Add SOCRATES gap-map to clinician PDF

Wire SOCRATES gap-map into record display

Merge branch 'socrates-experiment'

Pivot framing to endometriosis: Endometriosis UK 2026 stats, adapted scoping line

Write full README: what, guardrails, architecture, run instructions, roadmap

Add SOCRATES annotation layer: tags dimensions her words address, gap-maps what's not yet described

Wire citizen-science layer into app: opt-in consent, feedback loop, aggregate view

Add seeded synthetic dataset and aggregate view: descriptive counts only, honesty flag on every entry

Add citizen-science contribution filter: anonymised schema, privacy by design

Track elevenlabs dependency in uv manifest

Add ElevenLabs read-back: opt-in accessibility feature, Emilia voice, same-language response

Add 2-minute demo script with corrected Pitt hook and Q&A back pocket

Add Maai theme: clay/stone palette, Fraunces headings, centred masthead, parchment ground

Add paste-ready cardiac demo descriptions (EN/ZH) with rationale

Update placeholder to atypical female cardiac presentation

Add evidence base: BHF Bias and Biology briefing, key figures with citation notes

Update Space card to gender-neutral identity; fix metadata block position

Commit dependency updates (gradio, fpdf2) and whitespace tidy

Add HF Spaces config: requirements.txt and README metadata

Bundle Noto Sans SC font and use repo-relative path: PDFs now render on any machine

Add Gradio app: paste description in any language, get advocacy record and PDF

Add make_pdf(): advocacy record PDF with Unicode/CJK support and visible no-diagnosis guardrail

Enforce completeness in normalise(): surface untranslatable expressions verbatim instead of dropping them

Add build_record() chain: wires intake -> normalise into one advocacy record

Remove default main.py from uv init; chain entry point comes next

Add language detection to normalise(): surfaces detected language, verbatim untouched

Add multilingual test: proves non-English intake preserves verbatim, maps to English clinical terms

Add intake() core: captures verbatim description with language and timestamp provenance

Add normalise() hero agent: maps lived language to clinical terms, keeping verbatim

Initial commit