Kurt Stolle commited on
Commit
7f95801
·
1 Parent(s): 20e7eb7

Added manifest and download/prepare script

Browse files
Files changed (2) hide show
  1. manifest.csv +0 -0
  2. scripts/download.py +159 -0
manifest.csv ADDED
The diff for this file is too large to render. See raw diff
 
scripts/download.py ADDED
@@ -0,0 +1,159 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python
2
+ """
3
+ Downloads and extracts additional data from the Cityscapes dataset, moving the
4
+ required files to (existing) data directory.
5
+ """
6
+
7
+ import tempfile
8
+ from tqdm import tqdm
9
+ import os
10
+ import re
11
+ from pathlib import Path
12
+ import shutil
13
+ import zipfile
14
+ import argparse
15
+ import pandas as pd
16
+
17
+ from cityscapesscripts.download import downloader
18
+
19
+ PACKAGE_TO_VARIABLE = {
20
+ "vehicle_sequence": "vehicle",
21
+ "timestamp_sequence": "timestamp",
22
+ "leftImg8bit_sequence_trainvaltest": "image",
23
+ "camera_trainvaltest": "camera",
24
+ }
25
+
26
+
27
+ def parse_args() -> argparse.Namespace:
28
+ parser = argparse.ArgumentParser(
29
+ description="Download and extract the Cityscapes dataset."
30
+ )
31
+
32
+ parser.add_argument(
33
+ "downloads_dir",
34
+ type=Path,
35
+ help="Path to the directory where ZIP files are/will be stored (e.g., 'downloads').",
36
+ )
37
+ parser.add_argument(
38
+ "data_dir",
39
+ type=Path,
40
+ help="Path to the directory where extracted files should be moved (e.g., 'data').",
41
+ )
42
+ parser.add_argument(
43
+ "manifest_file",
44
+ type=Path,
45
+ help="Path to the manifest file (e.g., 'manifest.csv').",
46
+ )
47
+ return parser.parse_args()
48
+
49
+
50
+ def read_manifest(manifest_file: str) -> pd.DataFrame:
51
+ """
52
+ Read the manifest file and return a DataFrame.
53
+ """
54
+ df = pd.read_csv(manifest_file, index_col="primary_key")
55
+ assert "split" in df.columns, "Missing 'split' column in manifest."
56
+ assert "sequence" in df.columns, "Missing 'sequence' column in manifest."
57
+ assert "frame" in df.columns, "Missing 'frame' column in manifest."
58
+ return df
59
+
60
+
61
+ def download_files(downloads_dir: str, zip_files: list[str]):
62
+ if not zip_files:
63
+ print("No files to download.")
64
+ return
65
+
66
+ session = downloader.login()
67
+
68
+ downloader.download_packages(
69
+ session=session,
70
+ package_names=zip_files,
71
+ destination_path=downloads_dir,
72
+ resume=True,
73
+ )
74
+
75
+
76
+ def unzip_and_move(
77
+ pkg: str,
78
+ downloads_dir: Path,
79
+ data_dir: Path,
80
+ manifest: pd.DataFrame,
81
+ source_split: str,
82
+ ):
83
+ """
84
+ Unzip, rename, and move the requested split files for a single package.
85
+ """
86
+ zip_path = downloads_dir / f"{pkg}.zip"
87
+ pkg_type = PACKAGE_TO_VARIABLE[pkg]
88
+
89
+ re_name = re.compile(r"([^_]*_[^_]*_[^_]*)_.*")
90
+
91
+ if not zip_path.is_file():
92
+ print(f"Warning: ZIP file not found => {zip_path}. Skipping...")
93
+ return
94
+
95
+ print(f"Processing: {zip_path}")
96
+
97
+ with tempfile.TemporaryDirectory() as td, zipfile.ZipFile(zip_path, "r") as zf:
98
+ names = zf.namelist()
99
+ for member in tqdm(names):
100
+ if f"/{source_split}/" not in member:
101
+ continue
102
+
103
+ # Extract the file and get its new path in the temporary directory
104
+ res_path = Path(zf.extract(member, td))
105
+
106
+ if res_path.is_dir():
107
+ continue
108
+
109
+ # Read the manifest to find the specification of this sample
110
+ primary_key = re_name.sub(r"\1", res_path.stem)
111
+ sample = manifest.loc[primary_key]
112
+
113
+ # New name is: split/<sequence>/<frame>.<type>.<ext>
114
+ new_path = (
115
+ data_dir
116
+ / sample["split"]
117
+ / "{:06d}".format(sample["sequence"])
118
+ / "{:06d}.{:s}{:s}".format(sample["frame"], pkg_type, res_path.suffix)
119
+ )
120
+ new_path.parent.mkdir(parents=True, exist_ok=True)
121
+
122
+ shutil.move(res_path, new_path)
123
+
124
+
125
+ def main():
126
+ args = parse_args()
127
+
128
+ # Create directories if not exists
129
+ for dir in (args.downloads_dir, args.data_dir):
130
+ os.makedirs(dir, exist_ok=True)
131
+
132
+ # Read the manifest file
133
+ manifest = read_manifest(args.manifest_file)
134
+
135
+ # Prepare list of package zip files
136
+ packages = list(PACKAGE_TO_VARIABLE.keys())
137
+ zip_files = [f"{pkg}.zip" for pkg in packages]
138
+
139
+ # Filter out existing ZIPs
140
+ download_list = []
141
+ for zf in zip_files:
142
+ zf_path = os.path.join(args.downloads_dir, zf)
143
+ if os.path.isfile(zf_path):
144
+ print(f"Already downloaded: {zf_path}")
145
+ else:
146
+ download_list.append(zf)
147
+
148
+ # Download missing ZIP files using csDownload directly
149
+ download_files(args.downloads_dir, download_list)
150
+
151
+ # Unzip, rename, and move each package
152
+ for pkg in packages:
153
+ unzip_and_move(pkg, args.downloads_dir, args.data_dir, manifest, "val")
154
+
155
+ print(f"Successfully processed: {packages}")
156
+
157
+
158
+ if __name__ == "__main__":
159
+ main()