backend/main.py

@@ -251,6 +251,13 @@ async def metadata_backfill_wizard_page(request: Request):
     return templates.TemplateResponse("admin/metadata_backfill.html", {"request": request})
 
 
+@app.get("/settings/developer/project-tidy", response_class=HTMLResponse)
+async def project_tidy_page(request: Request):
+    """Tidy duplicate-looking projects: detect by fuzzy name match, merge
+    by clicking through pairs (Phase 5b)."""
+    return templates.TemplateResponse("admin/project_tidy.html", {"request": request})
+
+
 @app.get("/modems", response_class=HTMLResponse)
 async def modems_page(request: Request):
     """Field modems management dashboard"""

backend/routers/projects.py

@@ -729,6 +729,49 @@ async def project_merge_preview(
     }
 
 
+@router.get("/admin/duplicate_pairs")
+async def get_duplicate_pairs(
+    threshold: float = 0.85,
+    max_pairs: int = 200,
+    db: Session = Depends(get_db),
+):
+    """Return all active-project pairs whose names fuzzy-match above the
+    threshold.  Used by the Tidy page to surface duplicates that would
+    otherwise have to be hunted down one at a time.
+
+    Each pair carries a suggested merge-target with the reasoning so the
+    operator can decide direction with one click.
+    """
+    from backend.services import project_tidy as pt
+    pairs = pt.find_duplicate_pairs(db, threshold=threshold, max_pairs=max_pairs)
+
+    def _ps(p):
+        return {
+            "id":               p.id,
+            "name":             p.name,
+            "project_number":   p.project_number,
+            "client_name":      p.client_name,
+            "source":           p.source,
+            "status":           p.status,
+            "location_count":   p.location_count,
+            "assignment_count": p.assignment_count,
+        }
+
+    return {
+        "pairs": [
+            {
+                "a":                   _ps(pair.a),
+                "b":                   _ps(pair.b),
+                "score":               round(pair.score, 3),
+                "suggested_target_id": pair.suggested_target_id,
+                "reason":              pair.reason,
+            }
+            for pair in pairs
+        ],
+        "threshold": threshold,
+    }
+
+
 @router.post("/{source_id}/merge_into")
 async def project_merge_execute(
     source_id: str,

