CSV export format description

Important foreword

To understand how CSV export is crafted, we strongly suggest you read Structure of a result. This describes how Asqatasun results are organised. Once you get this, export formats will be quite straightforward to understand.

Structure of CSV export

An audit exported in CSV format is composed of one line per test. Each line contains the following fields:

Theme, Criterion, Test, Test Label, Level, Result, Details

First two lines are used to describe URL evaluated, level and date.

Note: Some fields (for instance Details) carry data that are themselves in a CSV-like format nested in the overall CSV export. Data may be nested up to 4 levels of CSV-like. (This will be extensively describded below.)

Theme

Theme is the name of the topic.

Example for RGAA 3 referential:

  • Images
  • Frames
  • Colours

Criterion

Criterion is the reference of the criterion

Example for RGAA 3 referential:

  • 1.1
  • 1.2
  • 8.9
  • 13.7

Test

Test is the reference of the test.

Example for RGAA 3 referential:

  • 1.1.1
  • 1.2.1
  • 8.9.1
  • 13.7.1

Test Label

Test Label is the complete description fo the test (from the official referential).

Example for test 1.1.1 from RGAA 3 referential:

Does each decorative image without legend (tag img) and with an alt attribute meet all these requirements?

  • the content of the alt attribute is empty (alt="")
  • the decorative image has no title attribute

Level

Level is the priority of the test.

Examples for RGAA 3 referential:

  • LEVEL_1 (equivalent of priority “A”)
  • LEVEL_2 (equivalent of priority “AA”)
  • LEVEL_3 (equivalent of priority “AAA”)

Result

Result is the actual result of the test. It can be one of:

  • Passed
  • Failed
  • Not Applicable
  • Pre-Qualified
  • Not Tested

Details

Details is organised in a CSV-like format.

Separator used is character ø:

Process remarks count ø Process remarks header ø Process remarks content

Let’s get in the internals of Details.

Process remarks counts

Process remarks count gives the number of instances of Process remarks content. Its value can be from 0 to N (N being an integer).

Process remarks headers

Process remarks headers is a string defining the structure of Process remarks content (that will be detailed below).

Its value is:

{key; result; target; evidenceElementCounter; evidenceElementStructure; evidenceElementList}

Process remarks content

Process remarks content is organised in a CSV-like format this way:

  • Separator is character ;
  • Data are enclosed into curly brackets { and }

Data organisation for Process remarks content was in fact describded with the headers (just above). We place it again here:

{key; result; target; evidenceElementCounter; evidenceElementStructure; evidenceElementList}

Let’s get in the internals of Process remarks content

Key

Key is the “message key” describing the Process Remark.

Here are a few examples of Key values:

  • AltMissing
  • CheckNatureOfElementWithNotEmptyAltAttribute
  • TitleAttributeMissing
  • CheckTheContrastOfImagesWithText

You may browse whole list of Process Remarks keys

Result of Process Remark

Result corresponds to the result of the Process Remark. Its value can be failed or nmi (nmi, “need more information”, was the old name for Pre-Qualified).

Target

Target is the element targeted by the Process Remark. This is usually an HTML element.

EvidenceElementCounter

EvidenceElementCounter is the number of Evidence Elements present is EvidenceElementList

EvidenceElementStructure

EvidenceElementStructure is the description (headers) of the CSV-like structure used for the Evidence Elements.

Data are enclosed with brackets ([ and ]), and fields are separated by pipes |.

Note: it is important to understand this structure varies and depends on each test.

EvidenceElementList

EvidenceElementList is a CSV-like structure used to store all the Evidence Elements from a given Process Remark.

Data are enclosed with brackets ([ and ]), and fields are separated by pipes |.

As already said, the actual values inside EvidenceElementList depend on each test.