search

microsoft / TagAnomaly

Anomaly detection analysis and labeling tool, specifically for multiple time series (one time series per category)
MIT License
320 stars 61 forks source link

How do I annotate after loading the data #19

Closed EmmaGuo910 closed 3 years ago

EmmaGuo910 commented 4 years ago

There is no button "Add selected points" on the web page to annotate an anomaly. How should I annotate? image image

omri374 commented 4 years ago

Hi @EmmaGuo910, thanks for your issue. Are you using the Docker image on Dockerhub, or have cloned the repo and ran it locally?

EmmaGuo910 commented 4 years ago

Hi @EmmaGuo910, thanks for your issue. Are you using the Docker image on Dockerhub, or have cloned the repo and ran it locally?

I ran it locally in RStudio

omri374 commented 4 years ago

Could you please verify you have this line in ui.R? https://github.com/microsoft/TagAnomaly/blob/0022f1666ab9efb2f7aa9eb1f52bcc007f5cb2d8/ui.R#L65

EmmaGuo910 commented 4 years ago

Could you please verify you have this line in ui.R? https://github.com/microsoft/TagAnomaly/blob/0022f1666ab9efb2f7aa9eb1f52bcc007f5cb2d8/ui.R#L65

I verified there was that code image image

omri374 commented 4 years ago

It seems like you are loading an old version... Could it be? For example, the line following add and delete says "Currently marked points" but on the screenshot it says "Selected points" (which was the old version. See PR #16)

I would try to do a fresh clone of the repo and try again.

EmmaGuo910 commented 4 years ago

It seems like you are loading an old version... Could it be? For example, the line following add and delete says "Currently marked points" but on the screenshot it says "Selected points" (which was the old version. See PR #16)

I would try to do a fresh clone of the repo and try again.

Thanks. It worked after I switched branch to 1515e42f618d2c73bff116f6cabe63ef799e54c6.

However, I was confused by the issue followed. I downloaded the labels set after added points to candidate list. The csv file showed that the category remains normal. (I see from documentation that I can merge several csv files afterwards). If that so, "Add selected points" is meaningless, I suppose, because I could just selected points in the graph and download. It will generate same result after merging csv files. image image

omri374 commented 4 years ago

The new functionality allows you to select multiple locations on the graph and download them combined. If you have anomalies at the beginning of your time series, and at the end, you might need to perform multiple selections.

EmmaGuo910 commented 4 years ago

I understand. Thanks a lot.

EmmaGuo910 commented 4 years ago

I understand. Thanks a lot.

EmmaGuo910 commented 4 years ago

I understand. Thanks a lot.

dean-sh commented 3 years ago

Hey @omri374, I'm having the same issue, not being able to label the data after loading it. I tried running master branch, or 1515e42 as you suggested, and also run the image from DockerHub.

I don't have the option to add the candidates to the desired category.

How can I add the labels? image

omri374 commented 3 years ago

Hey @dean-sh, I'm not sure why this is happening. Could you too verify you have this line in ui.R? https://github.com/microsoft/TagAnomaly/blob/0022f1666ab9efb2f7aa9eb1f52bcc007f5cb2d8/ui.R#L65

dean-sh commented 3 years ago

Hey @dean-sh, I'm not sure why this is happening. Could you too verify you have this line in ui.R? https://github.com/microsoft/TagAnomaly/blob/0022f1666ab9efb2f7aa9eb1f52bcc007f5cb2d8/ui.R#L65

Indeed I do.

omri374 commented 3 years ago

I could not reproduce this on two different versions of R. If anybody has this issue, please provide reproduction steps (including OS, R version and shiny version)

omri374 commented 3 years ago

Found the root cause for this. There were two ui.R and server.R files on the repo. Fixed on #22