CUOCO

Cuoco is a tool for automatic processing of data.

Example

Json example:

{
  "input_format": "csv",
  "output_format": "csv",
  "new_fileName": "new file",
  "new_file_route": "path/you/want/to/save/the/file",
  "index": "True",
  "header": "yes",
  "separator": ",",
  "num_nans": "mean",
  "str_nans": "yes",
  "caps": "lower",
  "normalize_method": "min_max",
  "normalize": [
    "Age"
  ],
  "balance_data": "yes",
  "balance_params": {
    "balance_method": "random",
    "y_col": "Age"
  }
}

Import the library

import cuoco
from cuoco import dataPipeline

Use the dataPipeline

dataPipeline.readJson('/content/biostats.csv', '/content/jsonTESTFILE.json')

Documentation

How it works: Cuoco uses a json created by the user to automatically apply data-processing functions to the desired dataset. The Json has the next values:

• input_format: format of the input dataset. Can be csv, parquet, orc and txt
• output_format: format of the resulted dataset. Can be csv, parquet, orc and txt
• new_fileName: name of the new dataset the DataChef will write
• new_file_route: route where to store the new data file
• index: if you want your final dataset to have a row index. Can be:
  • True
  • False
• header: if yor datasets has a header. Can be yes or none
• separator: the separator of your dataset. Only applies if its csv o txt format.
• num_nans: method you want to use against possible numerical nans (include empties). Can be:
  • drop: drop rows that contains nans
  • yes: dont do anything with rows that contains nans
  • mean: fill nans with the mean value of the column
  • median: fill nans with the median value of the column
  • mode: fill nans with the mode value of the column
• str_nans: method you want to use against possible string nans (include empties). Can be:
  • yes: keep nans columns
  • no: drop nans columns
• caps: method you want to use with strings that contains Upper and Lower case letters:
  • no: dont do anything
  • upper: put all strings of string columns to uppercase
  • lower: put all strings of string columns to lowercase
• normalize_method: method to use to normalize numerical columns. Can be:
  • no: dont normalize
  • max_abs: uses max absolute value to normalize
  • min_max: uses min - max value method to normalize
  • z_score: uses z-score value method to normalize
• normalize:
  • write the name of the columns you want to normalize
  • Note: if yor dataset does not have a header, you must write the columns's names you want to normalize in number format, if it has a header you must write the columns's names between ""
• balance_data: if you want to balance your data (recomended for AI datasets). Can be:
  • yes
  • no
• Inside balance_params there are two items:
  • balance_method: mehod you want for oversampling. Can be:
    • random: random oversampling
    • smote: perform SMOTE technique for oversampling.
  • y_col: column of the dataset you want to use as target for the balance