interested in helping in testing

[justinhjy1004]

hi, I saw your post on reddit on needing help testing things out.
this project seems to be quite cool, but I don't have an android device, though i have tried it out on my laptop.
my background is python, with interest in scientific computing, specifically with familiarity in np (from machine learning and deep learning) and sympy (used in algebraic geometry)
let me know how i can help, i am pretty new to this

[s243a]

Hello, thank you for your interest!

Getting started: The app runs in any modern browser — no Android needed. You can try it at [GitHub Pages link] or locally with npm run serve. NumPy and SymPy are both preloaded and ready to use.

How you can help: Example notebooks for Python would be very valuable — things like:

• A SymPy algebraic geometry walkthrough
• A NumPy/linear algebra tutorial
• A machine learning demo (e.g. gradient descent from scratch)

To create one, just write your cells in the app, then export via Menu > Export .ipynb. You can submit it as a PR or attach it to an issue. Test the export functionality before you get too far along though.

Regarding packages: Pyodide (our Python runtime) comes with many scientific packages pre-built (NumPy, SciPy, SymPy, scikit-learn, etc.). Pure-Python packages can also be installed at runtime via micropip.
Packages with C extensions need to be compiled to WebAssembly — if you need one that isn't available, let me know and I can look into building it.

Loading packages: Notebooks can be bundled as installable packages (Menu > Browse Packages). On the web version, packages must be hosted on the same domain due to browser CORS restrictions, but this isn't a limitation on Android or when running locally (the dev server proxies downloads through localhost to bypass CORS).

To run locally run the following command from the project root:

npm run serve

[justinhjy1004]

Sounds good and thank you! Just another quick question, do you have the feature of uploading csv files etc. at the moment?

[s243a]

Sounds good and thank you! Just another quick question, do you have the feature of uploading csv files etc. at the moment?

CSV upload works, but currently requires packaging it in a .zip file first. See: packages.md

Workaround for now:

Option 1: Create CSV programmatically

import sharedfs
sharedfs.write('/shared/data/mydata.csv', '''name,score
Alice,95
Bob,87''')

import pandas as pd
df = pd.read_csv('/shared/data/mydata.csv')
df.head()

Option 2: Package it with your notebook
Create a .zip with a scirepl.json manifest and your CSV files — see the package docs linked above.

Coming soon
I'll add direct CSV/data file upload (without needing to create a package) soon, since this is a common workflow. I'll update this issue when it's ready.

Another way you could help:

if you create tutorial notebooks (SymPy algebraic geometry, NumPy/ML demos, etc.), I'd love to feature them in promotional content—blog posts, videos, or demo showcases. Let me know if that interests you!

[s243a]

Hello @justinhjy1004,

File upload capabilities have been significantly improved to support CSVs and other data types. For CSVs the easiest way is to use the "Import File" option in the main menu. It will detect that CSVs are data files and upload them into your browser IndexedDB to /shared/data

Alternatively, you can go to "Files & Storage", in the main menu, where you can either specify the file path, or as shown below, select the folder via the file browser. In the file browser, we omit the "/shared" prefix because that is the root of the persistent storage. Kernel specific file systems are shown under "/mnt" but files located in /mnt here reside in memory and will not persist across browser sessions. Also files in file systems specific to kernels are not sharable across kernels for other languages.

I also created an example workbook under "Browse Packages & Workbooks", that uses CSV files. This workbook is a mix of R and Python.

Ensure you have the latest code

If you are running the app via the progressive web app, use Ctrl+Shift+R to refresh, this will help to ensure you have the latest code. A probably more robust way to do this is to select "Reload App" in the menu. This is a new feature that was added.

Other New Features

• Download capabilities: Download individual files or entire folders (exported as .zip)

• Zip extraction: Upload .zip files and choose to extract them into a folder or keep as a single file

• File deletion via the file browser: See the "x" button.

Cross-Kernel File Sharing

Files uploaded to SharedVFS are automatically synchronized across all kernels. For example:

• Upload a CSV via the UI → available to Python, R, and Prolog immediately

• Write a file from Python using open('/shared/data/output.csv', 'w') → R can read it with read.csv('/shared/data/output.csv')

Persistence

Files now persist to IndexedDB instead of localStorage, which removes the 5-10MB size limit and supports binary files natively. Your files survive page reloads automatically.

---

This should make creating those example notebooks much smoother! Let me know if you run into any issues with file handling, and I'm still very interested in featuring your SymPy algebraic geometry or NumPy tutorials if you'd like to create them.

---