Revision ae1dc1b14d1cab584f2ba6a18b0cb3aa00d7e429 authored by Kirill Sizov on 05 October 2023, 12:42:15 UTC, committed by GitHub on 05 October 2023, 12:42:15 UTC
When we export a task or project in CVAT format [our code
uses](https://github.com/opencv/cvat/blob/d497bb6a90dd3c9625c3c8e8f73278019e21983d/cvat/apps/dataset_manager/formats/cvat.py#L717)
the `group_by_frame` function. For tasks and for projects behaviour of
this function is different since for tasks [we perform sorting
frames](https://github.com/opencv/cvat/blob/d497bb6a90dd3c9625c3c8e8f73278019e21983d/cvat/apps/dataset_manager/bindings.py#L420-L421)
but for projects [we
don't](https://github.com/opencv/cvat/blob/d497bb6a90dd3c9625c3c8e8f73278019e21983d/cvat/apps/dataset_manager/bindings.py#L1152-L1154).
As a result, we can have a situation when our annotation file for the
project will have a list of images in unsorted form. So this PR adds
sorting frames during project export as well.
1 parent bfc86ae
History
File Mode Size
requirements
src
.gitignore -rw-r--r-- 807 bytes
MANIFEST.in -rw-r--r-- 47 bytes
README.md -rw-r--r-- 1.9 KB
developer_guide.md -rw-r--r-- 270 bytes
pyproject.toml -rw-r--r-- 383 bytes
setup.py -rw-r--r-- 1.9 KB

README.md

back to top