Upload 3 files

Dockerfile

@@ -0,0 +1,43 @@
+FROM python:3.10-slim
+
+# Install system dependencies
+
+RUN apt-get update && apt-get install -y \
+tesseract-ocr \
+tesseract-ocr-eng tesseract-ocr-hin tesseract-ocr-ara tesseract-ocr-spa \
+tesseract-ocr-ita tesseract-ocr-fra tesseract-ocr-rus \
+poppler-utils \
+libopencv-dev \
+&& rm -rf /var/lib/apt/lists/\*
+
+# Create working directory
+
+RUN mkdir -p /app WORKDIR /app
+
+# Copy files
+
+COPY . /app/
+
+# Debug: List files
+
+RUN ls -la /app
+
+# Verify files
+
+RUN if \[ ! -f /app/requirements.txt \]; then echo "requirements.txt not found" && exit 1; fi RUN if \[ ! -f /app/ocr_api.py \]; then echo "ocr_api.py not found" && exit 1; fi
+
+# Install Python dependencies
+
+RUN pip install --no-cache-dir -r /app/requirements.txt
+
+# Create non-root user
+
+RUN useradd -m appuser USER appuser
+
+# Expose port
+
+EXPOSE 8000
+
+# Run the app
+
+CMD \["uvicorn", "ocr_api:app", "--host", "0.0.0.0", "--port", "8000"\]

