datachain 0.6.6__tar.gz → 0.6.7__tar.gz

{datachain-0.6.6 → datachain-0.6.7}/.github/workflows/tests-studio.yml

@@ -52,14 +52,11 @@ jobs:
           STUDIO_READ_ACCESS_TOKEN: ${{ secrets.ITERATIVE_STUDIO_READ_ACCESS_TOKEN }}
         run: |
           echo "DataChain branch: $BRANCH"
-          if [[ "$BRANCH" == "main" ]]
-          then
-              STUDIO_BRANCH=develop
-          elif git ls-remote --heads https://"$STUDIO_READ_ACCESS_TOKEN"@github.com/iterative/studio.git "$BRANCH" | grep -F "$BRANCH" 2>&1>/dev/null
+          if git ls-remote --heads https://"$STUDIO_READ_ACCESS_TOKEN"@github.com/iterative/studio.git "$BRANCH" | grep -F "$BRANCH" 2>&1>/dev/null
           then
               STUDIO_BRANCH="$BRANCH"
           else
-              STUDIO_BRANCH=develop
+              STUDIO_BRANCH=main
           fi
           echo "STUDIO_BRANCH=$STUDIO_BRANCH" >> $GITHUB_ENV
           echo "Studio branch: $STUDIO_BRANCH"

{datachain-0.6.6/src/datachain.egg-info → datachain-0.6.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: datachain
-Version: 0.6.6
+Version: 0.6.7
 Summary: Wrangle unstructured AI data at scale
 Author-email: Dmitry Petrov <support@dvc.org>
 License: Apache-2.0

{datachain-0.6.6 → datachain-0.6.7}/src/datachain/catalog/catalog.py

@@ -58,7 +58,7 @@ from datachain.listing import Listing
 from datachain.node import DirType, Node, NodeWithPath
 from datachain.nodes_thread_pool import NodesThreadPool
 from datachain.remote.studio import StudioClient
-from datachain.sql.types import DateTime, SQLType, String
+from datachain.sql.types import DateTime, SQLType
 from datachain.utils import (
     DataChainDir,
     batched,
@@ -196,11 +196,6 @@ class DatasetRowsFetcher(NodesThreadPool):
         for c in [c for c, t in self.schema.items() if t == DateTime]:
             df[c] = pd.to_datetime(df[c], unit="s")
 
-        # strings are represented as binaries in parquet export so need to
-        # decode it back to strings
-        for c in [c for c, t in self.schema.items() if t == String]:
-            df[c] = df[c].str.decode("utf-8")
-
     def do_task(self, urls):
         import lz4.frame
         import pandas as pd
@@ -1403,6 +1398,7 @@ class Catalog:
             query_script=remote_dataset_version.query_script,
             create_rows=True,
             columns=columns,
+            feature_schema=remote_dataset_version.feature_schema,
             validate_version=False,
         )
 

{datachain-0.6.6 → datachain-0.6.7}/src/datachain/data_storage/schema.py

@@ -145,6 +145,8 @@ class DirExpansion:
 
 
 class DataTable:
+    MAX_RANDOM = 2**63 - 1
+
     def __init__(
         self,
         name: str,
@@ -269,8 +271,8 @@ class DataTable:
     def delete(self):
         return self.apply_conditions(self.table.delete())
 
-    @staticmethod
-    def sys_columns():
+    @classmethod
+    def sys_columns(cls):
         return [
             sa.Column("sys__id", Int, primary_key=True),
             sa.Column(

{datachain-0.6.6 → datachain-0.6.7}/src/datachain/sql/types.py

@@ -440,6 +440,8 @@ class TypeReadConverter:
 
     def json(self, value):
         if isinstance(value, str):
+            if value == "":
+                return {}
             return orjson.loads(value)
         return value
 

{datachain-0.6.6 → datachain-0.6.7/src/datachain.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: datachain
-Version: 0.6.6
+Version: 0.6.7
 Summary: Wrangle unstructured AI data at scale
 Author-email: Dmitry Petrov <support@dvc.org>
 License: Apache-2.0

{datachain-0.6.6 → datachain-0.6.7}/tests/func/test_datasets.py

@@ -5,7 +5,7 @@ from unittest.mock import ANY
 import pytest
 import sqlalchemy as sa
 
-from datachain.data_storage.sqlite import SQLiteWarehouse
+from datachain.data_storage.schema import DataTable
 from datachain.dataset import DatasetDependencyType, DatasetStatus
 from datachain.error import (
     DatasetInvalidVersionError,
@@ -827,10 +827,7 @@ def test_row_random(cloud_test_catalog):
     # Random values are unique
     assert len(set(random_values)) == len(random_values)
 
-    if isinstance(catalog.warehouse, SQLiteWarehouse):
-        RAND_MAX = 2**63  # noqa: N806
-    else:
-        RAND_MAX = 2**64  # noqa: N806
+    RAND_MAX = DataTable.MAX_RANDOM  # noqa: N806
 
     # Values are drawn uniformly from range(2**63)
     assert 0 <= min(random_values) < 0.4 * RAND_MAX

{datachain-0.6.6 → datachain-0.6.7}/tests/func/test_pull.py

@@ -49,19 +49,17 @@ def dog_entries_parquet_lz4(dog_entries) -> bytes:
         """
         adapted = {}
         for k, v in row.items():
-            if isinstance(v, str):
-                adapted[k] = v.encode("utf-8")
-            elif isinstance(v, datetime):
+            if isinstance(v, datetime):
                 adapted[k] = v.timestamp()
             elif v is None:
-                adapted[k] = b""
+                adapted[k] = ""
             else:
                 adapted[k] = v
 
         adapted["sys__id"] = 1
         adapted["sys__rand"] = 1
-        adapted["file__location"] = b""
-        adapted["file__source"] = b"s3://dogs"
+        adapted["file__location"] = ""
+        adapted["file__source"] = "s3://dogs"
         return adapted
 
     dog_entries = [_adapt_row(e) for e in dog_entries]