ -1390,12 +1379,12 @@ class Catalog:
         dataset = self.get_dataset(name)
         return self.warehouse.dataset_table_export_file_names(dataset, version)
 
-    def dataset_stats(self, name: str, version: int) -> DatasetStats:
+    def dataset_stats(self, name: str, version: Optional[int]) -> DatasetStats:
         """
         Returns tuple with dataset stats: total number of rows and total dataset size.
         """
         dataset = self.get_dataset(name)
-        dataset_version = dataset.get_version(version)
+        dataset_version = dataset.get_version(version or dataset.latest_version)
         return DatasetStats(
             num_objects=dataset_version.num_objects,
             size=dataset_version.size,

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/client/azure.py

@@ -4,7 +4,6 @@ from adlfs import AzureBlobFileSystem
 from tqdm import tqdm
 
 from datachain.lib.file import File
-from datachain.node import Entry
 
 from .fsspec import DELIMITER, Client, ResultQueue
 
@@ -14,17 +13,6 @@ class AzureClient(Client):
     PREFIX = "az://"
     protocol = "az"
 
-    def convert_info(self, v: dict[str, Any], path: str) -> Entry:
-        version_id = v.get("version_id")
-        return Entry.from_file(
-            path=path,
-            etag=v.get("etag", "").strip('"'),
-            version=version_id or "",
-            is_latest=version_id is None or bool(v.get("is_current_version")),
-            last_modified=v["last_modified"],
-            size=v.get("size", ""),
-        )
-
     def info_to_file(self, v: dict[str, Any], path: str) -> File:
         version_id = v.get("version_id")
         return File(
@@ -57,7 +45,7 @@ class AzureClient(Client):
                                 continue
                             info = (await self.fs._details([b]))[0]
                             entries.append(
-                                self.convert_info(info, self.rel_path(info["name"]))
+                                self.info_to_file(info, self.rel_path(info["name"]))
                             )
                         if entries:
                             await result_queue.put(entries)

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/client/fsspec.py

@@ -29,7 +29,7 @@ from tqdm import tqdm
 from datachain.cache import DataChainCache, UniqueId
 from datachain.client.fileslice import FileSlice, FileWrapper
 from datachain.error import ClientError as DataChainClientError
-from datachain.node import Entry
+from datachain.lib.file import File
 from datachain.nodes_fetcher import NodesFetcher
 from datachain.nodes_thread_pool import NodeChunk
 from datachain.storage import StorageURI
@@ -45,7 +45,7 @@ DELIMITER = "/"  # Path delimiter.
 
 DATA_SOURCE_URI_PATTERN = re.compile(r"^[\w]+:\/\/.*$")
 
-ResultQueue = asyncio.Queue[Optional[Sequence[Entry]]]
+ResultQueue = asyncio.Queue[Optional[Sequence[File]]]
 
 
 def _is_win_local_path(uri: str) -> bool:
@@ -116,15 +116,16 @@ class Client(ABC):
         return DATA_SOURCE_URI_PATTERN.match(name) is not None
 
     @staticmethod
-    def parse_url(
-        source: str,
-        cache: DataChainCache,
-        **kwargs,
-    ) -> tuple["Client", str]:
+    def parse_url(source: str) -> tuple[StorageURI, str]:
+        cls = Client.get_implementation(source)
+        storage_name, rel_path = cls.split_url(source)
+        return cls.get_uri(storage_name), rel_path
+
+    @staticmethod
+    def get_client(source: str, cache: DataChainCache, **kwargs) -> "Client":
         cls = Client.get_implementation(source)
-        storage_url, rel_path = cls.split_url(source)
-        client = cls.from_name(storage_url, cache, kwargs)
-        return client, rel_path
+        storage_url, _ = cls.split_url(source)
+        return cls.from_name(storage_url, cache, kwargs)
 
