Backup and Data Retention

Backup and Data Retention

What a project contains

A project is a single folder. Everything about one examination lives inside it:

Demo Exam\
├── project.json          the project's identity and settings (format version 2)
├── database.sqlite       candidates, recognised results, decisions, marks
├── logs\                 what OMRFlow did on this project
├── answer_keys\
├── candidate_lists\      imported attendance workbooks
└── exports\              generated report workbooks

Treat the whole folder as confidential. database.sqlite holds candidate identifiers, recognised answers, attendance states and marks; the imported workbooks hold names and roll numbers. This is why nothing from a project belongs in a public issue — see SUPPORT.md.

Imported scans are referenced by content hash and provenance rather than copied wholesale, so the folder your scans live in matters too. If you move or delete the source images, OMRFlow can tell they are missing or have changed, but it cannot re-read them.

Where else OMRFlow keeps things

What	Where
Application settings (worker count, reviewer name, recent projects)	%LOCALAPPDATA%\OMRFlow\
Application log	%LOCALAPPDATA%\OMRFlow\logs\

Neither contains candidate data. Neither is removed by uninstalling.

Backing up

Through OMRFlow

Application menu → Tools → Project Health / Recovery… takes a snapshot of the project database using SQLite's own online backup API. This is safe against a database that is open, which a plain file copy is not: copying database.sqlite while OMRFlow has it open can produce a file that looks fine and is subtly broken.

A manifest is written only after the backup has finished and been hashed, so an interrupted backup can never be mistaken for a complete one.

The same dialog restores a backup.

By hand

With OMRFlow closed, copy the entire project folder. Closing matters for the same reason as above.

Keep the source scans too, if you may need to reprocess.

What to back up, and when

When	What
Before upgrading OMRFlow — always, on a prerelease	The whole project folder
After processing a batch, before resolving conflicts	The whole project folder
After scoring, before generating reports	At least a database backup
After generating final reports	The folder, plus the generated workbooks stored separately

The pattern worth keeping: back up at each point where the next step would be expensive to redo.

Retention

OMRFlow deletes nothing on its own.

• Corrections are append-only. Resolving a conflict does not overwrite what recognition read; both are kept, with who decided and why.
• Reprocessing archives the superseded reading before resetting a sheet, so a sheet processed twice keeps both readings on record.
• Reports are written to exports\ and are never overwritten silently.

How long to keep a project after an examination is your institution's decision. OMRFlow imposes no retention policy and will not remove anything for you.

What uninstalling removes

	Removed?
The application files	Yes
Your projects	No
Your settings and logs	No

There is deliberately no "also delete my data" option in the uninstaller: it is too easy to click, and examination data is too expensive to recreate. See Uninstalling.

Before reinstalling or moving machines

Copy:

1. Every project folder you need.
2. The source scan folders, if you might reprocess.
3. %LOCALAPPDATA%\OMRFlow\ if you want your settings — optional; it is only preferences.

Then see Moving Projects, which covers what OMRFlow does when a project or its scans appear at a different path.

Related

• Upgrading OMRFlow
• Moving Projects
• Recovery After Interrupted Processing
• Project Format