mldev-reporting 0.1.1

An extantion library for creatind reports for MLDev

MLDev Reporting

Welcome to MLDev reporting - library for creating html reports for your experiment. The library provides functionaity for creating reports using Markdown templates and data from your experiment.

Start creating insightful reports today and unlock a clearer understanding of your ML experiments with MLDev reporting.

User Guide

First Step

Install MLDev Reporting using pip with comand

pip install mldev_reporting

Don't forget to add MLDev Reorting to your .mldev/stages/__init__.py file

import mldev_reporting

Second Step

Create an experement using experiment.yml

pipeline: !GenericPipeline
  runs:
    - !BasicStage
      name: report

      env:
        PYTHONPATH: '${env.PYTHONPATH}:./.mldev'

      script:
        - mldev run -f ./report.yml

Third Step

Create a report.yml to configire your report

report: &report !Report
  name: "index"
  output_dir: "light_report"
  template: "./template.md"
  theme: "light"
  lang: "ru"
  report_model:


pipeline: !GenericPipeline
  runs:
    - *report

name is the name of your html file. output_dir is a folder where your report will be located. template is a link to your tamplate file. theme is a theme of your report. It can be light | dark | cards. lang is the report language. It can be ru or en. report_model is where your report data will go.

Fourth Step

Create a template for the report. repo is an object wrom with you can access your report data

# Me cool report

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla sem elit, tincidunt {{repo.data.result}} consectetur aliquam eget, ornare eget nunc. Aliquam erat purus, rutrum in erat nec, sodales commodo augue. Nullam ante nibh, accumsan sit amet eleifend nec, porttitor vel tellus. Sed dapibus at sapien vel mattis. Ut commodo quam tortor, quis porta erat convallis quis.

## Subtitle 1
{{repo.table}}

## Subtitle 2
{{repo.graphic}}

Fifth Step

Specify your report data. You can do it ather by hand or by loading data from file using !LoadFile

loadfile: &load_json_file !LoadFile
  url: "index.json"

loadfile: &load_cvs_file !LoadFile
  url: "index.csv"

Now you can add your data to report_model

report: &report !Report
  name: "index"
  output_dir: "light_report"
  template: "./template.md"
  theme: "light"
  lang: "ru"
  report_model:
    data: *load_json_file

To create table use this comands

report: &report !Report
  name: "index"
  output_dir: "light_report"
  template: "./template.md"
  theme: "light"
  lang: "ru"
  report_model:
    data: *load_json_file
    table: !Table
      title: "Example Table"
      data: *load_cvs_file

To create graphic use this comands

report: &report !Report
  name: "index"
  output_dir: "light_report"
  template: "./template.md"
  theme: "light"
  lang: "ru"
  report_model:
    data: *load_json_file
    graphic: !Graphic
      datasets:
        - dataset: *load_cvs_file
          x: "x_axis"
          y: "y_axis"
          name: "Example Graphic"
          graphic_type: "line"

Sixth Step

Run the following comand to generate your report

mldev run -f experiment.yml

Documentation

Report Class

Field name	Type	Values	Default Value
name	string	Report file name	Required
output_dir	string	Path to the folder where the report will be stored	Required
template	string	Layout file path	Required
report_model	object	Report data	Required
theme	string	Topic	light
lang	string	Language	en

Table class

Field name	Type	Values	Default Value
title	string	Table name	Required
headers	string / array of strings	Object key if json file is used / Table column headers	Optional
rows	string / array of arrays of strings	Object key if json file is used / Table rows	Optional
data	link to data from file	Data from a file in json or csv format	Optional
widths	string	Column width as a percentage	Optional
header_rows	Integer	Line number to be used as header	Optional

Class Graphic

Field name	Type	Values	Default Value
datasets	array	Graph Datasets	Required
config	object	Graph Configuration	Empty object

More detailed information about the datasets can be found on the library page.

LoadFile class

Field name	Type	Values	Default Value
url	string	File path	Required
filter_fields	string array	A set of keys whose values ​​will be filtered (for json)	Optional
filter_columns	string array	A set of columns whose values ​​will be filtered (for csv)	Optional

Contributing

Please check the CONTRIBUTING.md guide if you'd like to participate in the project, ask a question or give a suggestion.

License

The software is licensed under Apache 2.0 license.