backend/services/project_tidy.py

@@ -0,0 +1,235 @@
+"""
+project_tidy.py — find duplicate-looking projects + offer bulk merge.
+
+The metadata-backfill parser is good at clustering events into candidate
+projects but doesn't compare its proposed project names against EACH OTHER
+(it only checks against existing terra-view projects).  After a bulk
+apply, you can end up with many near-duplicate projects — typo variants,
+abbreviation differences, etc.  This module surfaces them as pairs the
+operator can merge.
+
+Pairs vs clusters: a fully-connected group like (A, B, C) where each pair
+scores >= threshold becomes 3 pairs.  The operator has to do 2 merges to
+fully consolidate.  We don't try to be smarter about transitive grouping —
+in practice operators want to review the highest-similarity pair first
+anyway, and the list re-computes after each merge.
+
+Public API:
+  find_duplicate_pairs(db, *, threshold=0.85, max_pairs=200) → list[DuplicatePair]
+"""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass
+from typing import Optional
+
+import rapidfuzz
+from sqlalchemy import func
+from sqlalchemy.orm import Session
+
+from backend.models import (
+    Project,
+    MonitoringLocation,
+    UnitAssignment,
+)
+from backend.services.metadata_backfill import _normalise as _meta_normalise
+
+log = logging.getLogger("backend.services.project_tidy")
+
+
+DEFAULT_THRESHOLD       = 0.85   # WRatio similarity above which we surface a pair
+DEFAULT_MAX_PAIRS       = 200    # Cap the result list to keep response small
+MIN_NORMALISED_LENGTH   = 4      # Skip projects whose normalised name is too short
+                                 # to fuzzy-match safely (avoids "1" / "1" pairs).
+
+
+@dataclass
+class ProjectSummary:
+    id:                str
+    name:              str
+    project_number:    Optional[str]
+    client_name:       Optional[str]
+    source:            str            # 'manual' | 'metadata_backfill' | ...
+    status:            str
+    location_count:    int
+    assignment_count:  int
+    event_count_total: int            # approx — sum across assignments
+
+
+@dataclass
+class DuplicatePair:
+    a:           ProjectSummary
+    b:           ProjectSummary
+    score:       float
+    suggested_target_id: str          # the recommended "keep" side
+    reason:      str                  # why we picked that target
+
+
+# ── Helpers ──────────────────────────────────────────────────────────────────
+
+
+def _normalise_project_name(name: str) -> str:
+    """Project-name normalisation for tidy comparison.
+
+    Reuses the metadata_backfill normaliser (lowercase, punctuation→space,
+    collapse whitespace).  Returns "" for None or all-punctuation names.
+    """
+    return _meta_normalise(name)
+
+
+def _summarise_projects(db: Session) -> list[ProjectSummary]:
+    """One row per active project with cached counts.  Excludes deleted."""
+    projects = (
+        db.query(Project)
+        .filter(Project.status != "deleted")
+        .all()
+    )
+
+    # Bulk lookup: assignment counts + location counts per project.
+    loc_counts: dict[str, int] = dict(
+        db.query(MonitoringLocation.project_id, func.count(MonitoringLocation.id))
+        .filter(MonitoringLocation.project_id.in_([p.id for p in projects]) if projects else False)
+        .group_by(MonitoringLocation.project_id)
+        .all()
+    )
+    asgn_counts: dict[str, int] = dict(
+        db.query(UnitAssignment.project_id, func.count(UnitAssignment.id))
+        .filter(UnitAssignment.project_id.in_([p.id for p in projects]) if projects else False)
+        .group_by(UnitAssignment.project_id)
+        .all()
+    )
+
+    summaries: list[ProjectSummary] = []
+    for p in projects:
+        summaries.append(ProjectSummary(
+            id                = p.id,
+            name              = p.name,
+            project_number    = p.project_number,
+            client_name       = p.client_name,
+            source            = None,                       # filled below per assignment
+            status            = p.status or "active",
+            location_count    = loc_counts.get(p.id, 0),
+            assignment_count  = asgn_counts.get(p.id, 0),
+            event_count_total = 0,                          # not cheap to compute here; left 0
+        ))
+
+    # Determine each project's dominant assignment source.  Used to break ties
+    # when picking the "keep" target — prefer manual over parser-created.
+    rows = (
