Commit ·
9047602
1
Parent(s): 56f4087
uP
Browse files
upload.py
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/usr/bin/env python3
|
| 2 |
+
from huggingface_hub import HfApi
|
| 3 |
+
from huggingface_hub import hf_hub_download
|
| 4 |
+
import huggingface_hub
|
| 5 |
+
from huggingface_hub import get_repo_discussions
|
| 6 |
+
from bs4 import BeautifulSoup
|
| 7 |
+
import json
|
| 8 |
+
|
| 9 |
+
# repo_id = "stabilityai/stable-diffusion"
|
| 10 |
+
repo_id = "huggingface-projects/diffuse-the-rest"
|
| 11 |
+
dataset_repo_id = "triple-t/dummy"
|
| 12 |
+
path = "/home/patrick/image_cache"
|
| 13 |
+
|
| 14 |
+
print("retrieve images 0...")
|
| 15 |
+
discussions_list = list(get_repo_discussions(repo_id=repo_id, repo_type="space"))
|
| 16 |
+
print("retrieve images 1...")
|
| 17 |
+
all_data = []
|
| 18 |
+
for i, disc in enumerate(discussions_list[:5]):
|
| 19 |
+
disc = huggingface_hub.get_discussion_details(repo_id=repo_id, repo_type="space", discussion_num=disc.num)
|
| 20 |
+
page = BeautifulSoup(disc.events[0]._event["data"]["latest"]["raw"])
|
| 21 |
+
image_urls = [link.get('src') for link in page.findAll('img')]
|
| 22 |
+
data = {
|
| 23 |
+
"discussion_number": i,
|
| 24 |
+
"data": {
|
| 25 |
+
"prompt": disc.title,
|
| 26 |
+
"images": image_urls,
|
| 27 |
+
}
|
| 28 |
+
}
|
| 29 |
+
if not image_urls:
|
| 30 |
+
continue
|
| 31 |
+
else:
|
| 32 |
+
all_data.append(data)
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
file_name = "_".join(repo_id.split("/")) + ".json"
|
| 36 |
+
api = HfApi()
|
| 37 |
+
|
| 38 |
+
path = hf_hub_download(repo_id=dataset_repo_id, filename=file_name, cache_dir=path, repo_type="dataset")
|
| 39 |
+
|
| 40 |
+
with open(path, "r") as f:
|
| 41 |
+
data = json.load(path)
|
| 42 |
+
|
| 43 |
+
data += all_data
|
| 44 |
+
import ipdb; ipdb.set_trace()
|
| 45 |
+
|
| 46 |
+
with open(path, "w") as f:
|
| 47 |
+
f.write(json.dumps(all_data, sort_keys=True, indent=4))
|
| 48 |
+
|
| 49 |
+
api.upload_file(
|
| 50 |
+
path_or_fileobj=path,
|
| 51 |
+
path_in_repo=file_name,
|
| 52 |
+
repo_id=dataset_repo_id,
|
| 53 |
+
repo_type="dataset",
|
| 54 |
+
)
|