Parvotec — Information System

How to track & share information

A reference for maintaining the parvotec knowledge base — adding new content, rebuilding views, and sharing access.

01 — Information Structure

Talks

transcripts/talks/
ASGCT 2026 presenter views. Generated from VTT + video frames. Rebuild: python3 build_talk_viewers.py

Paper Slides

transcripts/slides/
PDF → PNG + OCR per slide. Rebuild: python3 extract_pptx_ocr.py

Proposals & Reports

transcripts/*.html
Proposals (DE/EN), SAXS analysis, ASGCT recap. Edit YAML data → rebuild HTML.

02 — Adding New Content

New talk video (MP4)

1. Drop MP4 into ASGCT2026/
2. Run python3 transcribe_all.py → generates VTT
3. Run python3 build_talk_viewers.py → generates presenter HTML
4. Add entry to TALKS dict in build_talk_viewers.py

New paper (PDF)

1. Drop PDF into Machine learning for Rupert/
2. Run python3 extract_pptx_ocr.py → extracts slides to transcripts/slides/
3. Link from parvotec page on calyrai-web

Update key ideas

Run python3 extract_key_ideas.py — re-scores all VTT transcripts and injects top passages into talk pages.

Push everything live

git add -A && git commit -m "…" && git push origin main && git push origin main:gh-pages --force

03 — Sharing Access

Now (GitHub Pages)

URL: ruperttscheliessnig.github.io/parvotec-internal/

Gate: parvotec2026 (PBKDF2, client-side)
Share: send the URL + password directly

Future (Cloudflare Access)

URL: parvotec.calyrai.ai

Auth: Google / Microsoft login
Share: invite by e-mail in CF Zero Trust dashboard
Setup: add domain to CF → Pages → Access policy

04 — All Scripts

transcribe_all.py	Whisper: MP4 → VTT transcript files
extract_pptx_ocr.py	PDF/PPTX → slide PNGs + OCR text → slide viewer HTML
build_talk_viewers.py	VTT + video frames → click-through presenter HTML
extract_key_ideas.py	VTT → scored key passages → inject into talk pages
migrate_to_calyr_design.py	Migrate any HTML page to calyr.aí CSS (run once per page)
upload_to_youtube.py	OAuth2 upload 6 MP4s to YouTube (unlisted), injects IDs into HTML

← Transcripts Index · ← calyr.aí