fix main restart time

main.py

@@ -19,86 +19,40 @@ load_dotenv()
 BASE_DIR = '/tmp'  # Use direct /tmp path for Hugging Face
 log_dir = os.path.join(BASE_DIR, 'app_logs')
 cache_dir = os.path.join(BASE_DIR, 'app_cache')
+nltk_data_dir = os.path.join(BASE_DIR, 'nltk_data')
+gensim_data_dir = os.path.join(BASE_DIR, 'gensim-data')
+upload_dir = os.path.join(BASE_DIR, 'uploads')
+ans_image_dir = os.path.join(BASE_DIR, 'ans_image')
+images_dir = os.path.join(BASE_DIR, 'images')
+
+# Global variables for model caching
+global_models = {}
 
 def ensure_directory(path):
     """Create directory and ensure full permissions with better error handling"""
+    if os.path.exists(path):
+        return path
+        
     try:
-        # First try to create the directory with full permissions
         os.makedirs(path, mode=0o777, exist_ok=True)
-        
-        # Double-check the permissions and try to set them again if needed
-        current_mode = os.stat(path).st_mode & 0o777
-        if current_mode != 0o777:
-            os.chmod(path, 0o777)
-            
-        # Verify the directory is writable
-        test_file = os.path.join(path, '.test_write')
-        try:
-            with open(test_file, 'w') as f:
-                f.write('test')
-            os.remove(test_file)
-        except Exception as e:
-            raise Exception(f"Directory {path} is not writable: {str(e)}")
-            
+        return path
     except Exception as e:
-        # If we can't set full permissions, try with more restrictive ones
         try:
             os.makedirs(path, mode=0o755, exist_ok=True)
-            os.chmod(path, 0o755)
-            
-            # Verify the directory is at least readable
-            if not os.access(path, os.R_OK | os.W_OK):
-                raise Exception(f"Cannot read/write to directory: {path}")
-                
+            return path
         except Exception as nested_e:
-            raise Exception(f"Could not create or set permissions for {path}: {str(nested_e)}")
-            
-    return path
-
-# Create necessary directories with full permissions
-ensure_directory(log_dir)
-ensure_directory(cache_dir)
-
-# Create a log file with timestamp and ensure it's writable
-log_file = os.path.join(log_dir, f'app_{datetime.now().strftime("%Y%m%d_%H%M%S")}.log')
-try:
-    # Create log file with full permissions
-    with open(log_file, 'w') as f:
-        pass
-    os.chmod(log_file, 0o666)
-except Exception as e:
-    print(f"Warning: Could not create log file: {e}")
-    log_file = os.path.join(BASE_DIR, 'app.log')
-    with open(log_file, 'w') as f:
-        pass
-    os.chmod(log_file, 0o666)
-
-# Set up logging to both file and console
-try:
-    logging.basicConfig(
-        level=logging.INFO,
-        format='%(asctime)s - %(levelname)s - %(message)s',
-        handlers=[
-            logging.FileHandler(log_file),
-            logging.StreamHandler(sys.stdout)
-        ]
-    )
-except Exception as e:
-    print(f"Warning: Could not set up file logging: {e}")
-    # Fallback to console-only logging
-    logging.basicConfig(
-        level=logging.INFO,
-        format='%(asctime)s - %(levelname)s - %(message)s',
-        handlers=[logging.StreamHandler(sys.stdout)]
-    )
-
-# Create a custom logger
+            print(f"Warning: Could not create directory {path}: {nested_e}")
+            return path
+
+# Simplified logging setup
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(levelname)s - %(message)s',
+    handlers=[logging.StreamHandler(sys.stdout)]
+)
 logger = logging.getLogger(__name__)
 
-# Add a print function that also logs
 def log_print(message, level="INFO"):
-    # Use the original print function to avoid recursion
-    builtins.print(message, flush=True)
     if level == "INFO":
         logger.info(message)
     elif level == "ERROR":
@@ -106,24 +60,24 @@ def log_print(message, level="INFO"):
     elif level == "WARNING":
         logger.warning(message)
 
