app.py

import gradio as gr
from transformers import AutoModel, AutoTokenizer
from PIL import Image
import torch

# Load the model
tokenizer = AutoTokenizer.from_pretrained("ucaslcl/GOT-OCR2_0", trust_remote_code=True)
model = AutoModel.from_pretrained(
    "ucaslcl/GOT-OCR2_0",
    trust_remote_code=True,
    low_cpu_mem_usage=True,
    use_safetensors=True,
    device_map="cuda",
    pad_token_id=tokenizer.eos_token_id
)
model = model.eval().cuda()

# Run inference
def run_ocr(image, task="Plain Text OCR", ocr_type="ocr", ocr_box="", ocr_color="red"):
    if isinstance(image, str):
        image = Image.open(image).convert("RGB")
    elif isinstance(image, Image.Image):
        image = image.convert("RGB")
    else:
        return "Invalid image input."
    
    result = model.chat(tokenizer, image, ocr_type=ocr_type)
    return result

# Gradio UI with API exposed
iface = gr.Interface(
    fn=run_ocr,
    inputs=[
        gr.Image(type="filepath", label="Image"),
        gr.Dropdown(choices=["Plain Text OCR", "Format Text OCR", "Fine-grained OCR (Box)", "Fine-grained OCR (Color)", "Multi-crop OCR", "Multi-page OCR"], value="Plain Text OCR", label="Task"),
        gr.Dropdown(choices=["ocr", "format"], value="ocr", label="OCR Type"),
        gr.Textbox(label="OCR Box", placeholder="Optional, e.g. [100,100,200,200]"),
        gr.Dropdown(choices=["red", "green", "blue"], value="red", label="OCR Color")
    ],
    outputs=gr.Textbox(label="OCR Output"),
    allow_flagging="never",
    allow_api=True   # ✅ This line enables API access!
)

iface.launch()