[GH-ISSUE #5164] [Design] Science Studio tabular import + column typing (GraphPad-like stats prep) #3172

Closed
opened 2026-10-31 00:37:30 +00:00 by bsakaguchi · 1 comment
Owner

Originally created by @ganttmeredith on GitHub (Oct 19, 2026).
Original GitHub issue: https://github.com/morluto/rea/issues/5164

Decision needed

Should Science Studio gain a local tabular import path for Excel (.xlsx/.xls), CSV, and delimited TXT that returns a typed column profile (continuous / categorical / string / …) before GraphPad-like tests and plots are offered?

Please approve, narrow, or decline before any integration implementation.

Proposal doc (draft PR): see linked PR once opened — docs/proposals/science-studio-tabular-data-typing.md.

Context and evidence

  • Science Studio already has a CSV-only stats service (/analyze/csv) that exposes pandas numeric columns.
  • Phase 2 clean-room stats are sketched in local investigation notes aligned with 05-stats-module.md (published stats definitions + scipy/matplotlib/plotnine — not Prism RE for MVP).
  • Users need Excel/TXT and semantic types so test/plot pickers can enable only applicable analyses, with caller overrides when inference is wrong.

Expected interaction with targets

  • Read existing files or captured data (caller-provided uploads in Science Studio)
  • Launch or attach to a process
  • Inject instrumentation or run scripts/code
  • Change process memory or registers
  • Write or change files/configuration (optional local analysis cache in the stats service — to confirm)
  • Contact network services
  • Recover protected code or work around protections

Boundary details

  • In scope: local parse, typed column profile, overrides, size limits TBD.
  • Out of scope: Prism binary RE, proprietary Prism project formats, copying Prism stats implementations, executing Excel macros.
  • REA impact is documentation/proposal unless maintainers later want an MCP table-inspect helper (separate decision).

Options under discussion

  1. Approve v1 vocabulary {continuous, categorical, string, unknown} only.
  2. Also include boolean / datetime in v1.
  3. Defer Excel; CSV/TXT only first.
  4. Decline / keep CSV numeric-only.

Before submitting

  • I searched related issues/PRs conceptually (stats Phase 2); this is a new typed-import proposal.
  • I understand contribution scope must be reviewed before integration.
  • No credentials or private samples included.
Originally created by @ganttmeredith on GitHub (Oct 19, 2026). Original GitHub issue: https://github.com/morluto/rea/issues/5164 ## Decision needed Should Science Studio gain a **local tabular import** path for **Excel (`.xlsx`/`.xls`), CSV, and delimited TXT** that returns a **typed column profile** (continuous / categorical / string / …) before GraphPad-like tests and plots are offered? **Please approve, narrow, or decline before any integration implementation.** Proposal doc (draft PR): see linked PR once opened — `docs/proposals/science-studio-tabular-data-typing.md`. ## Context and evidence - Science Studio already has a CSV-only stats service (`/analyze/csv`) that exposes pandas numeric columns. - Phase 2 clean-room stats are sketched in local investigation notes aligned with `05-stats-module.md` (published stats definitions + scipy/matplotlib/plotnine — **not** Prism RE for MVP). - Users need Excel/TXT and semantic types so test/plot pickers can enable only applicable analyses, with caller overrides when inference is wrong. ## Expected interaction with targets - [x] Read existing files or captured data (caller-provided uploads in Science Studio) - [ ] Launch or attach to a process - [ ] Inject instrumentation or run scripts/code - [ ] Change process memory or registers - [x] Write or change files/configuration (optional local analysis cache in the stats service — to confirm) - [ ] Contact network services - [ ] Recover protected code or work around protections ## Boundary details - **In scope:** local parse, typed column profile, overrides, size limits TBD. - **Out of scope:** Prism binary RE, proprietary Prism project formats, copying Prism stats implementations, executing Excel macros. - REA impact is documentation/proposal unless maintainers later want an MCP table-inspect helper (separate decision). ## Options under discussion 1. Approve v1 vocabulary `{continuous, categorical, string, unknown}` only. 2. Also include `boolean` / `datetime` in v1. 3. Defer Excel; CSV/TXT only first. 4. Decline / keep CSV numeric-only. ## Before submitting - [x] I searched related issues/PRs conceptually (stats Phase 2); this is a new typed-import proposal. - [x] I understand contribution scope must be reviewed before integration. - [x] No credentials or private samples included.
Author
Owner

@ganttmeredith commented on GitHub (Oct 19, 2026):

Closing: filed against public REA by mistake. Science Studio is a separate local product; this design discussion does not belong on morluto/rea.

<!-- gh-comment-id:6258765163 --> @ganttmeredith commented on GitHub (Oct 19, 2026): Closing: filed against public REA by mistake. Science Studio is a separate local product; this design discussion does not belong on morluto/rea.
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
morluto/rea#3172
No description provided.