+        db.query(UnitAssignment.project_id, UnitAssignment.source, func.count(UnitAssignment.id))
+        .group_by(UnitAssignment.project_id, UnitAssignment.source)
+        .all()
+    )
+    by_proj_src: dict[str, dict[str, int]] = {}
+    for proj_id, src, cnt in rows:
+        by_proj_src.setdefault(proj_id, {})[src or "manual"] = cnt
+    for s in summaries:
+        src_map = by_proj_src.get(s.id, {})
+        if not src_map:
+            s.source = "manual"
+        else:
+            # Dominant source (most assignments).
+            s.source = max(src_map.items(), key=lambda kv: kv[1])[0]
+
+    return summaries
+
+
+def _pick_target(a: ProjectSummary, b: ProjectSummary) -> tuple[str, str]:
+    """Decide which project should be the merge target (the one we keep).
+
+    Priorities (in order):
+      1. The one with `source='manual'` over `source='metadata_backfill'`
+         — operator-curated projects beat parser-created ones.
+      2. The one with a populated `project_number`.
+      3. The one with more locations (more curation history).
+      4. The one with more assignments.
+      5. The one with the shorter, cleaner name (tiebreaker).
+
+    Returns (target_id, reason_string).
+    """
+    # 1. Source provenance.
+    a_manual = a.source == "manual"
+    b_manual = b.source == "manual"
+    if a_manual and not b_manual:
+        return a.id, "A is manually-created; B is parser-created"
+    if b_manual and not a_manual:
+        return b.id, "B is manually-created; A is parser-created"
+
+    # 2. project_number populated.
+    if a.project_number and not b.project_number:
+        return a.id, "A has a project_number; B doesn't"
+    if b.project_number and not a.project_number:
+        return b.id, "B has a project_number; A doesn't"
+
+    # 3. More locations.
+    if a.location_count > b.location_count:
+        return a.id, f"A has more locations ({a.location_count} vs {b.location_count})"
+    if b.location_count > a.location_count:
+        return b.id, f"B has more locations ({b.location_count} vs {a.location_count})"
+
+    # 4. More assignments.
+    if a.assignment_count > b.assignment_count:
+        return a.id, f"A has more assignments ({a.assignment_count} vs {b.assignment_count})"
+    if b.assignment_count > a.assignment_count:
+        return b.id, f"B has more assignments ({b.assignment_count} vs {a.assignment_count})"
+
+    # 5. Shorter name (less likely to have baked-in junk).
+    if len(a.name) <= len(b.name):
+        return a.id, "A has the shorter / cleaner name"
+    return b.id, "B has the shorter / cleaner name"
+
+
+# ── Public ───────────────────────────────────────────────────────────────────
+
+
+def find_duplicate_pairs(
+    db: Session,
+    *,
+    threshold: float = DEFAULT_THRESHOLD,
+    max_pairs: int = DEFAULT_MAX_PAIRS,
+) -> list[DuplicatePair]:
+    """Compute all project-pair similarities above `threshold`.
+
+    O(N^2) over the project count — fine up to ~500 projects; beyond that
+    we'd want a blocked / token-indexed approach.  In practice
+    `metadata_backfill` projects tend to share tokens, so a simple
+    pre-filter (skip pairs that share NO tokens) would cheaply cut the
+    inner loop.  Deferred until profiling motivates it.
+    """
+    summaries = _summarise_projects(db)
+
+    # Pre-compute normalised names; skip too-short ones.
+    norm_by_id: dict[str, str] = {}
+    candidates: list[ProjectSummary] = []
+    for s in summaries:
+        n = _normalise_project_name(s.name)
+        if len(n) < MIN_NORMALISED_LENGTH:
+            continue
+        norm_by_id[s.id] = n
+        candidates.append(s)
+
+    pairs: list[DuplicatePair] = []
+    n = len(candidates)
+    for i in range(n):
+        a = candidates[i]
+        a_norm = norm_by_id[a.id]
+        for j in range(i + 1, n):
+            b = candidates[j]
+            b_norm = norm_by_id[b.id]
+            score = rapidfuzz.fuzz.WRatio(a_norm, b_norm) / 100.0
+            if score < threshold:
+                continue
+            target_id, reason = _pick_target(a, b)
+            pairs.append(DuplicatePair(
+                a                   = a,
+                b                   = b,
+                score               = score,
+                suggested_target_id = target_id,
+                reason              = reason,
+            ))
+
+    # Sort by score desc, then by total content (more data → review first).
+    pairs.sort(key=lambda p: (-p.score, -(p.a.assignment_count + p.b.assignment_count)))
+
+    return pairs[:max_pairs]

