import json
from refer import REFER
import random
from pycocotools import mask as maskUtils
from tqdm import tqdm
def annToMask(mask_ann, h=None, w=None):
    if isinstance(mask_ann, list):
        
        rles = maskUtils.frPyObjects(mask_ann, h, w)
        rle = maskUtils.merge(rles)
    elif isinstance(mask_ann['counts'], list):
        # uncompressed RLE
        rle = maskUtils.frPyObjects(mask_ann, h, w)
    else:
        # rle
        rle = mask_ann
    mask = maskUtils.decode(rle)
    return mask


refer_api = REFER('/mnt/workspace/workgroup/yuanyq/code/LISA/dataset/refer_seg', 'refcocog', 'umd')
ref_ids_train = refer_api.getRefIds(split="val")
images_ids_train = refer_api.getImgIds(ref_ids=ref_ids_train)
refs_train = refer_api.loadRefs(ref_ids=ref_ids_train)
# loaded_images = refer_api.loadImgs(image_ids=images_ids_train)     
annotation = refer_api.Anns     
img2refs = {}
for ref in refs_train:
    image_id = ref["image_id"]
    img2refs[image_id] = img2refs.get(image_id, []) + [
        ref,
    ]

final_data = []
for idx in tqdm(img2refs):
    dic = {}
    data = img2refs[idx]
    img_idx = data[0]['image_id']
    image = refer_api.loadImgs(image_ids=img_idx)[0]    
    dic['image'] = 'refer_seg/images/mscoco/images/train2014/'+image['file_name']
    # dic['image'] = 'refer_seg/images/saiapr_tc-12/'+image['file_name']
    dic['height']= image['height']
    dic['width'] = image['width']
    cats = []
    masks = []
    for ann in data:
        ann_idx = ann['ann_id']
        annotation_ = annotation[ann_idx]
        cat_list = set()
        for cat in ann['sentences']:
            cat_list.add(cat['sent'])
        cat_list = list(cat_list)
        cats+=cat_list
        for i in range(len(cat_list)):
            masks.append(annotation_['segmentation'])   
            
    dic['cat'] = cats
    dic['masks'] = masks
    final_data.append(dic)
    
print(len(final_data))
with open('/mnt/workspace/workgroup/yuanyq/code/seg-llava/val/refcocog_val.json', 'w') as f:
    f.write(json.dumps(final_data))

# conversations.append({'from': 'human', 'value': random.choice(SHORT_QUESTION).format(class_name=text.lower())})
# conversations.append({'from': 'gpt', 'value': random.choice(ANSWER_LIST)})