"""Isolated CPU TensorFlow worker for both official LAION classifier heads."""
import json
import sys
from pathlib import Path
import numpy as np
import tensorflow as tf

ROOT=Path('/tmp/vivarium-safety')
tf.config.threading.set_intra_op_parallelism_threads(8)
tf.config.threading.set_inter_op_parallelism_threads(2)
binary=tf.saved_model.load(str(ROOT/'canonical-head/clip_autokeras_binary_nsfw')).signatures['serving_default']
effnet=tf.saved_model.load(str(ROOT/'efficientnet-savedmodel')).signatures['serving_default']
print(json.dumps({'status':'ready'}),flush=True)
for line in sys.stdin:
    request=json.loads(line)
    if request.get('stop'):break
    data=np.load(request['input'])
    embeddings=tf.convert_to_tensor(data['embeddings'],tf.float64)
    clip=next(iter(binary(input_1=embeddings).values())).numpy().reshape(-1)
    scores=[]
    for key in ['original','foreground']:
        pixels=tf.convert_to_tensor(data[key],tf.float32)/255.
        pixels=tf.image.resize(pixels,[260,260],antialias=True)
        logits=next(iter(effnet(input_1=pixels).values()))
        scores.append(tf.nn.softmax(logits,axis=-1).numpy())
    np.savez(request['output'],clip=clip,efficientnet=np.concatenate(scores,axis=0))
    print(json.dumps({'status':'scored','samples':len(clip),'output':request['output']}),flush=True)