#129 data recipe to augment new features #131
Conversation
data/augment.py
Outdated
| @@ -0,0 +1,758 @@ | |||
| """ | |||
|
|
|||
| This data recipe lets the user to augment new features to the dataset using the Augment Cloud Service. | |||
There was a problem hiding this comment.
I would add more description about this recipe from the perspective of DAI. Example starting with requirements.
SnowFlake,
DataSet
DAI
description about augmentation, where the output of augment will be persisted, and how it is consumed by DAI.
data/augment.py
Outdated
| 6. The recipe polls the API for the completion of the table creation | ||
| 6. The recipe exports the dataset back to user's snowflake account | ||
| 7. The recipe downloads, saves the dataset from snowflake into driverlessai instance and returns the file path | ||
| 8. A new dataset is created in DAI with the augmented columns |
There was a problem hiding this comment.
customer facing recipe i would use full product name
| return "", str(e) | ||
|
|
||
|
|
||
| class AugmentDataset(CustomData): |
There was a problem hiding this comment.
where is the CustomData used? I guess this is needed since data recipe?
There was a problem hiding this comment.
@surenH2oai CustomData is the base class and we are overriding the method create_data in the subclass 'AugmentDataset. DAI will find out the subclass that derives from CustomDatawhich in this case isAugmentDatasetand it will invoke thecreate_data` method to get the updated dataframe with original columns + augmented columns
2 participants
This PR adds the data recipe that lets user augment new features to the dataset by using the augment service
https://github.com/h2oai/h2oai/issues/20586