feat(layout-review): address review findings on pages 7-12

Find Duplicates (01_deduplicator):
- Delete the redundant outer Options wrapper; surface threshold +
  survivor rule directly, push the rest behind a single Advanced pane.
- Disambiguate competing primaries: top result is an auto-resolved
  preview (secondary download), review decisions are the single primary.
- Plain-English match labels (exact / approximate); clarify the third.
- Lift the match-card caption to a one-time instruction; note delimiter
  is delimited-text-only.

Quality Check (08_validator_reporter) — stub:
- Remove the dead disabled "Load rules file (JSON)" uploader so the
  stub invites a single action; keep the informative feature list.

Map Columns (05_column_mapper):
- Regroup schema -> mapping -> strategy/advanced (core task contiguous).
- Make preset-vs-Advanced precedence legible (Custom + modified marker).
- Adopt the compact file-intake banner; drop the duplicate resolved-
  mapping table; fix the add-row gutter style.

Combine Files (07_multi_file_merger) — stub:
- Actually disable the Merge CTA (add the disabled attribute).

PDF to CSV (10_pdf_extractor):
- Drop page/raw from the default preview to match export + fix the
  horizontal clip; surface raw via per-row affordance + overflow-x.
- Move the column selector above the download button; give auto-excluded
  rows a reason; align the files card to Home; de-dupe the row count.

Automated Workflows (09_pipeline_runner):
- Replace hand-edited JSON step config with per-step control expanders;
  JSON moved behind Advanced import/export.
- Editing the table marks the mode modified; fold the empty error column
  into the status pill; render summaries as plain English; collapse the
  explainer by default.

Cross-cutting items (stub standardization on page 10, shared disabled-
field token, remaining intake rollout) deferred to a holistic pass.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-06-08 16:35:46 +00:00
parent 563d845b70
commit cf31d9ef14
6 changed files with 302 additions and 183 deletions

View File

@@ -41,7 +41,8 @@
<button class="dt-btn dt-btn-tertiary" title="Remove"></button>
</div>
<!-- Delimiter selector (CSV) -->
<!-- Delimiter selector — delimited-text only (CSV/TSV); omitted for XLSX/XLS.
Shown here because the staged file is customers_export.csv. -->
<div class="dt-field" style="max-width:320px">
<label class="dt-label">Delimiter</label>
<div class="dt-select">Comma (,)</div>
@@ -67,32 +68,33 @@
</div>
</details>
<!-- Options expander -->
<!-- Basic controls (visible by default) -->
<div class="dt-cols-2">
<div class="dt-field"><label class="dt-label">Match threshold</label>
<div class="dt-slider"><div class="track"><div class="fill" style="width:70%"></div><div class="knob" style="left:70%"></div></div><div class="val">85</div></div>
<div class="dt-help-text">Higher means rows must look more alike to count as a duplicate.</div></div>
<div class="dt-field"><label class="dt-label">When duplicates are found, keep</label>
<div class="dt-select">the most-complete row</div>
<div class="dt-help-text">Which row survives in each group of duplicates.</div></div>
</div>
<!-- Advanced options (single expander; basics live above) -->
<details class="dt-expander">
<summary>Options</summary>
<summary>Advanced options</summary>
<div class="dt-expander-body">
<details class="dt-expander" style="margin-top:0">
<summary>Advanced Options</summary>
<div class="dt-expander-body">
<div class="dt-cols-2">
<div>
<div class="dt-field"><label class="dt-label">Match on columns</label>
<div class="dt-multiselect"><span class="dt-ms-placeholder">Leave empty to auto-detect</span></div></div>
<div class="dt-field"><label class="dt-label">Strong keys</label>
<div class="dt-multiselect"><span class="dt-ms-chip">email <span class="x"></span></span></div></div>
<div class="dt-field"><label class="dt-label">Fuzzy columns</label>
<div class="dt-multiselect"><span class="dt-ms-chip">name <span class="x"></span></span></div></div>
</div>
<div>
<div class="dt-field"><label class="dt-label">Fuzzy algorithm</label><div class="dt-select">jaro_winkler</div></div>
<div class="dt-field"><label class="dt-label">Similarity threshold</label>
<div class="dt-slider"><div class="track"><div class="fill" style="width:70%"></div><div class="knob" style="left:70%"></div></div><div class="val">85</div></div></div>
<div class="dt-field"><label class="dt-label">Survivor rule</label><div class="dt-select">most-complete</div></div>
</div>
</div>
<div class="dt-check on" style="margin-top:6px"><span class="box"><span class="dt-mi">check</span></span> Merge mode — fill missing fields in the surviving row</div>
<p class="dt-help-text" style="margin-top:0">Leave these empty to auto-detect which columns to compare. Otherwise, list the columns that must match <strong>exactly</strong> and the ones that only need to match <strong>approximately</strong> — together these are the columns used to find duplicates.</p>
<div class="dt-cols-2">
<div>
<div class="dt-field"><label class="dt-label">Columns that must match exactly</label>
<div class="dt-multiselect"><span class="dt-ms-chip">email <span class="x"></span></span></div></div>
<div class="dt-field"><label class="dt-label">Columns to match approximately</label>
<div class="dt-multiselect"><span class="dt-ms-chip">name <span class="x"></span></span></div></div>
</div>
</details>
<div>
<div class="dt-field"><label class="dt-label">Approximate-match algorithm</label><div class="dt-select">jaro_winkler</div></div>
</div>
</div>
<div class="dt-check on" style="margin-top:6px"><span class="box"><span class="dt-mi">check</span></span> Merge mode — fill missing fields in the surviving row</div>
</div>
</details>
@@ -109,8 +111,9 @@
<div class="dt-metric"><div class="label">Match groups</div><div class="value">147</div></div>
<div class="dt-metric"><div class="label">Rows kept</div><div class="value">18,130</div></div>
</div>
<p class="dt-caption">Preview of an auto-resolved run: each group keeps its auto-picked survivor. Review the groups below to override any pending picks before the final download.</p>
<div class="dt-btn-row" style="max-width:560px">
<button class="dt-btn dt-btn-primary">Download deduplicated CSV</button>
<button class="dt-btn">Download auto-resolved CSV</button>
<button class="dt-btn">Download removed rows</button>
</div>
@@ -123,6 +126,7 @@
<button class="dt-btn">Reject All</button>
<button class="dt-btn">Clear Decisions</button>
</div>
<p class="dt-caption" style="margin-top:8px">Differing columns are highlighted. The survivor row is kept; uncheck a row to split it out of the group.</p>
<!-- Match group card 1 -->
<div class="dt-match-card">
@@ -140,7 +144,6 @@
</tbody>
</table>
</div>
<p class="dt-caption">Differing columns highlighted. The survivor row is kept; uncheck rows to split the group.</p>
</div>
</div>
@@ -163,8 +166,8 @@
</div>
</div>
<p class="dt-caption" style="margin-top:14px">Decisions: 1 merged, 1 pending</p>
<button class="dt-btn dt-btn-primary dt-btn-block" style="margin-top:8px">Apply Review Decisions &amp; Download</button>
<p class="dt-caption" style="margin-top:14px">Decisions: 1 merged, 1 pending · Pending groups keep their auto-picked survivor unless you review them.</p>
<button class="dt-btn dt-btn-primary dt-btn-block" style="margin-top:8px">Apply Review Decisions &amp; Download Final CSV</button>
<!-- Processing log -->
<details class="dt-expander" style="margin-top:18px">