ENH: update to IDC index v24

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: fedorov

assets/ann_group_index.sql

@@ -63,7 +63,7 @@ SELECT
   group_item.AnnotationGroupAlgorithmIdentificationSequence[SAFE_OFFSET(0)].AlgorithmName AS AlgorithmName
 
 FROM
-  `bigquery-public-data.idc_v23.dicom_all` AS ann
+  `bigquery-public-data.idc_v24.dicom_all` AS ann
 CROSS JOIN
   UNNEST(ann.AnnotationGroupSequence) AS group_item
 WHERE

assets/ann_index.sql

@@ -21,7 +21,7 @@ SELECT
   ReferencedSeriesSequence[SAFE_OFFSET(0)].SeriesInstanceUID AS referenced_SeriesInstanceUID
 
 FROM
-  `bigquery-public-data.idc_v23.dicom_all` AS ann
+  `bigquery-public-data.idc_v24.dicom_all` AS ann
 WHERE
   # Microscopy Bulk Simple Annotations SOP Class UID - more reliable than Modality = "ANN"
   SOPClassUID = "1.2.840.10008.5.1.4.1.1.91.1"

assets/clinical_index.sql

@@ -25,6 +25,6 @@ SELECT
   # values encountered in the column
   `values`
 FROM
-  `bigquery-public-data.idc_v23_clinical.column_metadata`
+  `bigquery-public-data.idc_v24_clinical.column_metadata`
 ORDER BY
   collection_id, table_name

assets/contrast_index.sql

@@ -11,7 +11,7 @@ WITH contrast_data AS (
     ARRAY_AGG(DISTINCT ContrastBolusAgent IGNORE NULLS ORDER BY ContrastBolusAgent) AS ContrastBolusAgent,
     ARRAY_AGG(DISTINCT ContrastBolusIngredient IGNORE NULLS ORDER BY ContrastBolusIngredient) AS ContrastBolusIngredient,
     ARRAY_AGG(DISTINCT ContrastBolusRoute IGNORE NULLS ORDER BY ContrastBolusRoute) AS ContrastBolusRoute
-  FROM `bigquery-public-data.idc_v23.dicom_all`
+  FROM `bigquery-public-data.idc_v24.dicom_all`
   WHERE Modality IN ('CT', 'MR', 'PT', 'XA', 'RF')
   GROUP BY SeriesInstanceUID
 )

assets/rtstruct_index.sql

@@ -16,7 +16,7 @@ WITH
       roi.ROIName,
       roi.ROIGenerationAlgorithm
     FROM
-      `bigquery-public-data.idc_v23.dicom_all`
+      `bigquery-public-data.idc_v24.dicom_all`
     CROSS JOIN
       UNNEST(StructureSetROISequence) AS roi
     WHERE
@@ -29,7 +29,7 @@ WITH
       obs.ReferencedROINumber,
       obs.RTROIInterpretedType
     FROM
-      `bigquery-public-data.idc_v23.dicom_all`
+      `bigquery-public-data.idc_v24.dicom_all`
     CROSS JOIN
       UNNEST(RTROIObservationsSequence) AS obs
     WHERE
@@ -40,7 +40,7 @@ WITH
       SOPInstanceUID,
       rt_series.SeriesInstanceUID AS referenced_SeriesInstanceUID
     FROM
-      `bigquery-public-data.idc_v23.dicom_all`
+      `bigquery-public-data.idc_v24.dicom_all`
     CROSS JOIN
       UNNEST(ReferencedFrameOfReferenceSequence) AS ref_frame
     CROSS JOIN

