POC: embedded JupyterLite notebook escape hatch (#650) - #771
Draft
dougc95 wants to merge 2 commits into
Draft
Conversation
Wiring for the #650 opt-in escape hatch: build.sh/build.ps1 produce a JupyterLite (Pyodide) site with a pre-seeded sample notebook that calls $sql-run from the browser and charts with pandas/matplotlib — no server-side Python. The built site (70 MB default / 463 MB vendored) is gitignored and reproduced by the build script. Verified: build.sh runs `jupyter lite build --contents content` and produces a servable dist/ with the sample notebook indexed (needed jupyter-server, now in the install line). README documents the no-CDN reality (default fetches Pyodide from jsDelivr and fails the guard; the --pyodide vendored build is 0-off-origin at 463 MB) and the CSP/IndexedDB caveats for embedding. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PTFfaGwHDggEbrFtfejJtE
The built site's Lab UI with the pre-seeded hfs-sql-on-fhir.ipynb open and the Python (Pyodide) kernel idle — proving the build bakes in the sample notebook.
Open
6 tasks
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
dougc95
added a commit
that referenced
this pull request
Aug 28, 2026
…vs proposal (#650) Adds a "Proof of concept" subsection to §5 mapping each option to its built demonstrator (PRs #768-#771) with what each verified, and relabels the now-shown feasibility as [demonstrated] while keeping the product decision, correct Stage 2, Stage 3, and the open questions as [proposal].
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Proof-of-concept for the opt-in escape hatch in the #650 evaluation: a full Python notebook compiled to WebAssembly, running entirely in the browser (no server-side interpreter). The runner-up, not the primary recommendation, because of its size and shape.
What it is
poc/jupyterlite/— build wiring (build.sh/build.ps1) that produces a JupyterLite (Pyodide) site with a pre-seeded sample notebook that calls$sql-runfrom the browser and charts with pandas/matplotlib. The built site (70 MB default / 463 MB vendored) is gitignored and reproduced by the script, so the branch stays small.Verified
build.shrunsjupyter lite build --contents contentand produces a servabledist/with the sample notebook indexed; the screenshot is that site's Lab UI withhfs-sql-on-fhir.ipynbopen and the Python (Pyodide) kernel idle. (Baking in custom content needsjupyter-server— now in the install line.)The no-CDN reality (measured, evaluation §10)
build.shas-is)cdn.jsdelivr.net--pyodide <tarball>)Only the vendored build passes the HFS no-CDN guard, at ~463 MB (≈3× the largest binary in the workspace) — which is why JupyterLite is the opt-in escape hatch, never in the default image. README covers the CSP / IndexedDB caveats for embedding.
Run
Relates to #650.