TODOs.txt

General

1 Testing
2 Make a windows installer
3 Manual!


******** QCtool ********

1 Improvements
    Dates
    - Detect of dates (library DateParser, detect if different format of dates exist in a column) - need improvement
    - Calculate statistics for date type variables
    Mixed types
    - Detect if a variable have different type of values..(ie numerical and has also string values)
    - Calculate statistics for cases with mixed type variables by choosing one type and ignoring the other type values - ie agegroup variable in ppmi dataset
    - Detect numerical variables with sufix (units, % etc)
    - Add new column with a list of row ids with outliers, (add similar column about rows with mixed type values)
    - Guessing the variable types if there is no metadata file - need improvement
    - Do more sophisticated the variable type guessing using a given percentage threshold  
    - 3 options of input files (csv metadata, csv only) - ok
    - CLI option for exporting only tex file and not pdf (for reducing the docker image size) - ok
    - CLI add error-warnings in cli mode for giving [csv\dicom] keyword
    - CLI make the -col_var and col_type optional, give default values
    - CLI make --input_csv and --root_folder positional argument
    - CLI use click library for parsing arguments   
    - integrate the goodtables.io table schema json and datapackage json (for automated quality check pipeline)
    Report
    - add histogram plottings in pdf for numerical variables
    - add values mean(+-)std, mean(+-)2*std, mean(+-)3*std
    - replace "nan" with "not avavailable" in pdf report
    - replace Latex2pdf with html2pdf for smaller docker container

    Various
    - list the packages in requirements.txt
    - CHANGE LICENCE from Apache 2 GNU v3

    GUI
    - GUI replace "no pdf" check box with "pdf"
    - GUI add select export file button - ok
    - add high resolution check - οκ
    - add minimum number of slices check - οκ
    - add type images check - οκ

    Extra Features
    - Connection with Data Catalogue for obtaining the hospital's variable version (using REST API?)

2 Testing
    - set up CIRCLECI
    - load various format of csvs (diff sep, quoted values)
    - additional test for fields of numerical variables
    - fields for categorical variables
    - dataset statistics
    - use twine and TestPyPI for uploading the packaget to PyPI (pip)
    - use Mock class