     @classmethod
     def create_fs(cls, **kwargs) -> "AbstractFileSystem":
@@ -188,7 +189,7 @@ class Client(ABC):
 
     async def get_current_etag(self, uid: UniqueId) -> str:
         info = await self.fs._info(self.get_full_path(uid.path))
-        return self.convert_info(info, "").etag
+        return self.info_to_file(info, "").etag
 
     async def get_size(self, path: str) -> int:
         return await self.fs._size(path)
@@ -198,7 +199,7 @@ class Client(ABC):
 
     async def scandir(
         self, start_prefix: str, method: str = "default"
-    ) -> AsyncIterator[Sequence[Entry]]:
+    ) -> AsyncIterator[Sequence[File]]:
         try:
             impl = getattr(self, f"_fetch_{method}")
         except AttributeError:
@@ -264,7 +265,7 @@ class Client(ABC):
     ) -> None:
         await self._fetch_nested(start_prefix, result_queue)
 
-    async def _fetch_dir(self, prefix, pbar, result_queue) -> set[str]:
+    async def _fetch_dir(self, prefix, pbar, result_queue: ResultQueue) -> set[str]:
         path = f"{self.name}/{prefix}"
         infos = await self.ls_dir(path)
         files = []
@@ -277,7 +278,7 @@ class Client(ABC):
             if info["type"] == "directory":
                 subdirs.add(subprefix)
             else:
-                files.append(self.convert_info(info, subprefix))
+                files.append(self.info_to_file(info, subprefix))
         if files:
             await result_queue.put(files)
         found_count = len(subdirs) + len(files)
@@ -303,7 +304,7 @@ class Client(ABC):
         return f"{self.PREFIX}{self.name}/{rel_path}"
 
     @abstractmethod
-    def convert_info(self, v: dict[str, Any], parent: str) -> Entry: ...
+    def info_to_file(self, v: dict[str, Any], parent: str) -> File: ...
 
     def fetch_nodes(
         self,

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/client/gcs.py

@@ -10,7 +10,6 @@ from gcsfs import GCSFileSystem
 from tqdm import tqdm
 
 from datachain.lib.file import File
-from datachain.node import Entry
 
 from .fsspec import DELIMITER, Client, ResultQueue
 
@@ -108,19 +107,9 @@ class GCSClient(Client):
         finally:
             await page_queue.put(None)
 
-    def _entry_from_dict(self, d: dict[str, Any]) -> Entry:
+    def _entry_from_dict(self, d: dict[str, Any]) -> File:
         info = self.fs._process_object(self.name, d)
-        return self.convert_info(info, self.rel_path(info["name"]))
-
-    def convert_info(self, v: dict[str, Any], path: str) -> Entry:
-        return Entry.from_file(
-            path=path,
-            etag=v.get("etag", ""),
-            version=v.get("generation", ""),
-            is_latest=not v.get("timeDeleted"),
-            last_modified=self.parse_timestamp(v["updated"]),
-            size=v.get("size", ""),
-        )
+        return self.info_to_file(info, self.rel_path(info["name"]))
 
     def info_to_file(self, v: dict[str, Any], path: str) -> File:
         return File(

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/client/hf.py

@@ -5,7 +5,6 @@ from typing import Any, cast
 from huggingface_hub import HfFileSystem
 
 from datachain.lib.file import File
-from datachain.node import Entry
 
 from .fsspec import Client
 
@@ -22,15 +21,6 @@ class HfClient(Client):
 
         return cast(HfFileSystem, super().create_fs(**kwargs))
 
-    def convert_info(self, v: dict[str, Any], path: str) -> Entry:
-        return Entry.from_file(
-            path=path,
-            size=v["size"],
-            version=v["last_commit"].oid,
-            etag=v.get("blob_id", ""),
-            last_modified=v["last_commit"].date,
-        )
-
     def info_to_file(self, v: dict[str, Any], path: str) -> File:
         return File(
             path=path,

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/client/local.py

@@ -7,8 +7,8 @@ from urllib.parse import urlparse
 
 from fsspec.implementations.local import LocalFileSystem
 
+from datachain.cache import UniqueId
 from datachain.lib.file import File
-from datachain.node import Entry
 from datachain.storage import StorageURI
 
 from .fsspec import Client
@@ -114,9 +114,9 @@ class FileClient(Client):
             use_symlinks=use_symlinks,
         )
 