-# Set up all cache and data directories in /tmp with full permissions
-try:
-    nltk_data_dir = os.path.join(BASE_DIR, 'nltk_data')
-    gensim_data_dir = os.path.join(BASE_DIR, 'gensim-data')
-    upload_dir = os.path.join(BASE_DIR, 'uploads')
-    ans_image_dir = os.path.join(BASE_DIR, 'ans_image')
-    images_dir = os.path.join(BASE_DIR, 'images')
-
-    # Create all directories with full permissions
-    for directory in [nltk_data_dir, gensim_data_dir, upload_dir, ans_image_dir, images_dir]:
-        ensure_directory(directory)
-
-except Exception as e:
-    print(f"Warning: Could not create cache directories: {e}")
-    # Create directories in /tmp directly
-    for dirname in ['nltk_data', 'gensim-data', 'uploads', 'ans_image', 'images']:
-        path = os.path.join(BASE_DIR, dirname)
-        ensure_directory(path)
+# Initialize NLTK only if not already downloaded
+def initialize_nltk():
+    required_nltk_data = ['stopwords', 'punkt', 'wordnet']
+    for data in required_nltk_data:
+        try:
+            # Check if the data exists in the custom directory
+            nltk.data.find(os.path.join('tokenizers', data))
+        except LookupError:
+            try:
+                log_print(f"Downloading NLTK data: {data}")
+                nltk.download(data, download_dir=nltk_data_dir, quiet=True)
+            except Exception as e:
+                log_print(f"Error downloading NLTK data {data}: {e}", "ERROR")
+
+# Create essential directories only
+essential_dirs = [cache_dir, upload_dir, images_dir]
+for directory in essential_dirs:
+    ensure_directory(directory)
 
 # Set environment variables
 os.environ['HF_HOME'] = cache_dir
@@ -132,18 +86,12 @@ os.environ['GENSIM_DATA_DIR'] = gensim_data_dir
 # Add the custom directory to NLTK's search path
 nltk.data.path.insert(0, nltk_data_dir)
 
-# Download required NLTK data
-required_nltk_data = ['stopwords', 'punkt', 'wordnet']
-for data in required_nltk_data:
-    try:
-        log_print(f"Downloading NLTK data: {data}")
-        nltk.download(data, download_dir=nltk_data_dir)
-    except Exception as e:
-        log_print(f"Error downloading NLTK data {data}: {e}", "ERROR")
-        raise
+# Initialize NLTK in background
+from threading import Thread
+nltk_thread = Thread(target=initialize_nltk)
+nltk_thread.start()
 
 from flask import Flask, request, jsonify, render_template
-import json 
 import torch
 from werkzeug.utils import secure_filename
 
@@ -614,43 +562,52 @@ def allowed_file(filename, allowed_extensions):
     return '.' in filename and \
            filename.rsplit('.', 1)[1].lower() in allowed_extensions
 
+def get_or_load_model(model_name):
+    """Get a model from cache or load it if not present"""
+    if model_name not in global_models:
+        if model_name == 'fasttext':
+            from gensim.models.fasttext import FastText
+            global_models[model_name] = FastText.load(os.path.join(cache_dir, 'fasttext.model'))
+        elif model_name == 'llm':
+            # Add your LLM model loading here
+            pass
+    return global_models.get(model_name)
+
 def cleanup_temp_files():
     """Clean up temporary files with proper error handling"""
     try:
         if os.path.exists(images_dir):
-            # First ensure we have write permissions
-            try:
-                os.chmod(images_dir, 0o777)
-            except Exception as e:
-                log_print(f"Warning: Could not set directory permissions: {e}", "WARNING")
-            
-            # Try to remove files
             for file in os.listdir(images_dir):
-                file_path = os.path.join(images_dir, file)
                 try:
+                    file_path = os.path.join(images_dir, file)
                     if os.path.isfile(file_path):
-                        try:
-                            os.chmod(file_path, 0o666)  # Ensure we can delete the file
-                        except:
-                            pass
                         os.unlink(file_path)
                 except Exception as e:
                     log_print(f"Warning: Could not delete file {file_path}: {e}", "WARNING")
-                    continue
-            
-            # Recreate the directory with full permissions
-            ensure_directory(images_dir)
-            
-        log_print("Successfully cleaned up temporary files")
     except Exception as e:
         log_print(f"Error cleaning up temporary files: {e}", "ERROR")
 
 if __name__ == '__main__':
     try:
-        # Ensure all directories exist and have proper permissions before starting
-        for directory in [log_dir, cache_dir, nltk_data_dir, gensim_data_dir, upload_dir, ans_image_dir, images_dir]:
+        # Create essential directories
+        for directory in essential_dirs:
             ensure_directory(directory)
         
+        # Start the NLTK initialization in background
+        nltk_thread.start()
+        
+        # Pre-load models in background
+        def preload_models():
+            try:
+                get_or_load_model('fasttext')
+                # Add other model preloading here
+            except Exception as e:
+                log_print(f"Warning: Could not preload models: {e}", "WARNING")
+        
+        model_thread = Thread(target=preload_models)
+        model_thread.start()
+        
+        # Start the Flask app
         port = int(os.environ.get('PORT', 7860))
         app.run(host='0.0.0.0', port=port, debug=False)
     finally: