idc-index-data 23.2.4__tar.gz → 23.2.5__tar.gz

{idc_index_data-23.2.4 → idc_index_data-23.2.5}/.gitignore

@@ -28,21 +28,14 @@ MANIFEST
 
 # Generated data files (created during build)
 idc_index.csv.zip
-idc_index.parquet
-prior_versions_index.parquet
+*.parquet
 
 # Generated schema files (created during build)
 *_schema.json
 
-# Generated SQL files at project root (created during build)
-# Note: The source SQL files in scripts/sql and assets are tracked in git
-idc_index.sql
-prior_versions_index.sql
-collections_index.sql
-analysis_results_index.sql
-clinical_index.sql
-sm_index.sql
-sm_instance_index.sql
+# Generated SQL files (created during build)
+# Note: tracked SQL files already in the repo remain unaffected
+*.sql
 
 # Auto-generated version file (created by hatch-vcs)
 src/idc_index_data/_version.py

{idc_index_data-23.2.4 → idc_index_data-23.2.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: idc-index-data
-Version: 23.2.4
+Version: 23.2.5
 Summary: ImagingDataCommons index to query and download data.
 Project-URL: Homepage, https://github.com/ImagingDataCommons/idc-index-data
 Project-URL: Bug Tracker, https://github.com/ImagingDataCommons/idc-index-data/issues

{idc_index_data-23.2.4 → idc_index_data-23.2.5}/hatch_build.py

@@ -6,6 +6,7 @@ import os
 import subprocess
 import sys
 from pathlib import Path
+from typing import ClassVar
 
 from hatchling.builders.hooks.plugin.interface import BuildHookInterface
 
@@ -15,6 +16,24 @@ class IDCBuildHook(BuildHookInterface):
 
     PLUGIN_NAME = "custom"
 
+    PARQUET_WHITELIST: ClassVar[set[str]] = {
+        "idc_index.parquet",
+        "prior_versions_index.parquet",
+    }
+
+    def _prune_unwhitelisted_parquet_files(self) -> None:
+        """Remove parquet files not explicitly whitelisted from the package dir."""
+        package_dir = Path(self.root) / "src" / "idc_index_data"
+        if not package_dir.exists():
+            return
+
+        for parquet_file in package_dir.glob("*.parquet"):
+            if parquet_file.name not in self.PARQUET_WHITELIST:
+                parquet_file.unlink()
+                self.app.display_info(
+                    f"Removed non-whitelisted parquet: {parquet_file.name}"
+                )
+
     def initialize(self, version: str, build_data: dict) -> None:  # noqa: ARG002
         """
         Generate data files before build.
@@ -31,6 +50,9 @@ class IDCBuildHook(BuildHookInterface):
                 - artifacts: List of additional files to include
                 - force_include: Dict mapping source -> dest paths
         """
+        # Remove parquet files that are optional to reduce package size
+        self._prune_unwhitelisted_parquet_files()
+
         # 1. Validate environment
         if not os.environ.get("GCP_PROJECT"):
             self.app.display_warning("Skipping data generation: GCP_PROJECT not set")

{idc_index_data-23.2.4 → idc_index_data-23.2.5}/pyproject.toml

@@ -77,6 +77,7 @@ version-file = "src/idc_index_data/_version.py"
 
 [tool.hatch.build.targets.wheel]
 packages = ["src/idc_index_data"]
+ignore-vcs = true
 
 [tool.hatch.build.hooks.custom]
 module = "hatch_build:IDCBuildHook"

{idc_index_data-23.2.4 → idc_index_data-23.2.5}/src/idc_index_data/_version.py

@@ -28,7 +28,7 @@ version_tuple: VERSION_TUPLE
 commit_id: COMMIT_ID
 __commit_id__: COMMIT_ID
 
-__version__ = version = '23.2.4'
-__version_tuple__ = version_tuple = (23, 2, 4)
+__version__ = version = '23.2.5'
+__version_tuple__ = version_tuple = (23, 2, 5)
 
 __commit_id__ = commit_id = None