-    async def get_current_etag(self, uid) -> str:
+    async def get_current_etag(self, uid: UniqueId) -> str:
         info = self.fs.info(self.get_full_path(uid.path))
-        return self.convert_info(info, "").etag
+        return self.info_to_file(info, "").etag
 
     async def get_size(self, path: str) -> int:
         return self.fs.size(path)
@@ -136,15 +136,6 @@ class FileClient(Client):
             full_path += "/"
         return full_path
 
-    def convert_info(self, v: dict[str, Any], path: str) -> Entry:
-        return Entry.from_file(
-            path=path,
-            etag=v["mtime"].hex(),
-            is_latest=True,
-            last_modified=datetime.fromtimestamp(v["mtime"], timezone.utc),
-            size=v.get("size", ""),
-        )
-
     def info_to_file(self, v: dict[str, Any], path: str) -> File:
         return File(
             source=self.uri,

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/client/s3.py

@@ -1,12 +1,11 @@
 import asyncio
-from typing import Any, cast
+from typing import Any, Optional, cast
 
 from botocore.exceptions import NoCredentialsError
 from s3fs import S3FileSystem
 from tqdm import tqdm
 
 from datachain.lib.file import File
-from datachain.node import Entry
 
 from .fsspec import DELIMITER, Client, ResultQueue
 
@@ -111,8 +110,9 @@ class ClientS3(Client):
     ) -> None:
         await self._fetch_flat(start_prefix, result_queue)
 
-    def _entry_from_boto(self, v, bucket, versions=False):
-        return Entry.from_file(
+    def _entry_from_boto(self, v, bucket, versions=False) -> File:
+        return File(
+            source=self.uri,
             path=v["Key"],
             etag=v.get("ETag", "").strip('"'),
             version=ClientS3.clean_s3_version(v.get("VersionId", "")),
@@ -125,8 +125,8 @@ class ClientS3(Client):
         self,
         prefix,
         pbar,
-        result_queue,
-    ):
+        result_queue: ResultQueue,
+    ) -> set[str]:
         if prefix:
             prefix = prefix.lstrip(DELIMITER) + DELIMITER
         files = []
@@ -141,7 +141,7 @@ class ClientS3(Client):
             if info["type"] == "directory":
                 subdirs.add(subprefix)
             else:
-                files.append(self.convert_info(info, subprefix))
+                files.append(self.info_to_file(info, subprefix))
                 pbar.update()
             found = True
         if not found:
@@ -152,18 +152,8 @@ class ClientS3(Client):
         return subdirs
 
     @staticmethod
-    def clean_s3_version(ver):
-        return ver if ver != "null" else ""
-
-    def convert_info(self, v: dict[str, Any], path: str) -> Entry:
-        return Entry.from_file(
-            path=path,
-            etag=v.get("ETag", "").strip('"'),
-            version=ClientS3.clean_s3_version(v.get("VersionId", "")),
-            is_latest=v.get("IsLatest", True),
-            last_modified=v.get("LastModified", ""),
-            size=v["size"],
-        )
+    def clean_s3_version(ver: Optional[str]) -> str:
+        return ver if (ver is not None and ver != "null") else ""
 
     def info_to_file(self, v: dict[str, Any], path: str) -> File:
         return File(

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/data_storage/sqlite.py

@@ -43,6 +43,8 @@ if TYPE_CHECKING:
     from sqlalchemy.sql.elements import ColumnElement
     from sqlalchemy.types import TypeEngine
 
+    from datachain.lib.file import File
+
 
 logger = logging.getLogger("datachain")
 
@@ -58,6 +60,10 @@ quote_schema = sqlite_dialect.identifier_preparer.quote_schema
 quote = sqlite_dialect.identifier_preparer.quote
 
