Working with your environment#

Sometimes, you may need to have more information about your environment data in JAI. For this, JAI owns some managing methods that are shown in this section.


If you need to know which databases are in your environment, use j.names. It’ll show you the names of all your databases in JAI.

>>> # Example response supposing these collections are in the environment:
>>> j.names

['jai_database', 'jai_selfsupervised', 'jai_supervised']

To get more information about your collections, use This method will present all your collection names, types, last modifications, dependencies and sizes.


              name|           type|    last modified|          dependencies|   size|
      jai_database|           Text| 2021-10-26-20h22|                    []|  20640|
jai_selfsupervised| SelfSupervised| 2021-11-03-17h50|                    []|   8500|
    jai_supervised|     Supervised| 2021-11-03-17h50|  [jai_selfsupervised]|   8500|

The drawback of is that it can run slowly depending on the number of collections in your environment.


If you forgot what columns your database has, this information could be accessed by j.fields method.

>>> # Use your collection name
>>> j.fields(name='jai_database')

{'id': 'int64',
 'Column1': 'float64',
 'Column2': 'float64',
 'Column3': 'float64',
 'Column4': 'float64'}


To get what collection type your collection is, use j.get_dtype.

>>> # Use your collection name
>>> j.get_dtype(name='jai_selfsupervised')



However, if you need details of what parameters you choose to fit your collection, j.describe can bring it for you.

>>> # Use your collection name
>>> j.describe(name='jai_database')

{'name': 'california',
 'dtype': 'SelfSupervised',
 'state': 'active',
 'version': '2021-10-26-20h22',
 'has_filter': False,
 'model_hyperparams': {'batch_size': 512,
 'learning_rate': 0.01,
 'encoder_layer': '2LM',
 'decoder_layer': '2LM',
 'hidden_latent_dim': 64,
 'dropout_rate': 0.1,
 'momentum': 0.1,
 'pretraining_ratio': 0.1,
 'noise_level': 0.0,
 'training_type': 'contrastive'}

To recover the fit report for your collection, use

>>> # Use your collection name


If you need to remember how many ids your collection have, use j.ids.

>>> # Use your collection name
>>> j.ids(name='jai_database', mode='summarized') # default

['20640 items from 0 to 20639']

For more information about how to work with your environment, check JAI Python Class