Document pool

Your documents are stored in folders. Folders are paginated and you can move to the next or previous page using the navigation menu.

Folders

Use folders to organize your documents better. The root is the pool folder, which is created by default with each new project.

Create a new folder

Click on the folder you want to be the parent of the new folder. Click on the folder action Add new, write the name of the new folder and press the key .

Rename a folder

Click on the folder you want to rename. Click on the folder action Rename, write the new name of the new folder and press .

Remove a folder

Click on the folder you want to remove. Click on the folder action Remove. Please remember that all the documents stored in this folder will be also removed.

Upload text

In order to upload a new document, please select the folder you want to upload documents to and click on . Once clicked, a modal menu is displayed.

The different formats accepted are described here: Input formats

Upload files with predefined document labels

If you have document labels defined in your project, you can pre-annotate these labels for the document you want to upload. This is very handy if you have metadata (e.g. time stamp, type of document, industry, severity, etc.) available you want to have readily available for your annotators or your ML model.

For example, let's say your model use Webhooks to generate predictions once a document is uploaded. If the user has pre-annotated this document before, your model has valuable information to generate these predictions based on the pre-annotations. Language can significantly vary between departments, contexts, industries, time, etc., so you have an opportunity here to pick this info and generate better predictions accordingly.

This option in the user interface is only available when you upload files. Expand the Advanced menu and set the document labels.

Using the API you can automatically pre-annotate documents uploading together the document/text and the ann.json file with the annotations.

Upload pre-annotated documents

If you have pre-annotated documents, you can upload them directly to tagtog. You will need these two files:

The file wit the text content. The format of this file should be one of our supported input formats.

The file with the annotations. From the GUI, the only supported format for annotations is the ann.json.

Please remember to name both files the same, except for the extension. For example: mydoc.pdf and mydoc.ann.json. You can upload multiple pre-annotated documents at the same time. For example, 5 text files and 5 annotation files.

Please check the API for more options as replacing existing annotations.

Remove a document

You can remove a document on the web editor view or in the document list view by clicking on the remove button .

To remove documents in batch, you can use the search bar or the API for batch removal.

Manually confirmed documents

In the document list view, each document has a check mark, when green, it means the document is confirmed.

Manually confirmed documents are those with all annotations complete. Depending on the project, it can also mean that the annotations have been reviewed by a human, and they can be used as training data.

Confirm documents is helpful to keep the progress of the annotation tasks.