Update app.py

app.py

@@ -1,5 +1,6 @@
 import os
 import re
+import io
 import gradio as gr
 from typing import Tuple, Optional, Dict, List
 from deep_translator import GoogleTranslator
@@ -13,33 +14,60 @@ from PIL import Image
 # Load environment variables
 load_dotenv()
 
-def ocr_from_image(image_path: str) -> str:
-    """
-    Extracts text from an image using OCR (Optical Character Recognition).
-    
-    Args:
-        image_path (str): The path to the image file.
+class OCRProcessor:
+    def __init__(self):
+        # Fetch the API keys from environment variables (secrets)
+        self.api_keys = [
+            os.getenv("ocr_space_api_key1"),  # Get the first API key from environment variable
+            os.getenv("ocr_space_api_key2")   # Get the second API key from environment variable
+        ]
+
+    def ocr_from_image(self, image: Image.Image) -> str:
+        """
+        Extract text from an uploaded image using OCR (Optical Character Recognition).
         
-    Returns:
-        str: The extracted text from the image, or an error message if OCR fails.
-    """
-    try:
-        # Open the image file
-        image = Image.open(image_path)
+        Args:
+            image (Image.Image): The PIL image object from which text will be extracted.
         
-        # Use pytesseract to extract text
-        extracted_text = pytesseract.image_to_string(image)
+        Returns:
+            str: The extracted text from the image.
+        """
+        for api_key in self.api_keys:
+            try:
+                # Prepare the API endpoint and data
+                url = "https://api.ocr.space/parse/image"
+                payload = {
+                    'apikey': api_key,
+                    'language': 'eng',  # You can set this to the desired language
+                }
+                
+                # Convert the image to bytes (PIL image to byte array)
+                img_byte_arr = io.BytesIO()
+                image.save(img_byte_arr, format='PNG')
+                img_byte_arr = img_byte_arr.getvalue()
+                
+                # Make the API request
+                response = requests.post(url, data=payload, files={'file': img_byte_arr})
+                
+                # Parse the JSON response
+                result = response.json()
+                
+                # Check if the response is valid and contains parsed text
+                if 'ParsedResults' in result:
+                    extracted_text = result['ParsedResults'][0]['ParsedText']
+                    return extracted_text.strip()
+                else:
+                    # If the OCR response is empty or contains no parsed text, handle that
+                    error_message = result.get('ErrorMessage', 'Unknown error')
+                    print(f"Error from OCR API: {error_message}")
+                    return f"Error in OCR extraction: {error_message}"
 
-        # Clean up the extracted text (if needed, remove unnecessary spaces or special characters)
-        cleaned_text = extracted_text.strip()
-        
-        # Return the extracted text
-        return cleaned_text
+            except Exception as e:
+                # If an error occurs (e.g., network issues), print the error and try the next API key
+                print(f"Error using API key {api_key}: {str(e)}")
 
-    except Exception as e:
-        # Handle errors (e.g., file not found, OCR issues)
-        print(f"Error in OCR: {str(e)}")
-        return "Error in extracting text from image."
+        # If both API keys fail, return a final error message
+        return "Error in extracting text from image using both API keys."
 
 
 
@@ -398,17 +426,15 @@ class QuestifyAI:
         self.translation_manager = TranslationManager()
         self.language_manager = LanguageManager()
         self.model_manager = ModelManager(self.api_key_manager)
+        # Fetch OCR API keys from Hugging Face Space secrets
+        self.ocr_api_keys = [
+            os.getenv("ocr_space_api_key1"),  # Get the first OCR API key from environment variable
+            os.getenv("ocr_space_api_key2")   # Get the second OCR API key from environment variable
+        ]
+        self.api_key_index = 0  # Start with the first API key
 
     def structure_question(self, question: str) -> str:
-        """
-        Organize user input by removing unwanted symbols and ensuring clarity.
-        
-        Args:
-            question (str): The original input question
-        
-        Returns:
-            str: A structured, cleaned-up question
-        """
+        """Organize user input by removing unwanted symbols and ensuring clarity."""
         question = re.sub(r"[#/*\\]", "", question)
         question = question.strip()
         return question
@@ -424,17 +450,7 @@ class QuestifyAI:
         input_language: str, 
         question_type: str
     ) -> Tuple[str, str, str]:
-        """
-        Process a question through translation and model response pipeline.
-        
-        Args:
-            question (str): The input question
-            input_language (str): Language of the input question
-            question_type (str): Type of question (general, math, job)
-        
-        Returns:
-            Tuple of (English Question, English Answer, Translated Answer)
-        """
+        """Process a question through translation and model response pipeline."""
         # Clean and structure the question
         question = self.structure_question(question)
         
@@ -461,7 +477,7 @@ class QuestifyAI:
 
     def ocr_from_image(self, image: Image.Image) -> str:
         """
-        Extract text from an uploaded image using OCR (Optical Character Recognition).
+        Extract text from an uploaded image using OCR.space API.
         
         Args:
             image (Image.Image): The PIL image object from which text will be extracted.
@@ -470,12 +486,49 @@ class QuestifyAI:
             str: The extracted text from the image.
         """
         try:
-            # Use pytesseract to extract text
-            extracted_text = pytesseract.image_to_string(image)
-            return extracted_text.strip()
+            api_url = "https://api.ocr.space/parse/image"
+            current_api_key = self.ocr_api_keys[self.api_key_index]
+            
+            # Open the image and send it to the API
+            img_byte_arr = io.BytesIO()
+            image.save(img_byte_arr, format="PNG")
+            img_byte_arr = img_byte_arr.getvalue()
+
+            files = {'file': ('image.png', img_byte_arr, 'image/png')}
+            data = {'apikey': current_api_key}
+            
+            # Send the request to OCR.space API
+            response = requests.post(api_url, files=files, data=data)
+            response.raise_for_status()  # Check for errors in the request
+            
+            # Parse the JSON response
+            result = response.json()
+            
+            # Check if the OCR was successful
+            if result.get("OCRExitCode") == 1:
+                # Extract text from the response
+                extracted_text = result["ParsedResults"][0]["ParsedText"]
+                return extracted_text.strip()
+            else:
+                # If OCR failed, try switching the API key
+                self.switch_api_key()
+                return "Error in extracting text from image."
+
+        except requests.exceptions.RequestException as e:
+            # Handle request errors
+            print(f"Error in OCR request: {str(e)}")
+            self.switch_api_key()
+            return "Error in extracting text from image."
         except Exception as e:
+            # Handle general errors
             print(f"Error in OCR: {str(e)}")
+            self.switch_api_key()
             return "Error in extracting text from image."
+        
+    def switch_api_key(self):
+        """Switch to the next OCR API key when the current one reaches its limit or fails."""
+        self.api_key_index = (self.api_key_index + 1) % len(self.ocr_api_keys)
+        print(f"Switched to API Key {self.api_key_index + 1}")
 
     def launch_ui(self):
         """Launch Gradio web interface for QuestifyAI."""