Add detect_from_csvs and detect_from_dataframes methods to MultiTableMetadata

[amontanez24]

Problem Description

As a user, it would be convenient to be able to detect the metadata for a whole dataset at once, instead of one table at a time. This would speed up the process and also enable detection for relationships to be done.

Currently we have the following detect methods in MultiTableMetadata:

• detect_table_from_dataframe
• detect_table_from_csv

These methods only detect one table at a time which is inconvenient.. It also makes it impossible to detect information between tables like relationships. For this reason we want to add two new methods that allow us to detect an entire folder or dictionary of dataframes at once.

Expected behavior

• Add detect_from_csvs

  • parameters:
    • folder_name: full name of folder where the csvs are stored.
  • behavior: This method should loop through all the csvs in the folder and load them the same way detect_table_from_csv does. Use the name of the csv as the table name (ie. if the file is adults.csv then the table name is adults).

• Add detect_from_dataframes

  • parameters:
    • data: dictionary mapping table names to dataframes.
  • behavior: This method should loop through the dataframes in the dictionary and detect them one at a time similarly to detect_table_from_dataframe.

metadata.detect_from_csv(
   folder_name='data/'
)

metadata.detect_from_dataframes(
   data={
       'guests': guests_table,
       'hotels': hotels_table
   }
)