AisingioroHao0 commited on
Commit ·
c860465
1
Parent(s): 4031149
add code
Browse files
comics.py
ADDED
|
@@ -0,0 +1,97 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from logging import config
|
| 2 |
+
import datasets
|
| 3 |
+
import os
|
| 4 |
+
from PIL import Image
|
| 5 |
+
import re
|
| 6 |
+
|
| 7 |
+
|
| 8 |
+
class ComicsConfig(datasets.BuilderConfig):
|
| 9 |
+
def __init__(self, **kwargs):
|
| 10 |
+
super(ComicsConfig, self).__init__(**kwargs)
|
| 11 |
+
|
| 12 |
+
|
| 13 |
+
class Comics(datasets.GeneratorBasedBuilder):
|
| 14 |
+
def _split_generators(self, dl_manager: datasets.DownloadManager):
|
| 15 |
+
return [
|
| 16 |
+
datasets.SplitGenerator(
|
| 17 |
+
name=datasets.Split.TRAIN,
|
| 18 |
+
gen_kwargs={},
|
| 19 |
+
)
|
| 20 |
+
]
|
| 21 |
+
|
| 22 |
+
BUILDER_CONFIGS = [
|
| 23 |
+
ComicsConfig(
|
| 24 |
+
name="chapter split",
|
| 25 |
+
description="comic dataset,item is a chapter",
|
| 26 |
+
),
|
| 27 |
+
ComicsConfig(
|
| 28 |
+
name="page split",
|
| 29 |
+
description="comic dataset,item is a page",
|
| 30 |
+
),
|
| 31 |
+
]
|
| 32 |
+
|
| 33 |
+
def _info(self):
|
| 34 |
+
if self.config.name == "page split":
|
| 35 |
+
return datasets.Features(
|
| 36 |
+
{
|
| 37 |
+
"comic_name": datasets.Value("string"),
|
| 38 |
+
"chapter_name": datasets.Value("string"),
|
| 39 |
+
"cover_image": datasets.features.Image(),
|
| 40 |
+
"image": datasets.features.Image(),
|
| 41 |
+
"prompt": datasets.Value("string"),
|
| 42 |
+
}
|
| 43 |
+
)
|
| 44 |
+
else:
|
| 45 |
+
return datasets.Features(
|
| 46 |
+
{
|
| 47 |
+
"comic_name": datasets.Value("string"),
|
| 48 |
+
"chapter_name": datasets.Value("string"),
|
| 49 |
+
"cover_image": datasets.features.Image(),
|
| 50 |
+
"images": datasets.features.Sequence(
|
| 51 |
+
{
|
| 52 |
+
"idx": datasets.Value("int32"),
|
| 53 |
+
"image": datasets.features.Image(),
|
| 54 |
+
}
|
| 55 |
+
),
|
| 56 |
+
"prompt": datasets.Value("string"),
|
| 57 |
+
}
|
| 58 |
+
)
|
| 59 |
+
|
| 60 |
+
def _generate_examples(self):
|
| 61 |
+
data_path = os.path.join(self.config.data_dir, "data")
|
| 62 |
+
for comic_name in os.listdir(data_path):
|
| 63 |
+
comic_path = os.path.join(data_path, comic_name)
|
| 64 |
+
cover_image = (
|
| 65 |
+
Image.open(os.path.join(comic_path, "cover.jpg"))
|
| 66 |
+
if os.path.exists(os.path.join(comic_path, "cover.jpg"))
|
| 67 |
+
else None
|
| 68 |
+
)
|
| 69 |
+
|
| 70 |
+
for chapter_name in os.listdir(comic_path):
|
| 71 |
+
chapter_path = os.path.join(comic_path, chapter_name)
|
| 72 |
+
if os.path.isfile(chapter_path):
|
| 73 |
+
continue
|
| 74 |
+
pages = list()
|
| 75 |
+
for idx, page_name in enumerate(os.listdir(chapter_path)):
|
| 76 |
+
page_path = os.path.join(chapter_path, page_name)
|
| 77 |
+
pages.append({"idx": idx, "image": Image.open(page_path)})
|
| 78 |
+
if self.config.name == "page split":
|
| 79 |
+
for page in pages:
|
| 80 |
+
yield comic_name + ":" + chapter_name + ":" + str(
|
| 81 |
+
page["idx"]
|
| 82 |
+
), {
|
| 83 |
+
"comic_name": comic_name,
|
| 84 |
+
"chapter_name": chapter_name,
|
| 85 |
+
"cover_image": cover_image,
|
| 86 |
+
"image": page["image"],
|
| 87 |
+
"idx": page["idx"],
|
| 88 |
+
"prompt": f"a image in the syle of manga.{comic_name}.{chapter_name}.{page['idx']}",
|
| 89 |
+
}
|
| 90 |
+
else:
|
| 91 |
+
yield comic_name + ":" + chapter_name + ":", {
|
| 92 |
+
"comic_name": comic_name,
|
| 93 |
+
"chapter_name": chapter_name,
|
| 94 |
+
"cover_image": cover_image,
|
| 95 |
+
"images": pages,
|
| 96 |
+
"prompt": f"a image in the syle of manga.{comic_name}.{chapter_name}",
|
| 97 |
+
}
|
upload.py
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import subprocess
|
| 3 |
+
from tqdm import tqdm
|
| 4 |
+
import git
|
| 5 |
+
|
| 6 |
+
repo_path = "."
|
| 7 |
+
os.chdir(repo_path)
|
| 8 |
+
repo = git.Repo(repo_path)
|
| 9 |
+
repo.index.commit("add images")
|
| 10 |
+
while True:
|
| 11 |
+
try:
|
| 12 |
+
repo.remote().push()
|
| 13 |
+
except Exception as e:
|
| 14 |
+
print(str(e))
|
| 15 |
+
else:
|
| 16 |
+
break
|
| 17 |
+
cnt = 0
|
| 18 |
+
for file_path in tqdm(repo.untracked_files):
|
| 19 |
+
if os.path.isfile(file_path):
|
| 20 |
+
repo.git().add([file_path])
|
| 21 |
+
cnt += 1
|
| 22 |
+
if cnt == 100:
|
| 23 |
+
repo.index.commit("add images")
|
| 24 |
+
while True:
|
| 25 |
+
try:
|
| 26 |
+
repo.remote().push()
|
| 27 |
+
except Exception as e:
|
| 28 |
+
print(str(e))
|
| 29 |
+
else:
|
| 30 |
+
break
|
| 31 |
+
cnt = 0
|
| 32 |
+
if cnt:
|
| 33 |
+
repo.index.commit("add images")
|
| 34 |
+
repo.remote().push()
|