classifyre-cli 0.7.3__tar.gz → 0.7.4__tar.gz

{classifyre_cli-0.7.3 → classifyre_cli-0.7.4}/.turbo/turbo-build.log

@@ -1,3 +1,3 @@
 $ uv sync
-Resolved 338 packages in 176ms
+Resolved 341 packages in 545ms
 Checked 52 packages in 1ms

{classifyre_cli-0.7.3 → classifyre_cli-0.7.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: classifyre-cli
-Version: 0.7.3
+Version: 0.7.4
 Summary: Classifyre CLI — scan and classify unstructured data sources
 License: MIT
 Keywords: data,ingestion,metadata,pii,secrets,unstructured

{classifyre_cli-0.7.3 → classifyre_cli-0.7.4}/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@classifyre/cli",
-  "version": "0.7.3",
+  "version": "0.7.4",
   "private": true,
   "scripts": {
     "build": "uv sync",

{classifyre_cli-0.7.3 → classifyre_cli-0.7.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "classifyre-cli"
-version = "0.7.3"
+version = "0.7.4"
 description = "Classifyre CLI — scan and classify unstructured data sources"
 readme = "README.md"
 # Upper bound is load-bearing: onnxruntime publishes no cp315 wheels (notably
@@ -25,7 +25,10 @@ dependencies = [
 [tool.uv.sources]
 # Local editable install for development — ignored when published to PyPI
 classifyre-schemas = { path = "../../packages/schemas", editable = true }
-en-core-web-sm = { url = "https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.8.0/en_core_web_sm-3.8.0-py3-none-any.whl" }
+# Explosion's en_core_web_sm 3.8.0 wheel, unchanged (same sha256), from our own
+# Hugging Face copy pinned to a commit: GitHub's release download answered 504
+# for hours and failed every CI run and image build that needed it.
+en-core-web-sm = { url = "https://huggingface.co/classifyre/en_core_web_sm/resolve/36818c1a75b5bac93b4f0df26b76da8d1149ce8c/en_core_web_sm-3.8.0-py3-none-any.whl" }
 # CPU-only torch on Linux. The default PyPI Linux wheels bundle the full NVIDIA
 # CUDA stack (~5-7 GB) which is dead weight on our CPU-only CLI jobs (no GPU
 # requested). Pull torch/torchvision from the PyTorch CPU index on Linux so the

{classifyre_cli-0.7.3 → classifyre_cli-0.7.4}/src/sources/trino/source.py

@@ -32,13 +32,15 @@ _SKIPPED_CATALOGS = {"system"}
 _SKIPPED_SCHEMAS = {"information_schema", "system"}
 
 # Trino connector name -> the table format its tables are stored in.
+# A catalog's connector names the format of every table in it. Anything else,
+# Hive's plain tables included, is ``native`` (as the Hive source reports them).
+# The ``lakehouse`` connector mixes formats per table, so it is ``native`` here
+# until its tables' own ``type`` property is read.
 _CONNECTOR_FORMATS = {
     "iceberg": "iceberg",
     "delta_lake": "delta",
     "delta-lake": "delta",
     "hudi": "hudi",
-    "hive": "hive",
-    "lakehouse": "lakehouse",
 }
 
 

{classifyre_cli-0.7.3 → classifyre_cli-0.7.4}/src/utils/lakehouse_storage.py

@@ -18,6 +18,7 @@ What each provider supports (see the lakehouse docs page for the tested matrix):
 
 from __future__ import annotations
 
+import os
 from pathlib import Path
 from urllib.parse import urlsplit
 
@@ -118,6 +119,11 @@ def azure_table_storage(
     object_store: dict[str, str] = {"azure_storage_account_name": account}
     if emulated:
         object_store |= {"azure_storage_use_emulator": "true", "allow_http": "true"}
+        # In emulator mode the Rust object_store readers (Lance, Hudi, and the
+        # Delta kernel inside DuckDB, which turns emulator mode on by itself for
+        # the Azurite account) take the endpoint only from this variable, and
+        # otherwise go to 127.0.0.1:10000 whatever account_url says.
+        os.environ["AZURITE_BLOB_STORAGE_URL"] = f"{parsed.scheme}://{parsed.netloc}"
     elif not blob_endpoint.endswith(".blob.core.windows.net"):
         object_store["azure_storage_endpoint"] = blob_endpoint  # OneLake, sovereign clouds
 

{classifyre_cli-0.7.3 → classifyre_cli-0.7.4}/tests/test_lakehouse_catalog_sources.py

@@ -3,7 +3,7 @@ Athena, BigQuery, Redshift.
 
 DuckLake runs against a real local lake (DuckDB's ``ducklake`` extension). The
 others talk to services, so their drivers are faked here and the lab
-(``showcases/lakehouse-lab``) runs them against real Polaris, Gravitino and Trino.
+(``development/lakehouse-lab``) runs them against real Polaris, Gravitino and Trino.
 """
 
 from __future__ import annotations

{classifyre_cli-0.7.3 → classifyre_cli-0.7.4}/tests/utils/test_lakehouse_tables.py

@@ -9,6 +9,7 @@ Paimon tables are written per test when their readers are installed (the
 
 from __future__ import annotations
 
+import os
 import shutil
 from pathlib import Path
 
@@ -301,3 +302,41 @@ def test_iceberg_scan_omits_version_when_unknown(monkeypatch: pytest.MonkeyPatch
 
 def test_iceberg_scan_pins_version_when_known(monkeypatch: pytest.MonkeyPatch) -> None:
     assert "version = 'v1'" in _iceberg_scan_sql("v1", monkeypatch)
+
+
+def test_azurite_endpoint_reaches_the_rust_readers(monkeypatch: pytest.MonkeyPatch) -> None:
+    # Emulator mode in object_store ignores account_url; without this the Lance
+    # and Delta readers went to 127.0.0.1:10000 from a pod that names Azurite.
+    from src.utils.lakehouse_storage import azure_table_storage
+
+    monkeypatch.delenv("AZURITE_BLOB_STORAGE_URL", raising=False)
+    storage = azure_table_storage(
+        account_url="http://lab-azurite.example.svc:10000/devstoreaccount1",
+        container="lake",
+        connection_string=None,
+        account_key="a2V5",
+        sas_token=None,
+        client_id=None,
+        client_secret=None,
+        tenant_id=None,
+    )
+    assert storage.object_store_options["azure_storage_use_emulator"] == "true"
+    assert os.environ["AZURITE_BLOB_STORAGE_URL"] == "http://lab-azurite.example.svc:10000"
+
+
+def test_real_azure_leaves_the_emulator_alone(monkeypatch: pytest.MonkeyPatch) -> None:
+    from src.utils.lakehouse_storage import azure_table_storage
+
+    monkeypatch.delenv("AZURITE_BLOB_STORAGE_URL", raising=False)
+    storage = azure_table_storage(
+        account_url="https://acme.blob.core.windows.net",
+        container="lake",
+        connection_string=None,
+        account_key="a2V5",
+        sas_token=None,
+        client_id=None,
+        client_secret=None,
+        tenant_id=None,
+    )
+    assert "azure_storage_use_emulator" not in storage.object_store_options
+    assert "AZURITE_BLOB_STORAGE_URL" not in os.environ