Fix issue with yaml for downsampled sparsified versions

README.md

@@ -60,13 +60,13 @@ configs:
   - name: clustering_desc
     dtype: string
   - name: id
-    dtype: int64
+    dtype: int32
   - name: start idx
-    dtype: int64
+    dtype: int32
   - name: end idx
-    dtype: int64
+    dtype: int32
   - name: length
-    dtype: int64
+    dtype: int32
   - name: cluster_id
     dtype: int32
   - name: correlation to model
@@ -76,9 +76,9 @@ configs:
   - name: correlation achieved with tolerance
     dtype: string
   - name: MAE
-    dtype: float64
+    dtype: float32
   - name: relaxed MAE
-    dtype: float64
+    dtype: float32
 - config_name: raw_partial_data
   data_files:
   - split: exploratory
@@ -575,7 +575,20 @@ configs:
     path: "exploratory/irregular_p30/resampled_1min/*-data.parquet"
   - split: confirmatory
     path: "confirmatory/irregular_p30/resampled_1min/*-data.parquet"
-  features: *id003
+  features: &id005
+  - name: subject_id
+    dtype: string
+  - name: datetime
+    dtype: string
+    type: timestamp
+  - name: original_index
+    dtype: int64
+  - name: iob
+    dtype: float64
+  - name: cob
+    dtype: float64
+  - name: ig
+    dtype: float64
 - config_name: downsampled_partial_labels
   data_files:
   - split: exploratory
@@ -596,7 +609,7 @@ configs:
     path: "exploratory/irregular_p90/resampled_1min/*-data.parquet"
   - split: confirmatory
     path: "confirmatory/irregular_p90/resampled_1min/*-data.parquet"
-  features: *id003
+  features: *id005
 - config_name: downsampled_sparse_labels
   data_files:
   - split: exploratory
@@ -612,6 +625,7 @@ configs:
     path: "confirmatory/irregular_p90/resampled_1min/bad_partitions/*-labels.parquet"
   features: *id002
 ---
+
 # CSTS - Correlation Structures in Time Series
 
 ## Important Notice