assets/seg_index.sql

@@ -25,7 +25,7 @@ WITH
               SegmentSequence,
               SegmentationType
             FROM
-              `bigquery-public-data.idc_v23.dicom_metadata`
+              `bigquery-public-data.idc_v24.dicom_metadata`
             WHERE
               # more reliable than Modality = "SEG"
               SOPClassUID = "1.2.840.10008.5.1.4.1.1.66.4"
@@ -101,7 +101,7 @@ WITH
           SourceImageSequence[SAFE_OFFSET(0)].ReferencedSOPInstanceUID
             AS sis_one
         FROM
-          `bigquery-public-data.idc_v23.dicom_all`
+          `bigquery-public-data.idc_v24.dicom_all`
         WHERE
           Modality = "SEG"
           AND SOPClassUID = "1.2.840.10008.5.1.4.1.1.66.4"
@@ -126,7 +126,7 @@ WITH
     FROM
       coalesced_ref
     JOIN
-      `bigquery-public-data.idc_v23.dicom_all` AS dicom_all
+      `bigquery-public-data.idc_v24.dicom_all` AS dicom_all
       ON
         coalesced_ref.referenced_sop = dicom_all.SOPInstanceUID
     RIGHT JOIN

assets/sm_index.sql

@@ -5,7 +5,7 @@
 -- SELECT
 --   * EXCEPT(Modality)
 -- FROM
---   `bigquery-public-data.idc_v19.dicom_metadata_curated_series_level`
+--   `bigquery-public-data.idc_v24.dicom_metadata_curated_series_level`
 -- WHERE
 --   Modality = "SM"
 
@@ -38,7 +38,7 @@ WITH
 
 
   FROM
-    `bigquery-public-data.idc_v23.dicom_all` AS dicom_all
+    `bigquery-public-data.idc_v24.dicom_all` AS dicom_all
   GROUP BY
     SeriesInstanceUID
   ),
@@ -52,7 +52,7 @@ SpecimenPreparationSequence_unnested AS (
         concept_code_sequence.CodeMeaning AS ccs_cm,
         concept_code_sequence.CodingSchemeDesignator AS ccs_csd,
         concept_code_sequence.CodeValue AS ccs_val,
-      FROM `bigquery-public-data.idc_v23.dicom_all`,
+      FROM `bigquery-public-data.idc_v24.dicom_all`,
       UNNEST(SpecimenDescriptionSequence[SAFE_OFFSET(0)].SpecimenPreparationSequence) as preparation_unnest_step1,
       UNNEST(preparation_unnest_step1.SpecimenPreparationStepContentItemSequence) as preparation_unnest_step2,
       UNNEST(preparation_unnest_step2.ConceptNameCodeSequence) as concept_name_code_sequence,

assets/sm_instance_index.sql

@@ -16,7 +16,7 @@ WITH
     concept_code_sequence.CodingSchemeDesignator AS ccs_csd,
     concept_code_sequence.CodeValue AS ccs_val,
   FROM
-    `bigquery-public-data.idc_v23.dicom_all`,
+    `bigquery-public-data.idc_v24.dicom_all`,
     UNNEST(SpecimenDescriptionSequence[SAFE_OFFSET(0)].SpecimenPreparationSequence) AS preparation_unnest_step1,
     UNNEST(preparation_unnest_step1.SpecimenPreparationStepContentItemSequence) AS preparation_unnest_step2,
     UNNEST(preparation_unnest_step2.ConceptNameCodeSequence) AS concept_name_code_sequence,
@@ -144,7 +144,7 @@ SELECT
   # unique identifier of the instance within the IDC
   dicom_all.crdc_instance_uuid AS crdc_instance_uuid
 FROM
-  `bigquery-public-data.idc_v23.dicom_all` AS dicom_all
+  `bigquery-public-data.idc_v24.dicom_all` AS dicom_all
 LEFT JOIN
   slide_embedding
 ON

scripts/python/idc_index_data_manager.py

@@ -454,7 +454,7 @@ def retrieve_latest_idc_release_version(self) -> int:
         SELECT
             MAX(idc_version) AS latest_idc_release_version
         FROM
-            `bigquery-public-data.idc_v23.version_metadata`
+            `bigquery-public-data.idc_v24.version_metadata`
         """
         query_job = self.client.query(query)
         result = query_job.result()

scripts/sql/analysis_results_index.sql

@@ -43,6 +43,6 @@ SELECT
   # citation for the analysis results collection that should be used for acknowledgment
   Citation
 FROM
-  `bigquery-public-data.idc_v23.analysis_results_metadata`
+  `bigquery-public-data.idc_v24.analysis_results_metadata`
 ORDER BY
   analysis_result_id