templates/admin/metadata_backfill.html

@@ -595,9 +595,24 @@ async function _apply(clusterIds) {
         });
         if (!r.ok) throw new Error('HTTP ' + r.status);
         const d = await r.json();
-        const sub = `${d.applied} applied · ${d.project_ids_created.length} new project(s) · ${d.location_ids_created.length} new location(s)` + (d.failed.length ? ` · ${d.failed.length} failed` : '');
+        const failedCount = (d.failed || []).length;
-        _showToast(`${d.applied} cluster${d.applied === 1 ? '' : 's'} applied`, sub, d.failed.length ? 'error' : 'success');
+
+        // Three states:
+        //   total success — applied N, no failures              → green toast, 4s
+        //   partial       — applied N, M failures               → red toast + modal listing reasons
+        //   total failure — applied 0, failures                 → red toast + modal
+        if (failedCount === 0) {
+            const sub = `${d.applied} applied · ${d.project_ids_created.length} new project(s) · ${d.location_ids_created.length} new location(s)`;
+            _showToast(`${d.applied} cluster${d.applied === 1 ? '' : 's'} applied`, sub, 'success');
             _hideToast(4000);
+        } else {
+            const title = d.applied > 0
+                ? `${d.applied} applied, ${failedCount} failed`
+                : `Apply failed — ${failedCount} cluster${failedCount === 1 ? '' : 's'} could not be applied`;
+            _showToast(title, 'See the details panel.', 'error');
+            _hideToast(6000);
+            _showFailureDetails(d.failed);
+        }
         await runScan(true);  // refresh
     } catch (e) {
         _showToast('Apply failed', e.message, 'error');
@@ -605,6 +620,46 @@ async function _apply(clusterIds) {
     }
 }
 
+// Modal-ish panel that lists each failed cluster with its server-side
+// reason.  Common failure modes seen in dev: missing DB tables after a
+// stale schema, blocking conflicts that slipped past the front-end guard,
+// rapidfuzz/SQLAlchemy edge cases.  Operator can dismiss and either
+// retry the cluster, skip it, or fix the underlying issue.
+function _showFailureDetails(failed) {
+    let panel = document.getElementById('apply-failure-panel');
+    if (!panel) {
+        panel = document.createElement('div');
+        panel.id = 'apply-failure-panel';
+        panel.className = 'fixed bottom-6 left-6 right-6 sm:right-auto sm:max-w-xl bg-white dark:bg-slate-800 rounded-xl shadow-2xl border border-red-200 dark:border-red-800 p-4 z-40';
+        document.body.appendChild(panel);
+    }
+    const rows = failed.map(f => `
+        <li class="flex items-start gap-2 text-sm border-l-2 border-red-300 dark:border-red-700 pl-3 py-1">
+            <code class="font-mono text-xs text-gray-500 dark:text-gray-400">${(f.cluster_id || '').slice(0, 8)}…</code>
+            <span class="text-gray-800 dark:text-gray-200 flex-1">${_esc(f.reason || '(no reason)')}</span>
+        </li>
+    `).join('');
+    panel.innerHTML = `
+        <div class="flex items-start justify-between gap-3 mb-2">
+            <h4 class="font-semibold text-gray-900 dark:text-white">
+                <svg class="w-5 h-5 inline text-red-500 -mt-0.5 mr-1" fill="none" stroke="currentColor" viewBox="0 0 24 24"><path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M12 9v2m0 4h.01M21 12a9 9 0 11-18 0 9 9 0 0118 0z"/></svg>
+                ${failed.length} cluster${failed.length === 1 ? '' : 's'} failed to apply
+            </h4>
+            <button onclick="document.getElementById('apply-failure-panel').remove()"
+                    class="text-gray-400 hover:text-gray-600 dark:hover:text-gray-200">
+                <svg class="w-5 h-5" fill="none" stroke="currentColor" viewBox="0 0 24 24"><path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M6 18L18 6M6 6l12 12"/></svg>
+            </button>
+        </div>
+        <ul class="space-y-1 max-h-64 overflow-y-auto">${rows}</ul>
+        <p class="mt-3 text-xs text-gray-500 dark:text-gray-400">
+            Common causes: missing DB schema (run the migration sweep), blocking conflict
+            with an existing UnitAssignment, or a UNIQUE constraint collision on the
+            project name.  Re-scan and the failed clusters reappear as pending — fix the
+            underlying issue and retry.
+        </p>
+    `;
+}
+
 async function applyOne(clusterId)   { return _apply([clusterId]); }
 
