StudySprint v0.5.0

[0.5.0] — 2026-11-03

Fifth versioned state: MINOR feature batch (v0.5.0 batch, parts 1+2+3+4 of 4) — the calendar core
(build A, issue #4157) + maps and temple routes (build B,
issue #4158) + state recall and hardening (build C, issue
#4159) + ingestion on every surface (build D, issue #4242)
of RFP-D1-6.

Provider credit: command-dispatch (Swarm Optimizations &
Implementation Assurance seat), session
6628d41d-9519-4881-9063-3afe501df815.

Added

• Calendar core (study_calendar.py), RFP-D1-6 items 1-3:
  (1) calendar intake — class meetings {days, times,
  building/room} + assessments {quiz/midterm/final, date, time,
  location} normalized into records; (2) assessment-frequency
  engine — deterministic frequency ramps as an assessment
  approaches, with an explicit ramp + briefing for assessments at
  a SHIFTED time or place vs the class pattern; (3) quiet hours —
  notifications landing inside a scheduled class are suppressed
  and batched to the block's end. Claims ceiling (in receipts +
  FEATURES): logistics support — pacing, timing, no-surprises;
  no learning-gain claim.
• Tests: tests/test_study_calendar.py covers every public
  function in study_calendar.py.
• Geography (study_geo.py), RFP-D1-6 items 4-5, consuming the
  calendar core's records: (4) location records for classes +
  assessments (from the student's own calendar — open data first)
  • route/travel-time assistance to the assessment room (injected
    travel provider; offline coordinate estimate by default);
    (5) temple routes on real geography — the method of loci over
    the student's actual campus path, an audio-cue script for the
    walk, and voice-input customization (rename/move) as data.
    Ceilings: logistics support + the loci ceiling (ordered/concrete
    recall, not understanding).
• Tests: tests/test_study_geo.py covers every public function
  in study_geo.py.
• State recall + hardening (study_hardening.py), RFP-D1-6
  item 6 as its own mode: varied-context practice rotation +
  the Smith (1979) mental-reinstatement script (encoding
  specificity; context effect modest, meta d~0.28 — NOT
  single-room matching), and adaptive hardening as graduated
  stress-inoculation drills (SIT phases 1-3) — OPT-IN ONLY, off
  by default, the flag in every receipt; NO-GUILT by
  construction (no score/penalty fields exist; pressure never
  carries penalties). Ceiling: performance under
  stress/anxiety, not recall gains.
• Tests: tests/test_study_hardening.py covers every public
  function in study_hardening.py, incl. the opt-in refusal and
  the no-guilt field ban.
• Ingestion on every surface (build D, issue #4242), one core
  path (study_pipeline.ingest_any / run_syllabus_upload):
  syllabus text — typed or OCR'd from a photographed page
  (provenance source: ocr) — parses into study_calendar
  records {times · buildings · kinds} and places ONLY on the
  student's confirm (confirm-gated placement; no confirm ⇒ no
  placement); textbook uploads run the content flow (structure
  → term bank → SR cards). Surface legs: CLI ingest command
  on the demo/agent entry; a TUI Ingest screen (file-path entry
  • Confirm placement); the webApp Ingest view (file picker +
    camera capture — text parses in-browser under the same line
    grammar; photographs are recorded as awaiting core OCR, OCR
    does not run in the browser); the APK shell wires the system
    file chooser / camera to the bundle's inputs
    (WebChromeClient.onShowFileChooser).
• Tests: syllabus parse + confirm gate + OCR provenance +
  per-surface legs in tests/test_study_pipeline.py,
  tests/test_tui.py, tests/test_web_surface.py,
  tests/test_android_surface.py.