File size: 13,783 Bytes
3a6f18e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 |
import json
import argparse
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument('--review_data_path', type=str, default='./data')
parser.add_argument('--meta_data_path', type=str, default='./data')
parser.add_argument('--save_dir', type=str, default='./data/dataset')
parser.add_argument('--dataset', type=str, default='dataset')
return parser.parse_args()
args = parse_args()
import os
if not os.path.exists(args.save_dir + '/' + args.dataset):
os.mkdir(args.save_dir + '/' + args.dataset)
# /datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments
''' Extract interaction sequence '''
inters = {}
with open(args.review_data_path, 'r', encoding = 'utf-8') as file:
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments.json', 'r', encoding = 'utf-8') as file:
for line in file:
element = json.loads(line)
if element['reviewerID'] in inters:
inters[element['reviewerID']].append({'time': element['unixReviewTime'], 'item': element['asin']})
else:
inters[element['reviewerID']] = [{'time': element['unixReviewTime'], 'item': element['asin']}]
# Filter out sequence shorter than 5
filtered_inters = {key: value for key, value in inters.items() if len(value) > 4}
final_inters = {}
for key, value in filtered_inters.items():
# Sort items according to time
value.sort(key = lambda x: x['time'])
final_inters[key] = [x['item'] for x in value]
# Save interaction
with open(args.save_dir + '/' + args.dataset + '/' + args.dataset + '.inters.json', 'w', encoding = 'utf-8') as f:
json.dump(final_inters, f, ensure_ascii = False, indent = 4)
''' Extract user review '''
reviews = {}
with open(args.review_data_path, 'r', encoding = 'utf-8') as file:
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments.json', 'r', encoding = 'utf-8') as file:
for line in file:
element = json.loads(line)
if len(element.get('reviewText', '')) > 0:
reviews[element['reviewerID']] = {element['asin']: element['reviewText']}
else:
continue
with open(args.save_dir + '/' + args.dataset + '/' + args.dataset + '.reviews.json', 'w', encoding = 'utf-8') as f:
json.dump(reviews, f, ensure_ascii = False, indent = 4)
''' Extract item features '''
features = {}
with open(args.meta_data_path, 'r', encoding = 'utf-8') as file:
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/meta_Musical_Instruments.json', 'r', encoding = 'utf-8') as file:
for line in file:
element = json.loads(line)
if len(element.get('title', '')) > 0 and len(element.get('description', '')) > 0 and len(element.get('imageURL', '')) > 0 and len(element.get('imageURLHighRes', '')) > 0:
features[element['asin']] = {
'title': element['title'],
'description': element['description'],
'image': element['imageURL'],
'imageH': element['imageURLHighRes']}
else:
continue
with open(args.save_dir + '/' + args.dataset + '/' + args.dataset + '.features.json', 'w', encoding = 'utf-8') as f:
json.dump(features, f, ensure_ascii = False, indent = 4)
''' Filter out modality deficiency '''
with open(args.save_dir + '/' + args.dataset + '/' + args.dataset + '.inters.json', 'r', encoding = 'utf-8') as file:
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.inters.json', 'r', encoding = 'utf-8') as file:
inters = json.load(file)
with open(args.save_dir + '/' + args.dataset + '/' + args.dataset + '.features.json', 'r', encoding = 'utf-8') as file:
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.features.json', 'r', encoding = 'utf-8') as file:
features = json.load(file)
# Remove items lacking modalities
full_inters = {}
number = 1
for key, value in inters.items():
full_value = []
for item in value:
if features.get(item, None) is None:
continue
elif len(features[item]['title']) == 0:
continue
elif len(' '.join(features[item]['description'])[:-1]) == 0:
continue
else:
try:
open_image = Image.open(requests.get(features[item]['imageH'][0], stream = True).raw)
except PIL.UnidentifiedImageError:
print('Caught UnidentifiedImageError.')
continue
except Exception as e:
print('Caught Exception.')
