Fixing the app

app.py

@@ -4,6 +4,8 @@ import spacy
 import shutil
 import pickle
 import random
+import hashlib
+
 import logging
 import asyncio
 import warnings
@@ -24,6 +26,11 @@ warnings.filterwarnings("ignore", category=UserWarning)
 # Constants
 TITLE = "🌐 Text2Graph: Extract Knowledge Graphs from Natural Language"
 SUBTITLE = "✨ Extract and visualize knowledge graphs from texts in any language!"
+MIN_CHARS = 20
+MAX_CHARS = 3500
+
+# Keep track of all processed texts
+doc_ids = []
 
 # Basic CSS for styling
 CUSTOM_CSS = """
@@ -99,7 +106,7 @@ def extract_kg(text="", model_name=MODEL_LIST[0], model=None):
         raise gr.Error("⚠️ Both text and model must be provided!")
     if not model:
         raise gr.Error("⚠️ Model must be provided!")
-
+    
     try:
         start_time = time.time()
         result = model.extract(text, model_name)
@@ -285,15 +292,6 @@ def process_and_visualize(text, model_name, progress=gr.Progress()):
     # Check if we're processing the first example for caching
     is_first_example = text == EXAMPLES[0][0]
 
-    # Clear the working directory if it exists
-    if os.path.exists(WORKING_DIR):
-        shutil.rmtree(WORKING_DIR)
-    os.makedirs(WORKING_DIR, exist_ok=True)
-    
-    # Initialize the LLMGraph model
-    model = LLMGraph()
-    asyncio.run(model.initialize_rag()) 
-
     # Try to load from cache if it's the first example
     if is_first_example and model_name == MODEL_LIST[0] and os.path.exists(EXAMPLE_CACHE_FILE):
         try:
@@ -306,6 +304,28 @@ def process_and_visualize(text, model_name, progress=gr.Progress()):
         except Exception as e:
             logging.error(f"Cache loading error: {str(e)}")
 
+    # Catch too long or too short text
+    if len(text) < MIN_CHARS:
+        raise gr.Error(f"⚠️ Text is too short! Please provide at least {MIN_CHARS} characters.")
+    if len(text) > MAX_CHARS:
+        raise gr.Error(f"⚠️ Text is too long! Please provide no more than {MAX_CHARS} characters.")
+    
+    if model_name == MODEL_LIST[1]:
+        # Compute the unique hash for the document
+        doc_id = hashlib.md5(text.strip().encode()).hexdigest()
+
+        if doc_id not in doc_ids:
+            doc_ids.append(doc_id)
+
+            # Clear the working directory if it exists
+            if os.path.exists(WORKING_DIR):
+                shutil.rmtree(WORKING_DIR)
+            os.makedirs(WORKING_DIR, exist_ok=True)
+            
+            # Initialize the LLMGraph model
+            model = LLMGraph()
+            asyncio.run(model.initialize_rag()) 
+
     # Continue with normal processing if cache fails
     progress(0, desc="Starting extraction...")
     json_data = extract_kg(text, model_name, model)