app.py

@@ -0,0 +1,173 @@
+from fastapi import FastAPI, File, UploadFile, HTTPException
+import pytesseract
+import cv2
+import os
+from PIL import Image
+import json
+import unicodedata
+from pdf2image import convert_from_bytes
+from pypdf import PdfReader
+import numpy as np
+from typing import List
+import io
+import logging
+import time
+import asyncio
+import psutil
+import cachetools
+import hashlib
+
+app = FastAPI()
+
+# Configure logging
+logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
+logger = logging.getLogger(__name__)
+
+# Set Tesseract path
+pytesseract.pytesseract.tesseract_cmd = "/usr/bin/tesseract"
+
+# In-memory cache for OCR results (1-hour TTL)
+cache = cachetools.TTLCache(maxsize=100, ttl=3600)
+
+def log_memory_usage():
+    """Log current memory usage."""
+    process = psutil.Process()
+    mem_info = process.memory_info()
+    return f"Memory usage: {mem_info.rss / 1024 / 1024:.2f} MB"
+
+def get_file_hash(pdf_bytes):
+    """Generate MD5 hash of PDF content."""
+    return hashlib.md5(pdf_bytes).hexdigest()
+
+async def process_page(img, page_idx):
+    """Process a single PDF page with OCR."""
+    start_time = time.time()
+    logger.info(f"Starting OCR for page {page_idx}, {log_memory_usage()}")
+    try:
+        img_cv = cv2.cvtColor(np.array(img), cv2.COLOR_RGB2BGR)
+        gray = cv2.cvtColor(img_cv, cv2.COLOR_BGR2GRAY)
+        img_pil = Image.fromarray(cv2.cvtColor(gray, cv2.COLOR_GRAY2RGB))
+        custom_config = r'--oem 1 --psm 6 -l eng+hin+ara+spa+ita+fra+rus'
+        page_text = pytesseract.image_to_string(img_pil, config=custom_config)
+        logger.info(f"Completed OCR for page {page_idx}, took {time.time() - start_time:.2f} seconds, {log_memory_usage()}")
+        return page_text + "\n"
+    except Exception as e:
+        logger.error(f"OCR failed for page {page_idx}: {str(e)}, {log_memory_usage()}")
+        return ""
+
+@app.post("/ocr")
+async def extract_text(files: List[UploadFile] = File(...)):
+    output_json = {
+        "success": True,
+        "message": "",
+        "data": []
+    }
+    success_count = 0
+    fail_count = 0
+
+    logger.info(f"Starting OCR for {len(files)} files, {log_memory_usage()}")
+
+    for file in files:
+        total_start_time = time.time()
+        logger.info(f"Processing file: {file.filename}, {log_memory_usage()}")
+
+        if not file.filename.lower().endswith('.pdf'):
+            fail_count += 1
+            output_json["data"].append({
+                "filename": file.filename,
+                "raw_text": "",
+                "error": "File is not a PDF"
+            })
+            logger.error(f"File {file.filename} is not a PDF")
+            continue
+
+        # Read PDF into memory
+        try:
+            pdf_start_time = time.time()
+            pdf_bytes = await file.read()
+            pdf_stream = io.BytesIO(pdf_bytes)
+            file_hash = get_file_hash(pdf_bytes)
+            logger.info(f"Read PDF {file.filename}, took {time.time() - pdf_start_time:.2f} seconds, size: {len(pdf_bytes)/1024:.2f} KB, {log_memory_usage()}")
+        except Exception as e:
+            fail_count += 1
+            output_json["data"].append({
+                "filename": file.filename,
+                "raw_text": "",
+                "error": f"Failed to read PDF: {str(e)}"
+            })
+            logger.error(f"Failed to read PDF {file.filename}: {str(e)}, {log_memory_usage()}")
+            continue
+
+        # Check cache
+        if file_hash in cache:
+            success_count += 1
+            output_json["data"].append({
+                "filename": file.filename,
+                "raw_text": cache[file_hash],
+                "error": ""
+            })
+            logger.info(f"Cache hit for {file.filename}, {log_memory_usage()}")
+            continue
+
+        # Initialize raw text
+        raw_text = ""
+        # Try extracting embedded text
+        try:
+            extract_start_time = time.time()
+            reader = PdfReader(pdf_stream)
+            for page in reader.pages:
+                text = page.extract_text()
+                if text:
+                    raw_text += text + "\n"
+            logger.info(f"Embedded text extraction for {file.filename}, took {time.time() - extract_start_time:.2f} seconds, text length: {len(raw_text)}, {log_memory_usage()}")
+        except Exception as e:
+            logger.warning(f"Embedded text extraction failed for {file.filename}: {str(e)}, {log_memory_usage()}")
+
+        # If no embedded text, perform OCR
+        if not raw_text.strip():
+            try:
+                convert_start_time = time.time()
+                images = convert_from_bytes(pdf_bytes, poppler_path="/usr/local/bin", dpi=100)
+                logger.info(f"PDF to images conversion for {file.filename}, {len(images)} pages, took {time.time() - convert_start_time:.2f} seconds, {log_memory_usage()}")
+
+                ocr_start_time = time.time()
+                tasks = [process_page(img, i) for i, img in enumerate(images)]
+                page_texts = await asyncio.gather(*tasks)
+                raw_text = "".join(page_texts)
+                logger.info(f"Total OCR for {file.filename}, took {time.time() - ocr_start_time:.2f} seconds, text length: {len(raw_text)}, {log_memory_usage()}")
+            except Exception as e:
+                fail_count += 1
+                output_json["data"].append({
+                    "filename": file.filename,
+                    "raw_text": "",
+                    "error": f"OCR failed: {str(e)}"
+                })
+                logger.error(f"OCR failed for {file.filename}: {str(e)}, {log_memory_usage()}")
+                continue
+
+        # Normalize text
+        try:
+            normalize_start_time = time.time()
+            raw_text = unicodedata.normalize('NFKC', raw_text)
+            raw_text = raw_text.encode().decode('utf-8')
+            logger.info(f"Text normalization for {file.filename}, took {time.time() - normalize_start_time:.2f} seconds, text length: {len(raw_text)}, {log_memory_usage()}")
+        except Exception as e:
+            logger.warning(f"Text normalization failed for {file.filename}: {str(e)}, {log_memory_usage()}")
+
+        # Cache and store result
+        cache[file_hash] = raw_text
+        success_count += 1
+        output_json["data"].append({
+            "filename": file.filename,
+            "raw_text": raw_text,
+            "error": ""
+        })
+
+        logger.info(f"Total processing for {file.filename}, took {time.time() - total_start_time:.2f} seconds, {log_memory_usage()}")
+
+    output_json["message"] = f"Processed {len(files)} PDFs. {success_count} succeeded, {fail_count} failed."
+    if fail_count > 0 and success_count == 0:
+        output_json["success"] = False
+
+    logger.info(f"Completed OCR for {len(files)} files, {success_count} succeeded, {fail_count} failed, {log_memory_usage()}")
+    return output_json

requirements.txt

@@ -0,0 +1,10 @@
+fastapi==0.115.0 
+uvicorn==0.30.6 
+pytesseract==0.3.13 
+opencv-python==4.10.0.84 
+pillow==10.4.0 
+pdf2image==1.17.0 
+pypdf==5.0.1 
+numpy==1.26.4 
+psutil==6.0.0 
+cachetools==5.5.0