 
+def _get_in_memory_uri():
+    return "file::memory:?cache=shared"
+
+
 def get_retry_sleep_sec(retry_count: int) -> int:
     return RETRY_START_SEC * (RETRY_FACTOR**retry_count)
 
@@ -119,7 +125,7 @@ class SQLiteDatabaseEngine(DatabaseEngine):
             if db_file == ":memory:":
                 # Enable multithreaded usage of the same in-memory db
                 db = sqlite3.connect(
-                    "file::memory:?cache=shared", uri=True, detect_types=DETECT_TYPES
+                    _get_in_memory_uri(), uri=True, detect_types=DETECT_TYPES
                 )
             else:
                 db = sqlite3.connect(
@@ -704,6 +710,9 @@ class SQLiteWarehouse(AbstractWarehouse):
 
         self.db.execute(insert_query)
 
+    def prepare_entries(self, entries: "Iterable[File]") -> Iterable[dict[str, Any]]:
+        return (e.model_dump() for e in entries)
+
     def insert_rows(self, table: Table, rows: Iterable[dict[str, Any]]) -> None:
         rows = list(rows)
         if not rows:

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/data_storage/warehouse.py

@@ -20,7 +20,7 @@ from datachain.client import Client
 from datachain.data_storage.schema import convert_rows_custom_column_types
 from datachain.data_storage.serializer import Serializable
 from datachain.dataset import DatasetRecord
-from datachain.node import DirType, DirTypeGroup, Entry, Node, NodeWithPath, get_path
+from datachain.node import DirType, DirTypeGroup, Node, NodeWithPath, get_path
 from datachain.sql.functions import path as pathfunc
 from datachain.sql.types import Int, SQLType
 from datachain.storage import StorageURI
@@ -34,6 +34,7 @@ if TYPE_CHECKING:
     from datachain.data_storage import AbstractIDGenerator, schema
     from datachain.data_storage.db_engine import DatabaseEngine
     from datachain.data_storage.schema import DataTable
+    from datachain.lib.file import File
 
 try:
     import numpy as np
@@ -401,25 +402,18 @@ class AbstractWarehouse(ABC, Serializable):
         expressions: tuple[_ColumnsClauseArgument[Any], ...] = (
             sa.func.count(table.c.sys__id),
         )
-        if "file__size" in table.columns:
-            expressions = (*expressions, sa.func.sum(table.c.file__size))
-        elif "size" in table.columns:
-            expressions = (*expressions, sa.func.sum(table.c.size))
+        size_columns = [
+            c for c in table.columns if c.name == "size" or c.name.endswith("__size")
+        ]
+        if size_columns:
+            expressions = (*expressions, sa.func.sum(sum(size_columns)))
         query = select(*expressions)
         ((nrows, *rest),) = self.db.execute(query)
-        return nrows, rest[0] if rest else None
-
-    def prepare_entries(
-        self, uri: str, entries: Iterable[Entry]
-    ) -> list[dict[str, Any]]:
-        """
-        Prepares bucket listing entry (row) for inserting into database
-        """
-
-        def _prepare_entry(entry: Entry):
-            return attrs.asdict(entry) | {"source": uri}
+        return nrows, rest[0] if rest else 0
 
-        return [_prepare_entry(e) for e in entries]
+    @abstractmethod
+    def prepare_entries(self, entries: "Iterable[File]") -> Iterable[dict[str, Any]]:
+        """Convert File entries so they can be passed on to `insert_rows()`"""
 
     @abstractmethod
     def insert_rows(self, table: Table, rows: Iterable[dict[str, Any]]) -> None:

{datachain-0.3.13 → datachain-0.3.15}/src/datachain/dataset.py

@@ -112,7 +112,7 @@ class DatasetDependency:
 
         if is_listing_dataset(dataset_name):
             dependency_type = DatasetDependencyType.STORAGE  # type: ignore[arg-type]
-            dependency_name = listing_uri_from_name(dataset_name)
+            dependency_name, _ = Client.parse_url(listing_uri_from_name(dataset_name))
 
         return cls(
             id,