Surface hidden counts for TopFiles and Collaborators in DevProfile

internal/report/profile_template.go

@@ -127,6 +127,9 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
       <span style="background:#0969da; color:#fff; border-radius:8px; padding:0 6px; font-size:10px;" title="{{.SharedFiles}} files / {{.SharedLines}} lines">{{.SharedFiles}} files · {{humanize .SharedLines}} lines</span>
     </span>
     {{end}}
+    {{if gt .Profile.CollaboratorsHidden 0}}
+    <span style="display:inline-flex; align-items:center; padding:3px 10px; font-size:11px; color:#656d76; font-style:italic;">+{{.Profile.CollaboratorsHidden}} more collaborators not shown</span>
+    {{end}}
   </div>
 </div>
 {{end}}
@@ -145,6 +148,9 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
   <td style="width:30%"><div style="display:flex;"><div class="bar bar-churn" style="width:{{pct .Churn $maxChurn}}%"></div></div></td>
 </tr>
 {{end}}
+{{if gt .Profile.TopFilesHidden 0}}
+<tr><td colspan="4" style="color:#656d76; font-style:italic; text-align:center;">+{{.Profile.TopFilesHidden}} more files not shown</td></tr>
+{{end}}
 </table>
 {{end}}
 

internal/report/report.go

@@ -54,6 +54,20 @@ type ReportData struct {
 	// so mature repos (linux-scale) don't blow up the HTML. nil when
 	// the dataset has no files.
 	Structure *TreeNode
+
+	// TotalDirectories / TotalExtensions / TotalBusFactorFiles are
+	// the full universe sizes (before top-N truncation) for sections
+	// whose denominators aren't in Summary. Templates render "20 of
+	// 127" headers so the reader sees the scale of what's been
+	// truncated. TotalBusFactorFiles specifically excludes files
+	// with empty devLines (pure-rename-only files post-ingest-fix)
+	// because BusFactor skips those — using Summary.TotalFiles here
+	// would make the header lie on rename-heavy repos. Summary
+	// already carries TotalDevs / TotalFiles / TotalCommits for the
+	// rest.
+	TotalDirectories    int
+	TotalExtensions     int
+	TotalBusFactorFiles int
 }
 
 // htmlTreeDepth caps the repo-structure tree baked into the HTML report.
@@ -359,6 +373,9 @@ func Generate(w io.Writer, ds *stats.Dataset, repoName string, topN int, sf stat
 		PatternGrid:          grid,
 		MaxPattern:           maxP,
 		Structure:            BuildRepoTree(stats.FileHotspots(ds, 0), htmlTreeDepth),
+		TotalDirectories:     stats.DirectoryCount(ds),
+		TotalExtensions:      stats.ExtensionCount(ds),
+		TotalBusFactorFiles:  stats.BusFactorCount(ds),
 	}
 	CapChildrenPerDir(data.Structure, htmlTreeMaxChildrenPerDir)
 

internal/report/template.go

@@ -185,7 +185,7 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
 {{end}}
 
 {{if .Contributors}}
-<h2>Top Contributors</h2>
+<h2>Top Contributors{{if lt (len .Contributors) .Summary.TotalDevs}} <span style="font-size:13px; color:#656d76; font-weight:normal;">{{thousands (len .Contributors)}} of {{thousands .Summary.TotalDevs}}</span>{{end}}</h2>
 <p class="hint">Ranked by commit count. High commit count with low lines may indicate small fixes; low count with high lines may indicate large features. · {{docRef "contributors"}}</p>
 <table>
 <tr><th>Name</th><th>Email</th><th>Commits</th><th></th><th>Additions</th><th>Deletions</th></tr>
@@ -204,7 +204,7 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
 {{end}}
 
 {{if .Hotspots}}
-<h2>File Hotspots</h2>
+<h2>File Hotspots{{if lt (len .Hotspots) .Summary.TotalFiles}} <span style="font-size:13px; color:#656d76; font-weight:normal;">{{thousands (len .Hotspots)}} of {{thousands .Summary.TotalFiles}}</span>{{end}}</h2>
 <p class="hint">Most frequently changed files. High churn with few devs = knowledge silo. High churn with many devs = shared bottleneck. · {{docRef "hotspots"}}</p>
 <table>
 <tr><th>Path</th><th>Commits</th><th>Churn</th><th></th><th>Devs</th></tr>
@@ -222,7 +222,7 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
 {{end}}
 
