Still debugging errors

This version includes:

Exception handling to catch and display errors.
Reduced the MAX_ROWS_PER_CHUNK to 50 to further mitigate the token limit issue.
Additional error messages to help you understand what's going wrong.

app.py

@@ -2,7 +2,6 @@ import streamlit as st
 import pandas as pd
 from io import StringIO
 from transformers import AutoTokenizer, AutoModelForTableQuestionAnswering
-import numpy as np
 
 # Initialize TAPAS model and tokenizer
 tokenizer = AutoTokenizer.from_pretrained("google/tapas-large-finetuned-wtq")
@@ -10,26 +9,20 @@ model = AutoModelForTableQuestionAnswering.from_pretrained("google/tapas-large-f
 
 def ask_llm_chunk(chunk, questions):
     chunk = chunk.astype(str)
+    try:
+        inputs = tokenizer(table=chunk, queries=questions, padding="max_length", return_tensors="pt")
+        if inputs["input_ids"].shape[1] > 512:
+            return ["Token limit exceeded for this chunk"] * len(questions)
+        outputs = model(**inputs)
+        predicted_answer_coordinates, predicted_aggregation_indices = tokenizer.convert_logits_to_predictions(
+            inputs,
+            outputs.logits.detach(),
+            outputs.logits_aggregation.detach()
+        )
+    except Exception as e:
+        st.write(f"An error occurred: {e}")
+        return ["Error processing this chunk"] * len(questions)
     
-    # Debugging statement to print chunk shape
-    print("Chunk shape:", chunk.shape)
-    print("Sample data:", chunk.head())
-    
-    # Count tokens
-    token_count = len(tokenizer.tokenize(str(chunk) + " ".join(questions)))
-    print("Token count:", token_count)
-    
-    if token_count > 512:
-        print("Warning: Token count exceeds maximum allowable sequence length.")
-        return ["Token limit exceeded for chunk"] * len(questions)
-        
-    inputs = tokenizer(table=chunk, queries=questions, padding="max_length", return_tensors="pt")
-    outputs = model(**inputs)
-    predicted_answer_coordinates, predicted_aggregation_indices = tokenizer.convert_logits_to_predictions(
-        inputs,
-        outputs.logits.detach(),
-        outputs.logits_aggregation.detach()
-    )
     answers = []
     for coordinates in predicted_answer_coordinates:
         if len(coordinates) == 1:
@@ -41,17 +34,22 @@ def ask_llm_chunk(chunk, questions):
             answers.append(", ".join(cell_values))
     return answers
 
-MAX_ROWS_PER_CHUNK = 200
+MAX_ROWS_PER_CHUNK = 50  # Reduced chunk size
 
 def summarize_map_reduce(data, questions):
-    dataframe = pd.read_csv(StringIO(data))
+    try:
+        dataframe = pd.read_csv(StringIO(data))
+    except Exception as e:
+        st.write(f"Error reading the CSV file: {e}")
+        return []
+        
     num_chunks = len(dataframe) // MAX_ROWS_PER_CHUNK + 1
     dataframe_chunks = np.array_split(dataframe, num_chunks)
     all_answers = []
     for chunk in dataframe_chunks:
         chunk_answers = ask_llm_chunk(chunk, questions)
         all_answers.extend(chunk_answers)
-    return all_answers  # For now, simply returning the answers for each chunk
+    return all_answers
 
 st.title("TAPAS Table Question Answering")
 
@@ -69,8 +67,11 @@ if csv_file is not None:
 
     if st.button("Submit"):
         if data and questions:
-            answers = summarize_map_reduce(data, questions)
-            st.write("Answers:")
-            for q, a in zip(questions, answers):
-                st.write(f"Question: {q}")
-                st.write(f"Answer: {a}")
+            try:
+                answers = summarize_map_reduce(data, questions)
+                st.write("Answers:")
+                for q, a in zip(questions, answers):
+                    st.write(f"Question: {q}")
+                    st.write(f"Answer: {a}")
+            except Exception as e:
+                st.write(f"An error occurred: {e}")