Dropping a session from annotations

[DerDiego13]

I used spans.correct and the closed then window without hitting save as I didn't want to save the annotations. However, they somehow got saved and are now added to the dataset. With "prodigy stats -ls" I can see all my sessions. Is there a way to drop the last one from the data base and then just to continue and redo them?

[ryanwesslen]

hi @DerDiego13,

Thanks for your question and welcome to the Prodigy community

You can run:

from prodigy.components.db import connect
db = connect()
db.drop_dataset("mydata-ryan")

Of course, it's good to be cautious, so you may want to run db-out before running to save a backup for your data.

Deleting certain annotation sessions

Yes, this definitely looks like your event_labels dataset has 12385 total examples. Maybe you accidentally added all of your annotations to the same dataset? To get to the bottom of this, you could run the db-out command, export it to a file and inspect it – maybe this will give you a better idea of what’s going on here? For example: prodigy db-out event_labels > event_labels.jsonl This will give you a JSONL file with all the records. You can edit it manually, or use a script to filter it. Wh…

Here's an example. Let's assume I'm saving my annotations into a dataset called mynerdata.

Let's assume these annotations have been saved from two sessions: diego (20 records annotated) and ryan (10 records annotated).

python -m prodigy ner.correct mynerdata en_core_web_sm data/news_headlines.jsonl --label ORG
Using 1 label(s): ORG

✨  Starting the web server at http://localhost:8080 ...
Open the app in your browser and start annotating!

✔ Saved 30 annotations to database SQLite
Dataset: mynerdata
Session ID: 2023-04-24_13-37-59

You can now run in Python using the database components:

>>> from prodigy.components.db import connect
>>> db = connect()
>>> len(db.get_dataset("mynerdata"))
30

Let's now view diego's and ryan's annotations:

>>> len(db.get_dataset_examples("mynerdata-diego"))
20
>>> len(db.get_dataset_examples("mynerdata-ryan"))
10

Now we'll use db.drop_dataset but for the session we want to drop (mynerdata-ryan):

>>> db.drop_dataset("mynerdata-ryan")
True
>>> len(db.get_dataset_examples("mynerdata"))
20
>>> len(db.get_dataset_examples("mynerdata-diego"))
20
>>> len(db.get_dataset_examples("mynerdata-ryan"))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: object of type 'NoneType' has no len()

So now mynerdata only has diego's annotations, but we can't find the session for ryan.

You can also use this same logic to clean up old sessions based on some criteria (e.g., remove any sessions without any annotations):

Dropping dataset from code doesn't properly delete examples

@Superscope Internally, sessions are also datasets and the individual examples you’re annotating are linked to both a regular dataset and a session dataset (and potentially other datasets). If I remember correctly, Prodigy isn’t currently purging empty sessions automatically – but you should be able to delete them just like regular datasets, by using the session name (timestamp). Alternatively, you can also write a custom Python script that retrieves and checks the sessions and deletes them (e.…

Hope this helps!

[DerDiego13]

Hi, thank you very much for the reply. I will have a look into it.
For now, I used "prodigy review" (as I needed to correct a few anyway) and will certainly come across the ones that were falsely added and then correct them along the way. Luckily, the dataset is not yet that long.