Experiment Assignment App

PHP/MySQL web app for assigning experiment access information to ZHAW students and tracking whether participation is counted for grading.

Structure

• index.html: student UI
• assets/: student CSS and JavaScript
• manage/: staff UI
• api/: student and staff JSON endpoints
• config/config.php: deployment database configuration
• database/schema.sql: V2 schema
• database/seed.sql: real course allowlist only
• database/seed_examples.sql: optional representative demo experiments and access data
• database/reset.sql: remove experiment setup and runtime data while preserving the global allowlist
• database/drop_tables.sql: drop all application tables in dependency order
• .agents/CONTEXT.md: current domain decisions
• .agents/PROJECT.md: milestone audit trail
• tests/: PHP checks

Database

Import the schema into an empty MySQL database:

mysql -u USER -p DATABASE < database/schema.sql
mysql -u USER -p DATABASE < database/seed.sql

seed.sql contains only the current real allowed_students list. Use database/seed_examples.sql only for a throwaway/demo database because it creates representative example experiments, conditions, selected participants, staff-prepared values, access pools, and slots.

Use database/reset.sql when you want to remove all experiment configuration and runtime data from a deployment while keeping the global allowed_students list.

Use database/drop_tables.sql only when you want to remove the full application schema before rebuilding it from schema.sql.

Configuration

Database deployment settings are stored in config/config.php. EXPERIMENT_DB_DSN remains available as an optional override for tests, especially the SQLite smoke test.

Student Flow

1. Student enters a @students.zhaw.ch email address.
2. The app loads all experiments visible to that student.
3. Closed experiments stay visible but cannot be opened.
4. Open experiments can be claimed once.
5. Existing claims are retrieved by email and show the same access information again.
6. Slot-based experiments require one slot choice with capacity checks.
7. Staff-entered appointment text appears in the access information when available.
8. Angerechnet is shown after staff confirms the participation.
9. The current email session is shown in the top navbar and can be ended with Beenden.

Staff Flow

The staff UI is at manage/index.html.

It supports:

• adding allowed students
• bulk-importing allowed students
• opening the dedicated global allowlist view from the editable student-count badge
• viewing and removing allowed students without participations
• creating and renaming experiments
• deleting setup experiments
• opening and closing experiments
• adding conditions
• deleting unused conditions
• selecting experiment participants from the global allowlist by all, seeded random subset, or manual email search
• clearing an experiment-specific participant selection before participations exist
• assigning selected participants to conditions manually or by seeded percentage randomization
• clearing condition assignments before participations exist
• defining access fields
• deleting access fields
• preparing bundled access pool rows through a guided CSV modal
• clearing unassigned access pool data
• preparing student-specific values for access fields sourced from staff entry before access is revealed
• viewing and deleting unassigned access pool rows
• manually assigning students
• viewing and removing experiment-specific eligibilities
• deterministic randomization across conditions
• creating time slots with capacity
• showing time-slot setup only for experiments marked as requiring time slots
• viewing slot choices by slot
• deleting unused time slots
• opening experiment-specific editing by clicking an experiment in the overview
• opening experiment-specific grading from the overview
• opening the Reports view from the navbar
• viewing one report row per globally allowed student with Kürzel plus one 0/1 approval column per experiment
• sorting report columns, filtering by Kürzel, and downloading the displayed report as CSV
• showing the access reveal time and compact access values in grading, with link fields rendered as labeled buttons
• filtering and sorting the grading table by each data column
• building a checked participation selection in the grading modal and applying bulk grading actions
• showing a navbar status indicator while backend requests are running
• listing registered students who have not opened access yet in the grading view
• setting appointment text
• resetting participations
• toggling Angerechnet
• navigating overview, experiment setup, and grading through the top workflow strip
• returning to the experiment overview through the navbar brand

Access fields that already back assigned runtime values cannot be deleted or structurally changed. Existing slot capacity cannot be reduced below the number of submitted slot choices.

For experiments with configured condition rows, access-pool imports are condition-scoped. Choose the target condition in the pool modal; the CSV for that condition includes both experiment-wide pool fields and fields specific to that condition. The experiment-wide pool option is used while no conditions exist.

Staff authentication is intentionally not implemented yet; the page relies on a hidden URL.

Tests

php tests/validation_test.php
php tests/text_quality_test.php
php tests/js_regression_test.php
php tests/api_smoke_test.php

js_regression_test.php catches focused management-client regressions that are not covered by JavaScript syntax checking alone. api_smoke_test.php uses a temporary SQLite database and skips when pdo_sqlite is unavailable. When SQLite support is available, it covers the student claim/retrieval flow, slot capacity enforcement, management setup actions, allowlist removal guards, participant selection and clearing, condition assignment and clearing, access-pool import, staff-entered access values, confirmation, bulk grading operations, appointment retrieval, reset, randomization, and the management approval report endpoint. The report coverage includes the distinction between opened access and confirmed Angerechnet approval.