Update app.py
Browse files
app.py
CHANGED
|
@@ -1,68 +1,68 @@
|
|
| 1 |
-
from fastapi import FastAPI, UploadFile, File, HTTPException
|
| 2 |
-
from fastapi.responses import JSONResponse
|
| 3 |
-
import easyocr
|
| 4 |
-
import numpy as np
|
| 5 |
-
from PIL import Image
|
| 6 |
-
import io
|
| 7 |
-
import uvicorn
|
| 8 |
-
import os
|
| 9 |
-
|
| 10 |
-
app = FastAPI()
|
| 11 |
-
|
| 12 |
-
# Inicializar el lector OCR
|
| 13 |
-
reader = easyocr.Reader(['es', 'en'])
|
| 14 |
-
|
| 15 |
-
@app.post("/ocr")
|
| 16 |
-
async def extract_text(file: UploadFile = File(...)):
|
| 17 |
-
"""
|
| 18 |
-
Extrae texto de una imagen usando OCR
|
| 19 |
-
"""
|
| 20 |
-
try:
|
| 21 |
-
# Verificar que el archivo sea una imagen
|
| 22 |
-
if not file.content_type.startswith('image/'):
|
| 23 |
-
raise HTTPException(status_code=400, detail="El archivo debe ser una imagen")
|
| 24 |
-
|
| 25 |
-
# Leer el archivo
|
| 26 |
-
contents = await file.read()
|
| 27 |
-
|
| 28 |
-
# Convertir a PIL Image
|
| 29 |
-
image = Image.open(io.BytesIO(contents))
|
| 30 |
-
|
| 31 |
-
# Convertir a RGB si es necesario
|
| 32 |
-
if image.mode != 'RGB':
|
| 33 |
-
image = image.convert('RGB')
|
| 34 |
-
|
| 35 |
-
# Convertir a numpy array
|
| 36 |
-
image_array = np.array(image)
|
| 37 |
-
|
| 38 |
-
# Realizar OCR
|
| 39 |
-
results = reader.readtext(image_array)
|
| 40 |
-
|
| 41 |
-
# Formatear resultados
|
| 42 |
-
extracted_text = []
|
| 43 |
-
for (bbox, text, confidence) in results:
|
| 44 |
-
extracted_text.append({
|
| 45 |
-
"text": text,
|
| 46 |
-
"confidence": float(confidence),
|
| 47 |
-
"bbox": bbox
|
| 48 |
-
})
|
| 49 |
-
|
| 50 |
-
# Texto completo concatenado
|
| 51 |
-
full_text = " ".join([item["text"] for item in extracted_text])
|
| 52 |
-
|
| 53 |
-
return JSONResponse(content={
|
| 54 |
-
"success": True,
|
| 55 |
-
"full_text": full_text,
|
| 56 |
-
"detailed_results": extracted_text
|
| 57 |
-
})
|
| 58 |
-
|
| 59 |
-
except Exception as e:
|
| 60 |
-
raise HTTPException(status_code=500, detail=f"Error procesando la imagen: {str(e)}")
|
| 61 |
-
|
| 62 |
-
@app.get("/")
|
| 63 |
-
async def root():
|
| 64 |
-
return {"message": "OCR API funcionando", "endpoint": "/ocr"}
|
| 65 |
-
|
| 66 |
-
if __name__ == "__main__":
|
| 67 |
-
port = int(os.environ.get("PORT", 7860))
|
| 68 |
-
uvicorn.run(app, host="0.0.0.0", port=port)
|
|
|
|
| 1 |
+
from fastapi import FastAPI, UploadFile, File, HTTPException
|
| 2 |
+
from fastapi.responses import JSONResponse
|
| 3 |
+
import easyocr
|
| 4 |
+
import numpy as np
|
| 5 |
+
from PIL import Image
|
| 6 |
+
import io
|
| 7 |
+
import uvicorn
|
| 8 |
+
import os
|
| 9 |
+
|
| 10 |
+
app = FastAPI()
|
| 11 |
+
|
| 12 |
+
# Inicializar el lector OCR
|
| 13 |
+
reader = easyocr.Reader(['es', 'en'], model_storage_directory='/app/model', user_network_directory='/app/user_network', recog_network='custom')
|
| 14 |
+
|
| 15 |
+
@app.post("/ocr")
|
| 16 |
+
async def extract_text(file: UploadFile = File(...)):
|
| 17 |
+
"""
|
| 18 |
+
Extrae texto de una imagen usando OCR
|
| 19 |
+
"""
|
| 20 |
+
try:
|
| 21 |
+
# Verificar que el archivo sea una imagen
|
| 22 |
+
if not file.content_type.startswith('image/'):
|
| 23 |
+
raise HTTPException(status_code=400, detail="El archivo debe ser una imagen")
|
| 24 |
+
|
| 25 |
+
# Leer el archivo
|
| 26 |
+
contents = await file.read()
|
| 27 |
+
|
| 28 |
+
# Convertir a PIL Image
|
| 29 |
+
image = Image.open(io.BytesIO(contents))
|
| 30 |
+
|
| 31 |
+
# Convertir a RGB si es necesario
|
| 32 |
+
if image.mode != 'RGB':
|
| 33 |
+
image = image.convert('RGB')
|
| 34 |
+
|
| 35 |
+
# Convertir a numpy array
|
| 36 |
+
image_array = np.array(image)
|
| 37 |
+
|
| 38 |
+
# Realizar OCR
|
| 39 |
+
results = reader.readtext(image_array)
|
| 40 |
+
|
| 41 |
+
# Formatear resultados
|
| 42 |
+
extracted_text = []
|
| 43 |
+
for (bbox, text, confidence) in results:
|
| 44 |
+
extracted_text.append({
|
| 45 |
+
"text": text,
|
| 46 |
+
"confidence": float(confidence),
|
| 47 |
+
"bbox": bbox
|
| 48 |
+
})
|
| 49 |
+
|
| 50 |
+
# Texto completo concatenado
|
| 51 |
+
full_text = " ".join([item["text"] for item in extracted_text])
|
| 52 |
+
|
| 53 |
+
return JSONResponse(content={
|
| 54 |
+
"success": True,
|
| 55 |
+
"full_text": full_text,
|
| 56 |
+
"detailed_results": extracted_text
|
| 57 |
+
})
|
| 58 |
+
|
| 59 |
+
except Exception as e:
|
| 60 |
+
raise HTTPException(status_code=500, detail=f"Error procesando la imagen: {str(e)}")
|
| 61 |
+
|
| 62 |
+
@app.get("/")
|
| 63 |
+
async def root():
|
| 64 |
+
return {"message": "OCR API funcionando", "endpoint": "/ocr"}
|
| 65 |
+
|
| 66 |
+
if __name__ == "__main__":
|
| 67 |
+
port = int(os.environ.get("PORT", 7860))
|
| 68 |
+
uvicorn.run(app, host="0.0.0.0", port=port)
|