the user selects the choice, then optionally the checkbox and then the Next button

app.py

@@ -8,11 +8,9 @@ CSV_FILENAME = "user_selections.csv"
 
 def assign_samples(csv_path):
     df = pd.read_csv(csv_path)
-
-    group_1 = df[(df["pool_pos"] == 1) & (~df["question_category"].str.endswith("_B"))].head(10)
-    group_2 = df[(df["pool_pos"] == 2) & (~df["question_category"].str.endswith("_B"))].head(10)
-    group_3 = df[(df["pool_pos"] == 3) & (~df["question_category"].str.endswith("_B"))].head(10)
-
+    group_1 = df[(df["pool_pos"] == 1) & (~df["question_category"].str.endswith("_B"))].head(5)
+    group_2 = df[(df["pool_pos"] == 2) & (~df["question_category"].str.endswith("_B"))].head(5)
+    group_3 = df[(df["pool_pos"] == 3) & (~df["question_category"].str.endswith("_B"))].head(5)
     return {
         "Bernardo": group_1,
         "Alessandro": group_1,
@@ -25,32 +23,32 @@ def assign_samples(csv_path):
     }
 
 def load_existing_annotations():
-    """Load the existing annotations from the HF dataset."""
     try:
         file_path = hf_hub_download(HF_REPO, CSV_FILENAME, repo_type="dataset", token=st.secrets["HF_TOKEN"])
         return pd.read_csv(file_path)
     except Exception:
-        return pd.DataFrame(columns=["username", "id"])  # Return empty DataFrame if not found
+        return pd.DataFrame(columns=["username", "id"])
 
-# Load datasets
 csv_file = "static/mc.csv"
 assignments = assign_samples(csv_file)
 existing_annotations = load_existing_annotations()
 
 valid_users = list(assignments.keys())
 
-# Initialize session state
 if "username" not in st.session_state:
     st.session_state.username = None
 if "index" not in st.session_state:
     st.session_state.index = 0
 if "results" not in st.session_state:
     st.session_state.results = []
+if "selected_answer" not in st.session_state:
+    st.session_state.selected_answer = None
+if "not_enough_info" not in st.session_state:
+    st.session_state.not_enough_info = False
 
 def update_name():
-    """Set username and reset index."""
     st.session_state.username = st.session_state.selected_user
-    st.session_state.index = 0  # Reset progress
+    st.session_state.index = 0
 
 if st.session_state.username is None:
     with st.form("user_form"):
@@ -59,12 +57,10 @@ if st.session_state.username is None:
         submit_button = st.form_submit_button("Start", on_click=update_name)
     st.stop()
 
-# Get assigned dataset and remove already labeled samples
 full_dataset = assignments[st.session_state.username].reset_index(drop=True)
 user_labeled_ids = existing_annotations[existing_annotations["username"] == st.session_state.username]["id"].tolist()
 dataset = full_dataset[~full_dataset["id"].isin(user_labeled_ids)].reset_index(drop=True)
 
-# If all samples are labeled, stop execution
 if dataset.empty:
     st.write("### Great! You have completed your assignment. 🎉")
     st.stop()
@@ -78,7 +74,11 @@ def push_to_hf_hub(csv_path):
     except Exception as e:
         print(f"Error pushing to HF: {e}")
 
-def save_choice(choice_index):
+def save_choice():
+    if st.session_state.selected_answer is None:
+        st.warning("Please select an answer before proceeding.")
+        return
+    
     sample = dataset.iloc[st.session_state.index]
     st.session_state.results.append({
         "username": st.session_state.username,
@@ -86,15 +86,16 @@ def save_choice(choice_index):
         "video_id": sample["video_id"],
         "answer1": sample["answer1"],
         "answer2": sample["answer2"],
-        "selected_answer": choice_index,
+        "selected_answer": st.session_state.selected_answer,
         "target": sample["target"],
-        "not_enough_info": not_enough_info
+        "not_enough_info": st.session_state.not_enough_info
     })
     
     st.session_state.index += 1
-    st.session_state.checkbox = False # reset the checkbox
+    st.session_state.selected_answer = None
+    st.session_state.not_enough_info = False
     
-    if st.session_state.index >= len(dataset):  # All remaining samples done
+    if st.session_state.index >= len(dataset):
         st.write("### Great! You have completed your assignment. 🎉")
         result_df = pd.DataFrame(st.session_state.results)
         csv_path = "user_selections.csv"
@@ -107,43 +108,37 @@ def save_choice(choice_index):
         st.stop()
     return
 
-# Select the current sample
 sample = dataset.iloc[st.session_state.index]
 
-# Title
 st.markdown("<h1 style='text-align: center; font-size: 50px;'>MAIA Sample</h1>", unsafe_allow_html=True)
-
-# Centered user name
 st.markdown(f"<h3 style='text-align: center;'>User: {st.session_state.username}</h3>", unsafe_allow_html=True)
+st.write("\n\n")
 
-st.write("\n\n")  # Add empty space
-
-# Instructions
 st.markdown("""
 ### Instructions:
 - Look at the video thumbnail, do not play it!
-- Select the checkbox if you think so.
-- Then choose one of the two answers.
+- Select one of the two answers.
+- If you think the frame does not provide enough information, select the checkbox.
+- Click 'Next' to proceed.
 """)
+st.write("---")
 
-st.write("---")  # Adds a horizontal divider for better separation
-
-# Display Video
 st.video(sample["video_url"])
 
-# Question Prompt
 st.markdown("<h4 style='text-align: center; margin-top: 20px;'>Scegli la descrizione corretta tra A e B</h4>", unsafe_allow_html=True)
 
-# Checkbox for uncertainty
-st.markdown("<div style='text-align: center;'>", unsafe_allow_html=True)
-not_enough_info = st.checkbox("The frame does not provide enough information to answer the question.", key='checkbox')
-st.markdown("</div>", unsafe_allow_html=True)
-
-st.write("\n")  # Add spacing before buttons
-
-# Buttons (Centered)
 col1, col2 = st.columns(2)
 with col1:
-    st.button(f"A: {sample.get('answer1', 'No answer1 available')}", on_click=lambda: save_choice(0))
+    if st.button(f"A: {sample.get('answer1', 'No answer1 available')}"):
+        st.session_state.selected_answer = 0
 with col2:
-    st.button(f"B: {sample.get('answer2', 'No answer2 available')}", on_click=lambda: save_choice(1))
+    if st.button(f"B: {sample.get('answer2', 'No answer2 available')}"):
+        st.session_state.selected_answer = 1
+
+st.markdown("<div style='text-align: center;'>", unsafe_allow_html=True)
+st.session_state.not_enough_info = st.checkbox("The frame does not provide enough information to answer the question.")
+st.markdown("</div>", unsafe_allow_html=True)
+st.write("\n")
+
+if st.button("Next"):
+    save_choice()