 async function applyBulkHighConfidence() {

templates/admin/project_tidy.html

@@ -0,0 +1,267 @@
+{% extends "base.html" %}
+
+{% block title %}Project Tidy - Seismo Fleet Manager{% endblock %}
+
+{% block content %}
+<!-- Breadcrumb -->
+<div class="mb-6">
+    <nav class="flex items-center space-x-2 text-sm">
+        <a href="/settings" class="text-seismo-orange hover:text-seismo-navy flex items-center">
+            <svg class="w-4 h-4 mr-1" fill="none" stroke="currentColor" viewBox="0 0 24 24">
+                <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M15 19l-7-7 7-7"></path>
+            </svg>
+            Settings
+        </a>
+        <svg class="w-4 h-4 text-gray-400" fill="none" stroke="currentColor" viewBox="0 0 24 24">
+            <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M9 5l7 7-7 7"></path>
+        </svg>
+        <span class="text-gray-900 dark:text-white font-medium">Project Tidy</span>
+    </nav>
+</div>
+
+<!-- Header -->
+<div class="mb-6">
+    <h1 class="text-3xl font-bold text-gray-900 dark:text-white">Project Tidy</h1>
+    <p class="text-gray-600 dark:text-gray-400 mt-1">
+        Find duplicate-looking projects via fuzzy name matching, then merge them with one click.
+        Useful after the metadata-backfill parser creates near-duplicates from operator name variations.
+    </p>
+</div>
+
+<!-- Controls -->
+<div class="bg-white dark:bg-slate-800 rounded-xl shadow-lg p-4 mb-5">
+    <div class="flex flex-wrap items-end gap-3">
+        <div class="flex flex-col gap-1">
+            <label class="text-xs text-gray-500 dark:text-gray-400">Similarity threshold</label>
+            <select id="threshold" onchange="runScan()"
+                    class="px-3 py-1.5 text-sm border border-gray-300 dark:border-gray-600 rounded-lg bg-white dark:bg-slate-700 text-gray-900 dark:text-white">
+                <option value="0.95">≥ 95% — near-identical only (typos)</option>
+                <option value="0.90">≥ 90% — close variants</option>
+                <option value="0.85" selected>≥ 85% — fuzzy match floor (recommended)</option>
+                <option value="0.80">≥ 80% — aggressive (more false positives)</option>
+            </select>
+        </div>
+        <button onclick="runScan()"
+                class="ml-auto px-4 py-1.5 text-sm bg-seismo-orange text-white rounded-lg hover:bg-seismo-navy transition-colors">
+            ↻ Scan for duplicates
+        </button>
+    </div>
+</div>
+
+<!-- Results -->
+<div id="results" class="space-y-3">
+    <div class="text-center py-12 text-gray-500 dark:text-gray-400">
+        Click "Scan for duplicates" to find pairs.
+    </div>
+</div>
+
+<!-- Apply progress toast -->
+<div id="tidy-toast" class="hidden fixed bottom-6 right-6 bg-white dark:bg-slate-800 rounded-xl shadow-2xl border border-gray-200 dark:border-gray-700 p-4 z-50 max-w-md">
+    <div class="flex items-center gap-3">
+        <div id="toast-icon" class="shrink-0">
+            <div class="animate-spin rounded-full h-6 w-6 border-b-2 border-seismo-orange"></div>
+        </div>
+        <div class="flex-1">
+            <p id="toast-message" class="text-sm font-medium text-gray-900 dark:text-white">Working…</p>
+            <p id="toast-sub" class="text-xs text-gray-500 dark:text-gray-400 mt-0.5"></p>
+        </div>
+    </div>
+</div>
+
+<script>
+let _pairs = [];
+
+function _esc(s) {
+    if (s == null) return '';
+    return String(s).replace(/&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;').replace(/"/g, '&quot;');
+}
+
+function _sourceBadge(source) {
+    if (source === 'metadata_backfill' || source === 'metadata_backfill_swap') {
