HumanSignal / HumanSignal/label-studio

UTF-8 caracter in filename

Open
#1,266 9 comments 0 reactions 1 assignee Claimed by @makseq View on GitHub
backend export problem
Dominant language
TypeScript
Stars
28.3k
Forks
3.7k
Avg merge
14h
Merged PRs (30d)
15

Description

**Describe the bug**
I'm using label-studio to annotate data for yolov5 training.

**To Reproduce**
After annotation I do an export in YOLO format which was working great until I started a new project. This time I don't get the same count of labels and images (252 images and 266 label text files after export).

**Expected behavior**
Same number of images and labels in YOLO export, with filename like "Capture_décran" or "Capture_d%C3%A9cran" as long as they match.

**Describe the bug**
I tracked down the problem and all the missing image files are all containing an "é" (which is very common in french).

The original image file name are starting by "Capture_décran" (which mean "screenshot") but the exported label text files are all "Capture_d%C3%A9cran"
It seems that the "é" caracter is causing the problem because C3A9 is the UTF-8 hex for it.

It look like this in Label Studio:
![Capture d’écran du 2021-08-10 09-29-07](https://user-images.githubusercontent.com/43412403/128826485-da83218d-2b4e-4106-af28-2b67224e283d.png)

But for the first one the exported .txt file is called "Capture_d%C3%A9cran_de_2021-06-07_12-05-52.txt"

**Environment:**
- OS: Ubuntu 20.04.2 LTS
- Label Studio Version : 1.1

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.