[Feature] Support for multi-page OMRs

[Udayraj123]

Is your feature request related to a problem? Please describe. We have discovered use cases for OMRs having multiple pages. Need a discussion on how to standardise the process of creating multi-page OMRs so that OMRChecker can identify that the two images are part of the same OMR sheet.

Discord message link: https://discord.com/channels/590134763784896514/1122965202997096549/1122966382137573596

Describe the solution you'd like We will rely on file naming conventions and file structure: Consider adding a special new file template-multi.json with a rough draft schema:

{
    "templateNames":{
        "page1": {
            "relative_path": "template-page1.json",
            "imageNameSelector": {
                "type": "OMR_ID_PAGE_NUMBER",
                "imageNameSuffix": "_1"
            }
        },
        "page2": {
            "relative_path": "template-page2.json",
            "imageNameSelector": {
                "type": "OMR_ID_PAGE_NUMBER",
                "imageNameSuffix": "_2"
            }
        }
    },
    "outputColumns": ["q1", "q2", "q3", "q4"]
}

Describe alternatives you've considered

• A central file could also be used. Same as above config, but instead of template path we directly pass full template json
• One more way is to add a QR code to indicate omr identifier

Additional context Many directions will be affected with change in the central schema for template files. Some items to start thinking upon:

• Strong Validation will be needed to avoid bad UX
  • Restrictions over particular plugins in inner template configs
  • Column overlaps
  • Presence of files
  • Absence of extra files
• Supporting regex and file naming patterns
• Concatenation config requirements
• Output file location and structure changes
• Evaluation logic changes

[ghplvh]

把图片拼成一张，不就行了，虽然不是很提倡，但是也是个办法

[Udayraj123]

@ghplvh English please?

[ghplvh]

@ghplvh English please?

Merge the images？

[Udayraj123]

@ghplvh merging the images can be desirable too, but if the number of pages is too high, we would prefer creating a single folder per OMR id

[Sadhgun]

Here's a brief idea: Having two markers - one for identifying the participant details and one for describing the page number. The marker for page number could also be used to make sure the paper is correctly aligned. Pre processor to be created - Person identifier to rename the omr sheet. Page identifier to sort the files into two folders with each having their template.json Two csv created will be merged using person identifier as common marker

[Udayraj123]

@Sadhgun we will discuss on discord and put the minutes of the call here.