Update app.py

app.py

@@ -35,7 +35,7 @@ uploaded_file = st.sidebar.file_uploader("Upload a medical text book (image)", t
 # Initialize the parser
 parser = LlamaParse(
     result_type="markdown",
-    parsing_instruction="You are given medical text book on medicine",
+    parsing_instruction="You are given a medical textbook on medicine",
     use_vendor_multimodal_model=True,
     vendor_multimodal_model_name="gpt-4o-mini-2024-07-18",
     show_progress=True,
@@ -46,14 +46,8 @@ parser = LlamaParse(
     language="en"
 )
 
-# Function to encode image to data URL
-def local_image_to_data_url(image_path):
-    mime_type, _ = guess_type(image_path)
-    if mime_type is None:
-        mime_type = 'image/png'
-    with open(image_path, "rb") as image_file:
-        base64_encoded_data = base64.b64encode(image_file.read()).decode('utf-8')
-    return f"data:{mime_type};base64,{base64_encoded_data}"
+# Initialize md_json_objs as an empty list
+md_json_objs = []
 
 # Upload and process file
 if uploaded_file:
@@ -70,6 +64,15 @@ if uploaded_file:
     st.write("File successfully processed!")
     st.write(f"Processed file: {uploaded_file.name}")
 
+# Function to encode image to data URL
+def local_image_to_data_url(image_path):
+    mime_type, _ = guess_type(image_path)
+    if mime_type is None:
+        mime_type = 'image/png'
+    with open(image_path, "rb") as image_file:
+        base64_encoded_data = base64.b64encode(image_file.read()).decode('utf-8')
+    return f"data:{mime_type};base64,{base64_encoded_data}"
+
 # Function to get sorted image files
 def get_page_number(file_name):
     match = re.search(r"-page-(\d+)\.jpg$", str(file_name))
@@ -97,8 +100,11 @@ def get_text_nodes(md_json_objs, image_dir) -> t.List[TextNode]:
             nodes.append(node)
     return nodes
 
-# Load text nodes
-text_nodes = get_text_nodes(md_json_objs, "data_images")
+# Load text nodes if md_json_objs is not empty
+if md_json_objs:
+    text_nodes = get_text_nodes(md_json_objs, "data_images")
+else:
+    text_nodes = []
 
 # Setup index and LLM
 embed_model = OpenAIEmbedding(model="text-embedding-3-large")
@@ -168,4 +174,4 @@ query_engine = MultimodalQueryEngine(QA_PROMPT, retriever, gpt_4o_mm)
 if query_text:
     st.write("Querying...")
     response = query_engine.custom_query(query_text)
-    st.markdown(response.response)
+    st.markdown(response.response)