Convert assessment question banks into LMS packages, practice quizzes, and readable review copies with one Rust engine for the native CLI and browser WebAssembly. Instructors keep one editable source while reusing assessments across learning systems.
The native converter builds from this source checkout. The shared native/Wasm refactor's current acceptance state is recorded in docs/active_plans/reports/shared_engine_delivery.md. No release installer is published yet. It preserves useful Python workflows while allowing harmless presentation differences. Table rendering always runs headlessly and automatically prefers Chromium headless shell; see docs/INSTALL.md for setup and known limitations. Review an imported package in its target LMS before relying on it for a course or exam.
Start with one tab-delimited Blackboard Question Upload (BBQ) text file. Select the output that serves the next teaching task.
| Teaching task | Native output | What it gives you |
|---|---|---|
| Import into Canvas | Canvas QTI 1.2 ZIP | A package ready for Canvas import review |
| Import into Blackboard | QTI 2.1 or Original pool ZIP | A Blackboard package or pool export |
| Let students practice | Self-grading HTML | A standalone browser quiz with immediate feedback |
| Review before release | Human-readable HTML | A readable copy of the questions and answers |
| Reuse elsewhere | Aiken, YAML, text2qti, BBQ, or BQGen | Text formats for Moodle, exams, and other tools |
| Hand questions to PLE | PLE Native JSON directory | One JSON document per question with associated media |
The same bank can produce several of these outputs in one command, so edits stay in the source questions instead of becoming separate copies in each learning system.
Install a Rust toolchain meeting rust-version in Cargo.toml.
From this repository, build the native commands:
cargo build --locked --release -p qpm-cli --binsCreate bbq-demo-questions.txt with this tab-delimited row:
MC Which base pairs with A? T correct C incorrect
Create a Canvas package, a Blackboard package, an instructor review copy, and a student practice quiz:
./target/release/bbq-converter \
--input bbq-demo-questions.txt \
--qti12 --qti21 --human --selftestThe command writes qti12-demo.zip, qti21-demo.zip, human-demo.html, and
selftest-demo.html. Open the self-test in a browser to try the learner-facing question, read the
human copy before publishing, then import the ZIP appropriate to the LMS.
Use the short flags for common destinations. --all selects all eleven registered writers; use
--format <engine-name> when the output does not have a shortcut.
# Canvas QTI 1.2
./target/release/bbq-converter --input bbq-genetics-questions.txt --qti12
# Blackboard QTI 2.1, with supported tables and static RDKit canvases rendered as PNGs
./target/release/bbq-converter --input bbq-genetics-questions.txt --qti21 --html-to-image
# Moodle Aiken text
./target/release/bbq-converter --input bbq-genetics-questions.txt --aiken
# PLE Native JSON and associated files
./target/release/bbq-converter --input bbq-genetics-questions.txt --format ple_native_json--html-to-image is available for Canvas QTI 1.2, Blackboard QTI 2.1, Blackboard Original
pool exports, and PLE Native JSON. Table rendering requires Chromium or Chrome. It needs the
optional RDKit shim
only when the bank contains an RDKit canvas. See
docs/HTML_TO_IMAGE.md for the supported content and setup.
The local @vosslab/qpm-wasm package exposes explicit initialization, format inventory,
conversion, and package checking. Its generated TypeScript API uses owned Uint8Array values
for input documents, companion images, and outputs. All four readers and eleven writers share
the native Rust implementation; native screenshot and molecule rendering stay in the CLI.
See docs/WASM_PACKAGE.md for building the package, typed calls, workers, and downloads. The package is private and has no publication step. The native CLI needs no Node or Wasm setup.
Use the companion command to see the formats compiled into your binary or to inspect a finished ZIP for package-integrity problems:
./target/release/qpm engines
./target/release/qpm check qti12-demo.zipcheck prints OK when it finds no integrity violations. It checks package structure and answer
bindings; it does not replace review of the imported questions in the target LMS.
- docs/INSTALL.md - source build, platform requirements, and the optional RDKit shim.
- docs/USAGE.md - CLI flags, output names, and practical conversion commands.
- docs/FORMATS.md - all eleven outputs, media behavior, and format limits.
- docs/QUESTION_TYPES.md - fields for the seven supported assessment types.
- docs/ENGINES.md - the current reader, writer, and media-capability inventory.
- docs/WASM_PACKAGE.md - local browser/Node package and byte API.
The maintained Python QTI Package Maker remains the reference implementation. Historical native verification is documented separately from shared engine acceptance; real LMS imports and release publication remain separate steps. Developers can read the completed port plan and refactor_progress.md.
This project is licensed under the GNU Lesser General Public License v3.0 or later. See LICENSE.LGPL-3.0.