continue
full_value.append(item)
if len(full_value) > 4:
print('Interaction', number)
full_inters[key] = full_value
number += 1
else:
continue
# Save full interaction sequences
with open(args.save_dir + '/' + args.dataset + '/' + args.dataset + '.inters.full.json', 'w', encoding = 'utf-8') as f:
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.inters.complete.json', 'w', encoding = 'utf-8') as f:
json.dump(full_inters, f, ensure_ascii = False, indent = 4)
# inters = {}
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments.json', 'r', encoding = 'utf-8') as file:
# for line in file:
# element = json.loads(line)
# if element['reviewerID'] in inters:
# inters[element['reviewerID']].append({'time': element['unixReviewTime'], 'item': element['asin']})
# else:
# inters[element['reviewerID']] = [{'time': element['unixReviewTime'], 'item': element['asin']}]
# filtered_inters = {key: value for key, value in inters.items() if len(value) > 4}
# final_inters = {}
# for key, value in filtered_inters.items():
# value.sort(key = lambda x: x['time'])
# final_inters[key] = [x['item'] for x in value]
# # print(len(final_inters))
# # length: 40673
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.inters.json', 'w', encoding = 'utf-8') as f:
# json.dump(final_inters, f, ensure_ascii = False, indent = 4)
# reviews = {}
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments.json', 'r', encoding = 'utf-8') as file:
# for line in file:
# element = json.loads(line)
# if len(element.get('reviewText', '')) > 0:
# reviews[element['reviewerID']] = {element['asin']: element['reviewText']}
# else:
# continue
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.reviews.json', 'w', encoding = 'utf-8') as f:
# json.dump(reviews, f, ensure_ascii = False, indent = 4)
# features = {}
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/meta_Musical_Instruments.json', 'r', encoding = 'utf-8') as file:
# for line in file:
# element = json.loads(line)
# if len(element.get('title', '')) > 0 and len(element.get('description', '')) > 0 and len(element.get('imageURL', '')) > 0 and len(element.get('imageURLHighRes', '')) > 0:
# features[element['asin']] = {
# 'title': element['title'],
# 'description': element['description'],
# 'image': element['imageURL'],
# 'imageH': element['imageURLHighRes']}
# else:
# continue
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.features.json', 'w', encoding = 'utf-8') as f:
# json.dump(features, f, ensure_ascii = False, indent = 4)
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.inters.json', 'r', encoding = 'utf-8') as file:
# inters = json.load(file)
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.features.json', 'r', encoding = 'utf-8') as file:
# features = json.load(file)
# full_inters = {}
# for key, value in inters.items():
# full_value = []
# for item in value:
# if features.get(item, None) is None:
# continue
# else:
# full_value.append(item)
# if len(full_value) > 4:
# full_inters[key] = full_value
# else:
# continue
# with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.inters.full.json', 'w', encoding = 'utf-8') as f:
# json.dump(full_inters, f, ensure_ascii = False, indent = 4)
import json
both, text, image, num = 0, 0, 0, 0
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/meta_Toys_and_Games.json', 'r', encoding = 'utf-8') as file:
for line in file:
element = json.loads(line)
text_image = text + image
num += 1
if len(element.get('title', '')) + len(element.get('description', '')) == 0:
text += 1
if len(element.get('imageURL', '')) + len(element.get('imageURLHighRes', '')) == 0:
image += 1
both = both + (text + image - text_image) // 2
print('item:', num)
print('both:', both)
print('text:', text)
print('image', image)
for value in full_inters.values():
all_items += value
items = list(set(all_items))
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.items.txt', 'w') as f:
for item in items:
f.write(item + '\n')
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.items.txt', 'r') as f:
items = f.read()
items = items.split('\n')
item_order_mapping = {}
item_order_reverse = {}
item_order = 0
for item in items:
item_order_mapping[item] = str(item_order)
item_order_reverse[str(item_order)] = item
item_order += 1
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.item_order_mapping.json', 'w', encoding = 'utf-8') as f:
json.dump(item_order_mapping, f, ensure_ascii = False, indent = 4)
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.item_order_reverse.json', 'w', encoding = 'utf-8') as f:
json.dump(item_order_reverse, f, ensure_ascii = False, indent = 4)
import json
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.inters.full.json', 'r', encoding = 'utf-8') as file:
inters = json.load(file)
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.features.json', 'r', encoding = 'utf-8') as file:
features = json.load(file)
all_items = []
for value in inters.values():
all_items += value
items = list(set(all_items))
full_features = {}
for item in items:
full_features[item] = features[item]
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.features.complete.json', 'w', encoding = 'utf-8') as f:
json.dump(full_features, f, ensure_ascii = False, indent = 4)
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.features.complete.json', 'r', encoding = 'utf-8') as f:
features = json.load(f)
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.item_order_mapping.json', 'r', encoding = 'utf-8') as f:
item_order_mapping = json.load(f)
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.inters.complete.json', 'r', encoding = 'utf-8') as f:
inters = json.load(f)
inter_reorder = {}
index = 0
for inter in inters:
inter_reorder[str(index)] = []
for item in inters[inter]:
inter_reorder[str(index)].append(int(item_order_mapping[item]))
index += 1
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.inters.numerical.json', 'w', encoding = 'utf-8') as f:
json.dump(inter_reorder, f, ensure_ascii = False, indent = 4)
feature_reorder = {}
for item in features:
feature_reorder[item_order_mapping[item]] = features[item]
with open('/datain/v-yinju/LLMBased_Multimodal_RS/Data/Musical_Instruments/Musical_Instruments.features.numerical.json', 'w', encoding = 'utf-8') as f:
json.dump(feature_reorder, f, ensure_ascii = False, indent = 4)
with open('/datain/v-yinju/rqvae-zzx/data/Instruments/Instruments.inter.json', 'r', encoding = 'utf-8') as f:
inters = json.load(f)
with open('/datain/v-yinju/rqvae-zzx/data/Instruments/Instruments.index.json', 'r', encoding = 'utf-8') as f:
indices = json.load(f)
with open('/datain/v-yinju/rqvae-zzx/data/Instruments/Instruments.item.json', 'r', encoding = 'utf-8') as f:
features = json.load(f)
for uid, inter in inters:
remapped_inter = []
for item in inter:
if str(item) in indices:
remapped_inter.append(''.join(indices[str(item)]))
if len(remapped_inter) >= 5:
remapped_inters[uid] = remapped_inter
from datasets import load_dataset, Dataset
amazon = load_dataset("pandas", data_files= '/datain/v-yinju/rqvae-zzx/data/Instruments/Instruments.inter.test.pkl')['train']
SYSTEM_PROMPT = """
Below is an instruction that describes a task. Write a response that appropriately completes the request. Respond in the following format:
<reasoning>
...
</reasoning>
<answer>
...
</answer>
"""
Amazon = amazon.map(lambda x: {
'prompt': [
{'role': 'system', 'content': SYSTEM_PROMPT},
{'role': 'user', 'content': x['instruction']}
],
'answer': x['response']
}) |