printing extracted text11

main.py

@@ -2,10 +2,18 @@ import os
 import tempfile
 import gc
 import psutil
+import time
 
-# Set up Hugging Face cache directory
+# Set environment variables before any other imports
 os.environ['TRANSFORMERS_CACHE'] = os.path.join(tempfile.gettempdir(), 'huggingface_cache')
-os.makedirs(os.environ['TRANSFORMERS_CACHE'], exist_ok=True)
+os.environ['HF_HOME'] = os.path.join(tempfile.gettempdir(), 'huggingface')
+os.environ['TORCH_HOME'] = os.path.join(tempfile.gettempdir(), 'torch')
+os.environ['XDG_CACHE_HOME'] = os.path.join(tempfile.gettempdir(), 'cache')
+
+# Create cache directories
+for cache_dir in [os.environ['TRANSFORMERS_CACHE'], os.environ['HF_HOME'], 
+                 os.environ['TORCH_HOME'], os.environ['XDG_CACHE_HOME']]:
+    os.makedirs(cache_dir, exist_ok=True)
 
 # Now import the rest of the dependencies
 import sys
@@ -23,49 +31,53 @@ import json
 import queue
 import threading
 from threading import Thread, Event
-import time
-import nltk
-import gensim
-from gensim.models import FastText
-from sentence_transformers import SentenceTransformer
-from transformers import pipeline
-import torch
-import torch.nn.functional as F
-from sklearn.feature_extraction.text import TfidfVectorizer
-from sklearn.metrics.pairwise import cosine_similarity
-import re
-import string
-import unicodedata
 import warnings
 from flask_cors import CORS
 from dotenv import load_dotenv
 warnings.filterwarnings('ignore')
 
-# Add the project root directory to Python path
-sys.path.append(os.path.dirname(os.path.abspath(__file__)))
-
-# Set up Hugging Face cache directory
-os.environ['TRANSFORMERS_CACHE'] = os.path.join(tempfile.gettempdir(), 'huggingface_cache')
-os.environ['HF_HOME'] = os.path.join(tempfile.gettempdir(), 'huggingface')
+# Import ML libraries with timeout protection
+def import_with_timeout(import_statement, timeout=30):
+    """Import a module with a timeout to prevent hanging"""
+    result = {'success': False, 'module': None, 'error': None}
+    
+    def _import():
+        try:
+            if isinstance(import_statement, str):
+                result['module'] = __import__(import_statement)
+            else:
+                exec(import_statement)
+            result['success'] = True
+        except Exception as e:
+            result['error'] = str(e)
+    
+    thread = Thread(target=_import)
+    thread.daemon = True
+    thread.start()
+    thread.join(timeout=timeout)
+    
+    if thread.is_alive():
+        return None, f"Import timed out after {timeout} seconds"
+    
+    return result['module'], result['error']
 
-# Create cache directory if it doesn't exist
-os.makedirs(os.environ['TRANSFORMERS_CACHE'], exist_ok=True)
-os.makedirs(os.environ['HF_HOME'], exist_ok=True)
+# Import ML libraries safely
+nltk, nltk_error = import_with_timeout('nltk')
+if nltk_error:
+    log_print(f"Warning: NLTK import failed: {nltk_error}", "WARNING")
 
-# Import HTR modules
-from HTR.app import extract_text_from_image
-from HTR.word import convert_image
-from HTR.strike import struck_images
-from HTR.hcr import text
-from HTR.spell_and_gramer_check import spell_grammer
+gensim, gensim_error = import_with_timeout('gensim')
+if gensim_error:
+    log_print(f"Warning: Gensim import failed: {gensim_error}", "WARNING")
 
-# Import utils
-from utils import notification_queue, log_print
+torch, torch_error = import_with_timeout('torch')
+if torch_error:
+    log_print(f"Warning: PyTorch import failed: {torch_error}", "WARNING")
 
-# Load environment variables
-load_dotenv()
+# Add the project root directory to Python path
+sys.path.append(os.path.dirname(os.path.abspath(__file__)))
 
-# Create directories in /tmp which is writable
+# Create cache directory if it doesn't exist
 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')
@@ -142,7 +154,7 @@ def get_or_load_model(model_name):
                 return None
         except Exception as e:
             log_print(f"Error loading {model_name} model: {e}", "ERROR")
-            raise
+            return None
     return global_models.get(model_name)
 
 def initialize_resources():
@@ -162,7 +174,8 @@ def initialize_resources():
                     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")
+                    log_print(f"Error downloading NLTK data {data}: {e}", "WARNING")
+                    continue
 
         # Initialize models
         try: