PDF/MVP Coverage Matrix

This matrix maps every listed PDF/MVP item from the CS14 acceptance checklist to current documentation coverage, expected product behavior, and final acceptance evidence.

#	PDF/MVP item	Documentation coverage	Expected product behavior	Acceptance evidence
1	Working deployed MVP	Local Setup & Live Deployment covers local smoke testing and the live Japan VM deployment.	Docker Compose runs PostgreSQL, FastAPI, Next.js, Caddy, and Cloudflare Tunnel; the public app is available at https://cs14.kazelis.top/.	Public /health, Swagger, frontend, seeded participant link, and docs site smoke checks pass.
2	Admin survey builder with create/edit surveys, blocks, and question types	Researcher Workflow documents create/edit, blocks, and question types.	Researcher can create surveys, add/edit/reorder content blocks, and configure free text, single choice, multiple choice, Likert, and rating questions.	Builder UI smoke test, persisted block changes, backend route tests, frontend build.
3	Social-media post template editor with feed posts, cards, counts, comments, source labels, and More Information buttons	Researcher Workflow and Social Post Fetching document the card editor and fetch/manual fallback rules.	Researcher can fetch metadata when possible, override all display fields, set engagement counts, write comment previews, and track More Information clicks.	Post card preview per condition/language, participant feed rendering, interaction tracking in export.
4	Preview/testing workflow	Researcher Workflow defines preview before publish and test-session discipline.	Researcher can preview selected condition/language without polluting final analytics, then perform a participant smoke run.	Preview session excluded or marked, condition/language rendering verified, share link copied only after review.
5	At least 2 languages end-to-end, translation CSV/JSON import/export, and participant language selection	Data Export & Translations gives explicit import/export and participant language-selection steps.	Admin exports translation templates, imports completed translations, and participant language choice drives rendered survey content.	CSV/JSON roundtrip, selected language stored on response, English/Chinese (plus Japanese, Korean, Spanish) checked where available.
6	Participant runner with clean UI, condition assignment, and randomisation	Researcher Workflow documents publish/share, participant run, and condition preview.	Participant opens one share link, chooses language, receives assigned condition, completes required blocks, and submits once.	Participant run smoke test, persisted response, condition assignment visible in analytics/export.
7	Camera calibration module with permission flow, guidance, face detection, dot sequence, score, and stored outcomes	Calibration & Privacy covers flow, troubleshooting, pass/fail interpretation, and privacy limits.	Participant grants camera permission, completes dot sequence, receives pass/acceptable/poor outcome, and calibration summary is stored without raw video.	Calibration session and points stored, numeric score/outcome exported, retry behavior checked.
8	Database schema for surveys, translations, anonymous participants, responses, calibration logs	Data Export & Translations and Calibration & Privacy document the researcher-visible data contract.	Data model supports surveys, posts, questions, translations, anonymous responses, interactions, calibration sessions, gaze/click summaries.	Alembic migrations, schema tests, privacy check that raw participant tokens/video are not exported.
9	CSV/JSON export with filters by survey, condition, and language	Data Export & Translations gives explicit CSV/JSON export and filter instructions.	Researcher exports owned survey data as CSV or JSON, filtered by survey, condition, language, completion status, calibration outcome, and preview inclusion.	Export files open cleanly, filters change row set, stable column/JSON schema, privacy exclusions verified.
10	Local and deployment setup docs plus researcher user guide	This VitePress site provides separate setup, deployment, researcher, export, fetching, calibration, and privacy pages.	A new team member can run locally, follow the researcher flow, understand live demo operations, and avoid unsupported fetch/export assumptions.	npm run build for docs, public docs at https://cs14-docs.kazelis.top/, and app smoke checklist before client demonstration.

Current Readiness Notes

• The docs site is intentionally separate from backend/frontend runtime code.
• Deployment instructions now describe the live Japan VM plus Cloudflare Tunnel setup.
• Product gaps should continue to be tracked in the main acceptance checklist and implementation issues.

Live URLs

• Participant/researcher app: https://cs14.kazelis.top/
• Public docs site: https://cs14-docs.kazelis.top/
• Published participant smoke link: https://cs14.kazelis.top/survey/CS14DEMO2026?lang=en

Client Feedback Status

Feedback item	Current status	Handoff note
Platform UI dropdown for X, Facebook, Truth Social, Instagram, Bluesky	Partial	Current code carries platform UI style through survey create/edit and the participant renderer has distinct feed skins. Final review should verify one published survey per style because older branches collapsed Truth Social and Bluesky back to X-style.
External post/headline/image/More Information opens in a new tab	Done	Participant click handling opens _blank synchronously before interaction logging and keeps the survey tab on the same route. Browser QA should verify headline, image, and More Information each create one intentional click interaction.
Survey-time gaze/attention data	Partial	Calibration sessions and post-calibration gaze/click capture are implemented, including participant-token hardening and attention summary plumbing. Public browser evidence still needs a real camera or controlled fake-camera pass to confirm the live survey-time stream end to end.
Camera calibration flow	Done with device-risk	Webcam permission, face stability, dot sequence, score, and stored outcomes are implemented. Demo readiness still depends on HTTPS camera permission and room lighting on the presenter device.
CSV/JSON export with useful filters	Partial	Backend/export paths expose CSV and JSON filters for group/language/status/calibration/preview. Final acceptance should verify that the visible analytics/export UI changes row sets and labels export-only filters honestly.
Live deployment and docs links	Done	App and docs are live at the URLs above; do not revert wording to "deployment postponed".

Transcript Handoff Notes

• 8:17-13:52: client asked which platform UI is shown and requested a researcher dropdown for Facebook/X/Truth Social/Instagram/Bluesky-style presentation.
• 14:28-15:26: client said post clicks, including More Information, must open a separate tab so the participant does not lose the survey page.
• 16:04-17:33: client asked to see CSV/export data.
• 17:33-18:24: client asked where survey-time eye-tracking data is, not only calibration output.
• 18:24 onward: client asked to put these comments into the status check form and send the platform/Cloudflare links.

External Click QA Checklist

• Start from the published participant smoke link above. If calibration blocks automation, complete calibration manually or use a local app run with a post that includes a hero image.
• Click the post headline, hero image, and More Information button separately.
• Expected result for each click: a separate tab opens to the original URL, the survey tab remains on /survey/..., and exactly one semantic click interaction is sent with the current participant_token.
• Expected raw click tracking: headline clicks are labelled headline, image clicks are labelled image, More Information clicks are labelled more_info, and action buttons keep their own labels.