Fix for https://github.com/bigscience-workshop/promptsource/issues/20

promptsource/promptsource.py

@@ -2,7 +2,8 @@
 import requests
 import streamlit as st
 from session import _get_state
-from utils import _ADDITIONAL_ENGLISH_DATSETS
+from utils import (_ADDITIONAL_ENGLISH_DATSETS, removeHyphen,
+                   renameDatasetColumn)
 
 from templates import Template, TemplateCollection
 
@@ -213,6 +214,10 @@ def list_datasets(template_collection, _priority_filter):
     )
     st.markdown(md)
 
+    st.sidebar.subheader("Dataset Schema")
+    st.sidebar.write(render_features(dataset.features))
+    dataset = renameDatasetColumn(dataset)
+
     dataset_templates = template_collection.get_dataset(dataset_key, conf_option.name if conf_option else None)
 
     template_list = dataset_templates.keys
@@ -230,6 +235,8 @@ def list_datasets(template_collection, _priority_filter):
     example_index = st.sidebar.slider("Select the example index", 0, len(dataset) - 1)
 
     example = dataset[example_index]
+    example = removeHyphen(example)
+
     st.sidebar.write(example)
 
     col1, _, col2 = st.beta_columns([18, 1, 6])

promptsource/utils.py

@@ -118,3 +118,23 @@
     "xtreme",
     "yelp_polarity",
 ]
+
+
+def removeHyphen(example):
+    example_clean = {}
+    for key in example.keys():
+        if "-" in key:
+            new_key = key.replace("-", "_")
+            example_clean[new_key] = example[key]
+        else:
+            example_clean[key] = example[key]
+    example = example_clean
+    return example
+
+
+def renameDatasetColumn(dataset):
+    col_names = dataset.column_names
+    for cols in col_names:
+        if "-" in cols:
+            dataset = dataset.rename_column(cols, cols.replace("-", "_"))
+    return dataset