Upload smoking_metadata.json with huggingface_hub

smoking_metadata.json

@@ -0,0 +1,33 @@
+{
+  "dataset_name": "smoking",
+  "download_urls": {
+    "GSE50660": "https://ftp.ncbi.nlm.nih.gov/geo/series/GSE50nnn/GSE50660/matrix/GSE50660_series_matrix.txt.gz",
+    "HumanMethylation450_manifest": "https://webdata.illumina.com/downloads/productfiles/humanmethylation450/humanmethylation450_15017482_v1-2.csv"
+  },
+  "download_timestamp": "2026-05-22T17:53:43.060338",
+  "statistics": {
+    "num_samples": 464,
+    "num_features": 20763,
+    "target_stats": {
+      "class_mapping": {
+        "never": 0,
+        "ever": 1
+      },
+      "num_classes": 2,
+      "class_names": [
+        "never",
+        "ever"
+      ],
+      "samples_per_class": {
+        "never": 179,
+        "ever": 285
+      },
+      "original_geo_class_counts": {
+        "never": 179,
+        "former": 263,
+        "current": 22
+      }
+    }
+  },
+  "preprocessing_notes": "GSE50660 Illumina 450k beta values mapped to genes using the HumanMethylation450 v1.2 manifest, restricted to probes annotated as TSS1500 or TSS200 promoter regions. For each gene, the candidate promoter probe with the minimum mean beta across never-smoker samples (original GEO smoking == 0) is kept as the gene-level feature. Values are then median-centered per gene across samples. The original 3-class GEO smoking status (0=never, 1=former, 2=current) is collapsed into a binary target: 0=never, 1=ever (former or current)."
+}