Deprecating NER Annotator in favour of Label Studio

[tecoholic]

I recently came across Label Studio while looking for a image labeling tool. It seems to cover every aspect of ML labeling from text to video.

I gave it a try. I believe the NER support of Label Studio is great replacement for NER Annotator. It provides a lot of export options compared to the SpaCy focused JSON output of NER Annotator.

The only drawback seems to be that it is a huge web application instead of a simple web+desktop application. However, the installation is a simple pip install label-studio away. So, I am NOT really concerned about NER Annotator being deprecated in its favour.

Key points

• NER Annotator isn't actively developed. Only sparingly maintained.
• Users can still download the existing versions from Github.
• Deprecation will archive this project on Github and make it read-only. If you want to for fork and develop your own features, you can still do it.

[tecoholic]

@alvi-khan If you feel like you want to keep this project open, kindly let me know. Otherwise, I will mark this repo as archived next week.

[alvi-khan]

@tecoholic From what I understand, your goal with building NER Annotator was to create a simple and user-friendly tool for a specific use case. It does that job very well, but I can also see how expanding it to cover a broader range of use cases would present challenges. Since I’m not currently working on ML projects, my own contributions to new features have naturally slowed down as well. Given that, I completely understand your decision to archive the project.

Label Studio does seem like a more comprehensive tool, but I believe the simple desktop application/hosted website format of NER Annotator is easier to use, especially for non-technical users. When preparing datasets in the past, I found that on-boarding people to the tool was incredibly smooth. Even if the project is archived, the existing versions remain functional, and there are no major issues with them. So, I don’t think archiving it will cause any significant problems for those who still want to use it.

[tecoholic]

@alvi-khan thank you I understand your point about being a simple tool that does one thing in a straightforward manner. I am not going to "Archive" it yet, but add a not on the top of README mentioning it is no longer actively developed.