+        return '<span class="px-1.5 py-0.5 rounded text-xs bg-amber-100 text-amber-800 dark:bg-amber-900/30 dark:text-amber-300" title="Auto-created by the metadata-backfill parser">parser</span>';
+    }
+    return '<span class="px-1.5 py-0.5 rounded text-xs bg-blue-100 text-blue-800 dark:bg-blue-900/30 dark:text-blue-300" title="Manually created via the UI">manual</span>';
+}
+
+function _showToast(message, sub, kind) {
+    const toast = document.getElementById('tidy-toast');
+    const icon  = document.getElementById('toast-icon');
+    document.getElementById('toast-message').textContent = message;
+    document.getElementById('toast-sub').textContent     = sub || '';
+    if (kind === 'success') {
+        icon.innerHTML = '<svg class="w-6 h-6 text-green-500" fill="none" stroke="currentColor" viewBox="0 0 24 24"><path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M5 13l4 4L19 7"></path></svg>';
+    } else if (kind === 'error') {
+        icon.innerHTML = '<svg class="w-6 h-6 text-red-500" fill="none" stroke="currentColor" viewBox="0 0 24 24"><path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M12 9v2m0 4h.01M21 12a9 9 0 11-18 0 9 9 0 0118 0z"></path></svg>';
+    } else {
+        icon.innerHTML = '<div class="animate-spin rounded-full h-6 w-6 border-b-2 border-seismo-orange"></div>';
+    }
+    toast.classList.remove('hidden');
+}
+
+function _hideToast(after) {
+    setTimeout(() => document.getElementById('tidy-toast').classList.add('hidden'), after || 3000);
+}
+
+async function runScan() {
+    const results = document.getElementById('results');
+    results.innerHTML = '<div class="text-center py-12 text-gray-500 dark:text-gray-400"><div class="animate-spin rounded-full h-8 w-8 border-b-2 border-seismo-orange mx-auto mb-3"></div>Scanning…</div>';
+    const threshold = document.getElementById('threshold').value;
+    try {
+        const r = await fetch(`/api/projects/admin/duplicate_pairs?threshold=${threshold}`);
+        if (!r.ok) throw new Error('HTTP ' + r.status);
+        const d = await r.json();
+        _pairs = d.pairs || [];
+        render();
+    } catch (e) {
+        results.innerHTML = `<div class="bg-white dark:bg-slate-800 rounded-xl shadow-lg p-6 text-center text-red-500">Scan failed: ${_esc(e.message)}</div>`;
+    }
+}
+
+function render() {
+    const results = document.getElementById('results');
+    if (_pairs.length === 0) {
+        results.innerHTML = `<div class="bg-white dark:bg-slate-800 rounded-xl shadow-lg p-8 text-center">
+            <svg class="w-16 h-16 mx-auto mb-4 text-green-500" fill="none" stroke="currentColor" viewBox="0 0 24 24">
+                <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M9 12l2 2 4-4m6 2a9 9 0 11-18 0 9 9 0 0118 0z"></path>
+            </svg>
+            <h3 class="text-lg font-semibold text-gray-900 dark:text-white mb-1">✨ No duplicates above the threshold</h3>
+            <p class="text-gray-500 dark:text-gray-400">Lower the threshold or call it good.</p>
+        </div>`;
+        return;
+    }
+
+    const summary = `<div class="bg-white dark:bg-slate-800 rounded-xl shadow-lg p-4 mb-3">
+        <div class="text-sm text-gray-700 dark:text-gray-300">
+            Found <strong>${_pairs.length}</strong> duplicate pair${_pairs.length === 1 ? '' : 's'}.
+            Review the suggested merge direction (arrow points at the target project to keep),
+            adjust if needed, then click <strong>Merge</strong>.
+        </div>
+    </div>`;
+    results.innerHTML = summary + _pairs.map(_renderPair).join('');
+}
+
+function _renderPair(pair, i) {
+    const sourceTarget = pair.suggested_target_id === pair.a.id ? 'a' : 'b';
+    return `<div class="bg-white dark:bg-slate-800 rounded-xl shadow-lg p-4" data-idx="${i}">
+        <div class="flex items-center justify-between mb-3">
+            <div class="flex items-center gap-2">
+                <span class="px-2 py-0.5 rounded text-xs font-medium bg-amber-100 text-amber-800 dark:bg-amber-900/30 dark:text-amber-300">${(pair.score * 100).toFixed(0)}% match</span>
+                <span class="text-xs text-gray-500 dark:text-gray-400">${_esc(pair.reason)}</span>
+            </div>
+            <div class="flex items-center gap-2">
+                <button onclick="confirmMerge(${i}, 'a_into_b')"
+                        class="px-3 py-1.5 text-xs rounded ${sourceTarget === 'b' ? 'bg-seismo-orange hover:bg-seismo-navy text-white font-medium' : 'border border-gray-300 dark:border-gray-600 text-gray-700 dark:text-gray-300 hover:bg-gray-50 dark:hover:bg-gray-700'}">
+                    Merge A → B
+                </button>
+                <button onclick="confirmMerge(${i}, 'b_into_a')"
+                        class="px-3 py-1.5 text-xs rounded ${sourceTarget === 'a' ? 'bg-seismo-orange hover:bg-seismo-navy text-white font-medium' : 'border border-gray-300 dark:border-gray-600 text-gray-700 dark:text-gray-300 hover:bg-gray-50 dark:hover:bg-gray-700'}">
+                    Merge B → A
+                </button>
+                <button onclick="dismissPair(${i})"
+                        title="Hide this pair (not actually a duplicate)"
+                        class="px-3 py-1.5 text-xs border border-gray-300 dark:border-gray-600 text-gray-500 dark:text-gray-400 rounded hover:bg-gray-50 dark:hover:bg-gray-700">
+                    Not a dup
+                </button>
+            </div>
+        </div>
+        <div class="grid grid-cols-2 gap-3">
+            ${_renderProject(pair.a, 'A', sourceTarget === 'a')}
+            ${_renderProject(pair.b, 'B', sourceTarget === 'b')}
+        </div>
+    </div>`;
+}
+
+function _renderProject(p, label, isTarget) {
+    const borderCls = isTarget ? 'border-seismo-orange ring-1 ring-seismo-orange/30' : 'border-gray-200 dark:border-gray-700';
+    return `<a href="/projects/${_esc(p.id)}" target="_blank"
+              class="block bg-gray-50 dark:bg-slate-900/50 rounded-lg p-3 border ${borderCls} hover:shadow-md transition-shadow">
+        <div class="flex items-start justify-between gap-2 mb-1">
+            <div class="text-xs text-gray-500 dark:text-gray-400">Project ${label}${isTarget ? ' · suggested target' : ''}</div>
+            ${_sourceBadge(p.source)}
+        </div>
+        <div class="font-semibold text-gray-900 dark:text-white text-sm">${_esc(p.name)}</div>
+        ${p.project_number ? `<div class="text-xs text-gray-500 dark:text-gray-400 mt-0.5">#${_esc(p.project_number)}</div>` : ''}
+        ${p.client_name ? `<div class="text-xs text-gray-500 dark:text-gray-400 mt-0.5">${_esc(p.client_name)}</div>` : ''}
+        <div class="flex items-center gap-3 text-xs text-gray-600 dark:text-gray-400 mt-2">
+            <span><strong>${p.location_count}</strong> location${p.location_count === 1 ? '' : 's'}</span>
+            <span><strong>${p.assignment_count}</strong> assignment${p.assignment_count === 1 ? '' : 's'}</span>
+        </div>
+    </a>`;
+}
+
+async function confirmMerge(idx, direction) {
+    const pair = _pairs[idx];
+    if (!pair) return;
+    let sourceId, targetId, sourceName, targetName;
+    if (direction === 'a_into_b') {
+        sourceId = pair.a.id; targetId = pair.b.id;
+        sourceName = pair.a.name; targetName = pair.b.name;
+    } else {
+        sourceId = pair.b.id; targetId = pair.a.id;
+        sourceName = pair.b.name; targetName = pair.a.name;
+    }
+
+    // Pull preview to surface conflicts / consolidation count BEFORE merging.
+    let preview;
+    try {
+        const r = await fetch(`/api/projects/${sourceId}/merge_preview?target_id=${targetId}`);
+        if (!r.ok) {
+            const err = await r.json().catch(() => ({detail: 'HTTP ' + r.status}));
+            throw new Error(err.detail || ('HTTP ' + r.status));
+        }
+        preview = await r.json();
+    } catch (e) {
+        alert('Preview failed: ' + e.message);
+        return;
+    }
+
+    const summary = [
+        `${preview.total_assignments_moving} assignment(s)`,
+        `${preview.total_sessions_moving} session(s)`,
+        `${preview.total_data_files_moving} data file(s)`,
+    ].join(', ');
+    let consolidation = '';
+    const consolidates = preview.location_plans.filter(p => p.action === 'consolidate').length;
+    if (consolidates > 0) {
+        consolidation = `\n\n${consolidates} same-named location(s) will be consolidated.`;
+    }
+    const ok = confirm(
+        `Merge "${sourceName}" into "${targetName}"?\n\n` +
+        `Will move: ${summary}.${consolidation}\n\n` +
+        `Source will be soft-deleted.  This is reversible only via direct DB edit.`
+    );
+    if (!ok) return;
+
+    _showToast(`Merging "${sourceName}" → "${targetName}"…`);
+    try {
+        const r = await fetch(`/api/projects/${sourceId}/merge_into?target_id=${targetId}`, { method: 'POST' });
+        if (!r.ok) {
+            const err = await r.json().catch(() => ({detail: 'HTTP ' + r.status}));
+            throw new Error(err.detail || ('HTTP ' + r.status));
+        }
+        const d = await r.json();
+        _showToast(`Merged into "${targetName}"`,
+            `${d.assignments_moved} assignment(s), ${d.locations_moved + d.locations_consolidated} location(s)`,
+            'success');
+        _hideToast(3500);
+        // Re-scan: list updates without the merged pair.
+        await runScan();
+    } catch (e) {
+        _showToast('Merge failed', e.message, 'error');
+        _hideToast(5000);
+    }
+}
+
+function dismissPair(idx) {
+    // Just hide locally for now; doesn't persist across re-scans.
+    // A persistent "ignore pair" feature would need a new table; defer.
+    _pairs.splice(idx, 1);
+    render();
+}
+
+// Auto-scan on load with default threshold.
+document.addEventListener('DOMContentLoaded', () => {
+    runScan();
+});
+</script>
+{% endblock %}

templates/settings.html

@@ -574,6 +574,20 @@
                         Open
                     </a>
                 </div>
+
+                <!-- Project Tidy (Phase 5b) -->
+                <div class="flex items-center justify-between p-4 bg-gray-50 dark:bg-slate-700 rounded-lg">
+                    <div>
+                        <div class="font-medium text-gray-900 dark:text-white">Project Tidy</div>
+                        <div class="text-sm text-gray-500 dark:text-gray-400 mt-0.5">
+                            Find duplicate-looking projects via fuzzy name match (typos, abbreviations, spacing variations) and bulk-merge them.
+                        </div>
+                    </div>
+                    <a href="/settings/developer/project-tidy"
+                       class="ml-6 px-4 py-2 bg-seismo-orange hover:bg-orange-600 text-white text-sm font-medium rounded-lg transition-colors whitespace-nowrap">
+                        Open
+                    </a>
+                </div>
             </div>
         </div>
     </div>