Skip to content

Tutorial part 6: the Analysis stage on coco_pharma_trial - #983

Merged
dwolfson merged 3 commits into
mainfrom
claude/tutorial-prose-part-6
Nov 4, 2026
Merged

dwolfson merged 3 commits into
mainfrom
claude/tutorial-prose-part-6

Conversation

@dwolfson

@dwolfson dwolfson commented Nov 3, 2026

Copy link
Copy Markdown
Owner

Part 6 of the Resource Explorer tutorial series, written from the running Explorer on coco_pharma_trial (4 Analysis).

  • Prose: walks the Analysis stage's Survey & analyses card and the questions it answers.
  • One press, owner-approved: Table Statistics Freshness › Run →, pressed once, local only. Its result is quoted: "215 of 266 tables (81%)", and step 6 names the 215 as the never-analyzed count.
  • Captures: S21, S21b and S22.
  • Outline: G33 and G34 (grades W57, W58).
  • Architect read-back: PASS. Its two requested changes are in (609e432).

Gate b284 green on main 0b456dd: harness 0 fail, boot 3, pytest 14,451 passed / 0 failed, trellis-analytics 383, tree 9bf24b34.

🤖 Generated with Claude Code

… funnel

Part 6 answers Erin's "already in the catalog?" on Discovery. The
catalog row reads a real absence, and the fingerprint row finds the
trial database a likely copy of the Quickstart's coco_pharma (the same
database registered twice). Assessment answers four questions with
nothing run on it. Analysis shows a never-measured row, then runs one
analysis on its own instead of the eight-step survey: Table Statistics
Freshness, pressed once on the project owner's word with the
Coordinator's CLEAR (2026-11-03T15:00:31Z, activity e5061e6c). That
turns two partial answers whole.

No Assessment survey is run: the Database Assessment Survey is offered
on Analysis and adds nothing on Assessment until W43 (Architect). The
outline records S21-S23, extends G13 and G23, and adds G31 (the LAST
SURVEY tile names a survey that did not run) and G32 (two wordings for
"not run"). The V6 narration draft is included; part 5's Next now links
here.

Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>
The Architect's read-back: the source line "statistics_freshness ·
measured · 215 of 266 tables (81%)" shows the never-analyzed count
under a "measured" label; all 266 relations were measured. Step 6 keeps
both quotes and says so, and the outline adds G33 (graded W57).

Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>
Step 4's documentation and well-modelled answers show the same Schema
Conventions summary. One clause now says so, from the Architect's reading
of the question catalogue (Q-0088, Q-0090), and the outline adds G34
(graded W58).

Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>
@dwolfson
dwolfson merged commit 62c5f5b into main Nov 4, 2026
4 of 5 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant