Skip to content

Commit 9f3377b

Browse files
author
Michael Aydinbas
committed
Untrack the wayfinder planning folders
The migration map is closed and the reporting map is planning material, not part of the pipeline. Both are now gitignored and kept outside the repository, which is public. The five files that linked into them no longer do: the links pointed at paths that would 404 for any reader, so they are now plain references to the migration decision record.
1 parent ab1b8a3 commit 9f3377b

6 files changed

Lines changed: 18 additions & 13 deletions

File tree

‎.gitignore‎

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -75,3 +75,5 @@ secrets/
7575
.opencode
7676
CLAUDE.local.md
7777
coverage.xml
78+
docs/wayfinder/
79+
docs/wayfinder-reporting/

‎docs/archive/MR_DESCRIPTION.md‎

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -19,7 +19,7 @@ first -- the cell-by-cell triage over 254 trackers, the production verification
1919
run, the dependency audit, the performance profile, R's retirement, and the
2020
documentation overhaul. What is listed at the end stayed open and never blocked
2121
it; it carries on under
22-
[docs/wayfinder/map.md](../wayfinder/map.md).
22+
the migration decision record, kept outside this repository.
2323

2424
---
2525

@@ -493,7 +493,8 @@ columns left for that month.
493493
## Where verification stands
494494

495495
Current baseline: `output/comparison/2026-08-21T215325Z`, 254 trackers.
496-
Earlier counts on the wayfinder map were measured against smaller tracker sets
496+
Earlier counts in the migration decision record were measured against smaller
497+
tracker sets
497498
and should be read as historical.
498499

499500
| Stage | Mismatches | Unclassified |
@@ -514,7 +515,7 @@ the pipeline *should* publish, not a difference nobody understands.
514515

515516
# Open at the time of the merge
516517

517-
**None of this blocked the merge.** It is the work the wayfinder map carries
518+
**None of this blocked the merge.** It is the work the migration record carries
518519
forward. This list is frozen as of 2026-08-24 and is not updated -- the map is
519520
the live record.
520521

@@ -813,4 +814,4 @@ Cleaned `unclassified` is back to **16**, with the four-stage totals unchanged.
813814
predating the current `src/` layout — not in git, pending a decision to delete.
814815

815816
Full history, per-ticket evidence and every decision:
816-
[docs/wayfinder/map.md](../wayfinder/map.md).
817+
the migration decision record, kept outside this repository.

‎docs/archive/PYTHON_IMPROVEMENTS.md‎

Lines changed: 4 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -3,11 +3,12 @@
33
> **ARCHIVED — frozen, and known to be incomplete.** See
44
> [`docs/archive/README.md`](README.md). This was a hand-maintained list
55
> written early in the migration, and the completeness audit
6-
> ([wayfinder ticket 7](../wayfinder/tickets/07-pipeline-completeness-audit.md))
6+
> (the migration completeness audit)
77
> found it **undercounting the known divergences**. It was superseded by two
88
> things that derive rather than list: the cause registry in
99
> `src/a4d/migration/compare.py`, where every divergence is tied to a mechanism
10-
> traced in real source workbooks, and the wayfinder map's Decisions-so-far.
10+
> traced in real source workbooks, and the migration's own decision record, which is kept outside this
11+
> repository.
1112
> Read those for the real inventory; read this for early context only. Frozen
1213
> at tag `migration-archive-frozen`.
1314
@@ -137,7 +138,7 @@ for h1, h2 in zip(header_1, header_2, strict=True):
137138

138139
**Status**: ✅ Improved in Python
139140

140-
Three distinct date-parsing patterns surfaced during product-pipeline diff investigation against R goldens. In each case Python yields a more correct result than R; no R-parity fix is warranted. Investigation: the underlying notebook (`Ali_internship/residual_dig.ipynb`) was never committed to the repo and no longer exists; the surviving record of this analysis is [`Product pipeline parity presentation.pdf`](Product%20pipeline%20parity%20presentation.pdf), kept until the automated comparison script (see [ticket 2](../wayfinder/tickets/02-documentation-strategy.md)) reproduces and supersedes its numbers.
141+
Three distinct date-parsing patterns surfaced during product-pipeline diff investigation against R goldens. In each case Python yields a more correct result than R; no R-parity fix is warranted. Investigation: the underlying notebook (`Ali_internship/residual_dig.ipynb`) was never committed to the repo and no longer exists; the surviving record of this analysis is [`Product pipeline parity presentation.pdf`](Product%20pipeline%20parity%20presentation.pdf), kept until the automated comparison script reproduces and supersedes its numbers.
141142

142143
### 5.1 "Sept" → "Sep" month abbreviation
143144

‎docs/archive/README.md‎

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -21,7 +21,7 @@ fact stated here should be assumed stale unless the code agrees.
2121
| What the pipeline does and how to run it | [`README.md`](../../README.md), [`SETUP.md`](../../SETUP.md) |
2222
| Module-by-module map, CLI, output layout | [`docs/CLAUDE.md`](../CLAUDE.md) |
2323
| Why a particular cleaning rule exists | the docstring on the rule itself |
24-
| The migration's decision history | [`docs/wayfinder/`](../wayfinder/map.md) |
24+
| The migration's decision history | Kept outside this repository |
2525
| The old R source | `git show r-archive-removed^:r-archive/R/<file>` |
2626

2727
## What is here
@@ -33,7 +33,7 @@ fact stated here should be assumed stale unless the code agrees.
3333
plus the R-to-Python step mapping.
3434
- **`PYTHON_IMPROVEMENTS.md`** — cases where Python was judged more correct
3535
than R. Superseded as a *catalogue* by the cause registry in
36-
`src/a4d/migration/compare.py` and by the wayfinder map's decisions, both of
36+
`src/a4d/migration/compare.py` and by the migration decision record, both of
3737
which are evidence-backed where this document is not.
3838
- **`PRODUCT_DATA_PIPELINE_FEATURE.md`** — the original proposal for the
3939
product arm, written before it was built. Already marked obsolete in its own
@@ -44,8 +44,9 @@ fact stated here should be assumed stale unless the code agrees.
4444
## Why they were kept rather than deleted
4545

4646
The migration's conclusions rest on cell-by-cell evidence gathered over 56
47-
wayfinder tickets. Those conclusions are now expressed where they belong — in
48-
the code's own docstrings, in the wayfinder map, and in the comparison tool's
47+
migration tickets. Those conclusions are now expressed where they belong — in
48+
the code's own docstrings, in the migration decision record, and in the
49+
comparison tool's
4950
cause registry. These documents are the working papers behind them. They cost
5051
nothing to keep and they are the only place some of the intermediate reasoning
5152
survives.

‎scripts/compare_outputs.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1068,7 +1068,7 @@ def _compare_arm(py_root: Path, r_dir: Path, py_dir: Path, stage: Stage) -> Dire
10681068
[bold]Cell divergence[/bold] -- rows matched across R and Python (via the arm's \
10691069
row-alignment key), diffed value by value. The most granular value comparison -- but it's only \
10701070
meaningful once Row-key divergence above confirms the key actually paired the rows; 0 here can \
1071-
mean "everything agreed" or "nothing was paired to compare" (ticket 15/17 on the wayfinder map).\
1071+
mean "everything agreed" or "nothing was paired to compare".\
10721072
"""
10731073

10741074

‎src/a4d/validate/snapshot.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -16,7 +16,7 @@
1616
*supposed* to move.
1717
1818
The digest is written beside the tracker corpus rather than committed, so no
19-
per-clinic shape is published; see ``docs/wayfinder/tickets/09-snapshot-
19+
per-clinic shape is published; see the planning record kept outside this
2020
regression-tests.md`` for that decision.
2121
"""
2222

0 commit comments

Comments
 (0)