Repository navigation
Home
OMRFlow processes optical-mark-recognition answer sheets: you design a template, scan the sheets, let OMRFlow read them, review what it could not decide, reconcile attendance against your candidate list, score against an answer key, and generate result workbooks built on your own attendance workbook.
Everything happens on your machine. Nothing is uploaded anywhere.
This screenshot predates the shell redesign: the stages, the panel and the cards are current, the two-band chrome above them is not. See The application window in the README.
Version
0.1.0-alpha.2. Core workflows are implemented and covered by an automated test suite, but real examination-data qualification is still in progress — no real attendance workbook and no real scanned cohort has been processed end to end.Independently verify any generated result before using it operationally. See Known Limitations for exactly what has and has not been validated.
| If you want to… | Read |
|---|---|
| Install OMRFlow | Installation |
| See it work, end to end, in about half an hour | Quick Start |
| Know what is and is not trustworthy yet | Known Limitations |
| Understand a particular stage | User Guide |
| Install a newer Alpha without losing work | Upgrading OMRFlow |
| Fix something that has gone wrong | Troubleshooting |
| Report a problem | SUPPORT.md |
| Contribute code | Contribution Guide |
OMRFlow's window presents nine stages left to right, in the order you work through them. You can move between them freely; each stage says what it is waiting for.
| Stage | What you do | |
|---|---|---|
| 1 | Project | Create or open a project; name the examination and define its sets |
| 2 | Template | Mark up a blank answer sheet so OMRFlow knows where the bubbles are |
| 3 | Calibrate | Check the template against real scans and tune recognition before a batch |
| 4 | Scan | Import scanned sheets and run recognition |
| 5 | Resolve | Decide the sheets recognition was unsure about |
| 6 | Attendance | Import the candidate list and reconcile it against the scripts |
| 7 | Answer Key | Enter or scan the key for each question-paper set, and verify it |
| 8 | Results | Configure marking and calculate scores and ranks |
| 9 | Reports | Generate roll-wise and merit-wise workbooks |
Stages 6 and 7 do not depend on each other and can be done in either order, or while scanning is still running.
These pages are the user-facing documentation and are the canonical
source: the GitHub Wiki is published from docs/wiki/ in the repository, so
what you are reading is version-controlled alongside the code that it
describes.
Some pages are deliberately short and hand off to a longer reference
document kept elsewhere in the repository — docs/reporting.md,
docs/scoring.md and so on. Those documents are detailed, current and
verified; duplicating them here would create two versions of the truth and
one of them would rot. Where that happens the page says so and links
directly.
Developer documentation — architecture, the data model, the recognition
engine, the testing strategy — stays in docs/ and is linked from the
Developers section.
OMRFlow is pre-1.0 and developed in numbered phases. Phases 0–10 are implemented; the current work is Phase 11, which takes the application from "implemented and synthetically tested" to a qualified stable release:
- 11A — Alpha release infrastructure. This release.
- 11B — Real-data qualification and Beta. Validating against real examination material.
-
11C — Release candidate and stable. Feature freeze, packaged-build
qualification,
v1.0.0.
See the Development Roadmap for the detail, and Release History for what has shipped.
Every OMRFlow release is archived on Zenodo.
DOI: https://doi.org/10.5281/zenodo.22943575
If you use OMRFlow in research, teaching, academic work, software, or another project, please cite the project using its Zenodo DOI. See Citation for a ready-made plain-text and BibTeX entry.
MIT. See LICENSE.