Update app.py

app.py

@@ -1,5 +1,8 @@
 import warnings
 warnings.simplefilter(action='ignore', category=FutureWarning)
+import logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
 
 import fitz
 import gradio as gr
@@ -20,7 +23,7 @@ llm = HuggingFaceEndpoint(
 llm_engine_hf = ChatHuggingFace(llm=llm)
 
 def read_pdf(file_path):
-    print("It is a PDF file")
+    logger.info("Reading a PDF file")
     try:
         pdf_document = fitz.open(file_path)
         text = ""
@@ -28,15 +31,28 @@ def read_pdf(file_path):
             page = pdf_document[page_num]
             text += page.get_text()
         
+        if not text.strip():
+            message = "PDF contains no text. It may be due to the PDF being password-protected, collapsed, or full of images."
+            logger.info(message)
+            return message
+        
         return text
+        
     except Exception as e:
-        print("Error reading file,", e)
+        error_message = f"Error reading PDF file: {e}"
+        logger.error(error_message)
+        return error_message
         
 def read_txt(file_path):
-    print("It is not a PDF file")
-    with open(file_path, "r", encoding="utf-8") as f:
-        text = f.read()
-    return text
+    logger.info("Reading a TXT file")
+    try:
+        with open(file_path, "r", encoding="utf-8") as f:
+            text = f.read()
+        return text
+    except Exception as e:
+        error_message = f"Error reading TXT file: {e}"
+        logger.error(error_message)
+        return error_message
 
 def summarize(file, n_words):
     global llm
@@ -47,11 +63,8 @@ def summarize(file, n_words):
     else:
         text = read_txt(file_path)
 
-    print("Length of text is ", len(text))
-    if len(text) > CONTEXT_WINDOW:
-        print(f"Slicing the first {CONTEXT_WINDOW} characters")
-        text = text[:CONTEXT_WINDOW]
-            
+    logger.info("Length of text is %d", len(text))
+                
     lang = detect(text[:CONTEXT_WINDOW])
     template_translate = '''
 Please carefully read the following document:
@@ -74,6 +87,8 @@ The summary should be in {LANG} language.
         formatted_prompt = prompt_summarize.format(TEXT=chunk, LANG=lang)
         summary = llm.invoke(formatted_prompt)
         summaries.append(summary)
+
+    logger.info(f"Chunked into {len(summaries)}.")
     
     final_summary = "\n\n".join(summaries)
     return final_summary