main.py

import os
os.environ["HF_HOME"] = "/.cache"

import re
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM


model_dir_large = 'edithram23/Redaction_Personal_info_v1'
tokenizer_large = AutoTokenizer.from_pretrained(model_dir_large)
model_large = AutoModelForSeq2SeqLM.from_pretrained(model_dir_large)

def mask_generation(text,model=model_large,tokenizer=tokenizer_large):
    import re
    inputs = ["Mask Generation: " + text+'.']
    inputs = tokenizer(inputs, max_length=512, truncation=True, return_tensors="pt")
    output = model.generate(**inputs, num_beams=8, do_sample=True, max_length=512)
    decoded_output = tokenizer.batch_decode(output, skip_special_tokens=True)[0]
    predicted_title = decoded_output.strip()
    pattern = r'\[.*?\]'
    # Replace all occurrences of the pattern with [redacted]
    redacted_text = re.sub(pattern, '[redacted]', predicted_title)
    return redacted_text

from fastapi import FastAPI
import uvicorn

app = FastAPI()

@app.get("/")
async def hello():
    return {"msg" : "Live"}

@app.post("/mask")
async def mask_input(query):
    output = mask_generation(query)
    return {"data" : output}

if __name__ == '__main__':
    os.environ["HF_HOME"] = "/.cache"
    uvicorn.run("main:app", host="0.0.0.0", port=7860, reload=True, workers=1)