[UPDATE, DOCS] update app.py

- add emoji to prediction results
- update space name

README.md

@@ -1,5 +1,5 @@
 ---
-title: Climate Plus Demo
+title: climate-plus demo
 emoji: 🌦️
 colorFrom: yellow
 colorTo: blue

app.py

@@ -71,7 +71,12 @@ def tcfd_classify(text):
 
 
 data1 = {
-    "example": ["Example 1", "Example 2", "Example 3", "Example 4"],
+    "example": [
+        "Example 1 (Sea ice has diminished much faster than scientists and climate models anticipated.)",
+        "Example 2 (Climate Models Have Overestimated Global Warming)",
+        "Example 3 (Climate skeptics argue temperature records have been adjusted in recent years to ...)",
+        "Example 4 (Humans are too insignificant to affect global climate.)",
+    ],
     "claim": [
         "Sea ice has diminished much faster than scientists and climate models anticipated.",
         "Climate Models Have Overestimated Global Warming",
@@ -88,7 +93,12 @@ data1 = {
 }
 
 data2 = {
-    "example": ["Example 1", "Example 2", "Example 3", "Example 4"],
+    "example": [
+        "Example 1 (As a global provider of transport and logistics services ...)",
+        "Example 2 (There are no sentences in the provided excerpts that disclose Scope 1 and Scope 2)",
+        "Example 3 (Our strategy needs to be resilient under a range of climate-related scenarios.)",
+        "Example 4 (AXA created a Group-level Responsible Investment Committee ...)",
+    ],
     "text": [
         "As a global provider of transport and logistics services, we are often called on for expert input and industry insights by government representatives.",
         "There are no sentences in the provided excerpts that disclose Scope 1 and Scope 2, and, if appropriate Scope 3 GHG emissions. The provided excerpts focus on other metrics and targets related to social impact investing, assets under management, and carbon footprint calculations.",
@@ -103,9 +113,27 @@ data2 = {
     ],
 }
 
+
+def get_pred_emoji(str1, str2, mode="factcheck"):
+    if mode == "factcheck":
+        if str1 == str2:
+            return "✅"
+        else:
+            return "❌"
+    elif mode == "tcfd":
+        if str1 == str2:
+            return "✅"
+        elif str1.split()[:-1] == str2.split()[:-1]:
+            return "🔧"
+        else:
+            return "❌"
+
+
 df1 = pd.DataFrame(data1)
 df2 = pd.DataFrame(data2)
 
+st.markdown("# climate-plus demo")
+st.markdown("This is a minimal example of two models we trained for `climate-plus` project. See the [GitHub repo](https://github.com/rexarski/climate-plus) for more details.")
 
 st.markdown("## Factchecking")
 
@@ -114,20 +142,32 @@ ex1_selected = st.selectbox(
 )
 selected_row1 = df1[df1["example"] == ex1_selected]
 
-st.markdown(f'**Claim**: {selected_row1["claim"].values[0]}')
-st.markdown(f'**Evidence**: {selected_row1["evidence"].values[0]}')
-st.markdown(f'**Label**: {selected_row1["label"].values[0]}')
-st.markdown("---")
+ex_claim = selected_row1["claim"].values[0]
+ex_evidence = selected_row1["evidence"].values[0]
+ex_label = selected_row1["label"].values[0]
+ex_pred = factcheck(
+    selected_row1["claim"].values[0], selected_row1["evidence"].values[0]
+)
+st.markdown(f"**Claim**: {ex_claim}")
+st.markdown(f"**Evidence**: {ex_evidence}")
+st.markdown(f"**Label**: {ex_label}")
 st.markdown(
-    f'**Prediction**: {factcheck(selected_row1["claim"].values[0], selected_row1["evidence"].values[0])}'
+    f'**Prediction**: {ex_pred} {get_pred_emoji(ex_label, ex_pred, mode="factcheck")}'
 )
 
+st.markdown("---")
+
 st.markdown("## TCFD disclosure classification")
 
 ex2_selected = st.selectbox("Select a TCFD disclosure example", df2["example"])
 selected_row2 = df2[df2["example"] == ex2_selected]
 
-st.markdown(f'**Text**: {selected_row2["text"].values[0]}')
-st.markdown(f'**Label**: {selected_row2["label"].values[0]}')
-st.markdown("---")
-st.markdown(f'**Prediction**: {tcfd_classify(selected_row2["text"].values[0])}')
+ex_text = selected_row2["text"].values[0]
+ex_label2 = selected_row2["label"].values[0]
+ex_pred2 = tcfd_classify(selected_row2["text"].values[0])
+
+st.markdown(f"**Text**: {ex_text}")
+st.markdown(f"**Label**: {ex_label2}")
+st.markdown(
+    f'**Prediction**: {ex_pred2} {get_pred_emoji(ex_label2, ex_pred2, mode="tcfd")}'
+)