Update app.py

app.py

@@ -1,3 +1,6 @@
+import os
+import asyncio
+from asyncio import BoundedSemaphore
 from fastapi import FastAPI, File, UploadFile
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.middleware.gzip import GZipMiddleware
@@ -16,41 +19,47 @@ app.add_middleware(
     allow_methods=["*"],
     allow_headers=["*"]
 )
+app.add_middleware(GZipMiddleware, minimum_size=1000)
 
 # Initialize models once at startup
 ocr_model = ocr_predictor(pretrained=True)
-paddle_ocr = PaddleOCR(lang='en', use_angle_cls=True)
-
-def ocr_with_doctr(file):
-    text_output = ''
-    doc = DocumentFile.from_pdf(file)
-    ocr_model = ocr_predictor(pretrained=True)
-    result = ocr_model(doc)
-    for page in result.pages:
-        for block in page.blocks:
-            for line in block.lines:
-                text_output += " ".join([word.value for word in line.words]) + "\n"
-    return text_output
-
-def ocr_with_paddle(img):
-    finaltext = ''
-    result = paddle_ocr.ocr(img)
-    for i in range(len(result[0])):
-        text = result[0][i][1][0]
-        finaltext += ' ' + text
-    return finaltext
-
-def generate_text_from_image(img):
-    return ocr_with_paddle(img)
+paddle_ocr = PaddleOCR(lang='en', use_angle_cls=True, use_gpu=False)
+
+# Create a bounded semaphore to limit concurrent requests
+semaphore = BoundedSemaphore(10)
+
+async def ocr_with_doctr(file):
+    async with semaphore:
+        doc = DocumentFile.from_pdf(file)
+        result = ocr_model(doc)
+        text_output = ''
+        for page in result.pages:
+            for block in page.blocks:
+                for line in block.lines:
+                    text_output += " ".join([word.value for word in line.words]) + "\n"
+        return text_output
+
+async def ocr_with_paddle(img):
+    async with semaphore:
+        result = paddle_ocr.ocr(img)
+        finaltext = ''
+        for i in range(len(result[0])):
+            text = result[0][i][1][0]
+            finaltext +='' + text
+        return finaltext
+
+async def generate_text_from_image(img):
+    return await ocr_with_paddle(img)
 
 @app.post("/ocr/")
 async def perform_ocr(file: UploadFile = File(...)):
     file_bytes = await file.read()
     if file.filename.endswith('.pdf'):
-        text_output = ocr_with_doctr(io.BytesIO(file_bytes))
+        text_output = await ocr_with_doctr(io.BytesIO(file_bytes))
     else:
-        img = np.array(Image.open(io.BytesIO(file_bytes)))
-        text_output = generate_text_from_image(img)
+        img = Image.open(io.BytesIO(file_bytes))
+        img.thumbnail((1024, 1024))  # Reduce image size
+        text_output = await generate_text_from_image(img)
     return {"ocr_text": text_output}
 
 @app.get("/test/")