 {{if .Directories}}
-<h2>Directories</h2>
+<h2>Directories{{if lt (len .Directories) .TotalDirectories}} <span style="font-size:13px; color:#656d76; font-weight:normal;">{{thousands (len .Directories)}} of {{thousands .TotalDirectories}}</span>{{end}}</h2>
 <p class="hint">Module-level health. <b>File touches</b> is the sum of per-file commit counts (one commit touching N files contributes N), not distinct commits. Low bus factor = knowledge concentrated in few people. · {{docRef "directories"}}</p>
 <table>
 <tr><th>Directory</th><th>File Touches</th><th>Churn</th><th>Files</th><th>Devs</th><th>Bus Factor</th></tr>
@@ -240,7 +240,7 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
 {{end}}
 
 {{if .Extensions}}
-<h2>Extensions</h2>
+<h2>Extensions{{if lt (len .Extensions) .TotalExtensions}} <span style="font-size:13px; color:#656d76; font-weight:normal;">{{thousands (len .Extensions)}} of {{thousands .TotalExtensions}}</span>{{end}}</h2>
 <p class="hint">File extensions ranked by <b>recent churn</b> — "where is the team spending effort now", not "what exists at HEAD". Cross-read with Directories: a repo with high <code>.yaml</code> recent churn concentrated in one dir is config-as-code; spread across many dirs is config sprawl. · {{docRef "extensions"}}</p>
 <table>
 <tr><th>Ext</th><th>Files</th><th>Churn</th><th>Recent Churn</th><th></th><th>Devs</th><th>First Seen</th><th>Last Seen</th></tr>
@@ -261,7 +261,7 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
 {{end}}
 
 {{if .ChurnRisk}}
-<h2>Churn Risk</h2>
+<h2>Churn Risk{{if lt (len .ChurnRisk) .Summary.TotalFiles}} <span style="font-size:13px; color:#656d76; font-weight:normal;">{{thousands (len .ChurnRisk)}} of {{thousands .Summary.TotalFiles}}</span>{{end}}</h2>
 <p class="hint">Files ranked by recent churn. Label classifies context so you can judge action: <b>legacy-hotspot</b> (old code + concentrated + declining) is the urgent alarm; <b>silo</b> suggests knowledge transfer; <b>active-core</b> is young code with a single author (often fine); <b>active</b> is shared healthy work; <b>cold</b> is quiet.{{if (index .ChurnRisk 0).AgePercentile}} <b>Age P__ / Trend P__</b> under the label show where this file sits in the repo's distribution: age P90 means older than 90% of tracked files; trend P10 means declining more sharply than 90%. Classification boundaries are the P75 age and P25 trend of this dataset (see {{docRef "churn-risk"}}).{{end}}</p>
 {{if .ChurnRiskLabelCounts}}
 <div class="churn-chips">
@@ -290,7 +290,7 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
 {{end}}
 
 {{if .BusFactor}}
-<h2>Bus Factor Risk</h2>
+<h2>Bus Factor Risk{{if lt (len .BusFactor) .TotalBusFactorFiles}} <span style="font-size:13px; color:#656d76; font-weight:normal;">{{thousands (len .BusFactor)}} of {{thousands .TotalBusFactorFiles}}</span>{{end}}</h2>
 <p class="hint">Files with fewest developers owning 80%+ of changes. Bus factor 1 = if that person leaves, nobody else knows the code. · {{docRef "bus-factor"}}</p>
 <table>
 <tr><th>Path</th><th>Bus Factor</th><th>Top Devs</th></tr>
@@ -337,7 +337,7 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
 {{end}}
 
 {{if .TopCommits}}
-<h2>Top Commits</h2>
+<h2>Top Commits{{if lt (len .TopCommits) .Summary.TotalCommits}} <span style="font-size:13px; color:#656d76; font-weight:normal;">{{thousands (len .TopCommits)}} of {{thousands .Summary.TotalCommits}}</span>{{end}}</h2>
 <p class="hint">Largest commits by lines changed. Unusually large commits may be imports, generated code, or risky big-bang changes worth reviewing. · {{docRef "top-commits"}}</p>
 <table>
 <tr><th>SHA</th><th>Author</th><th>Date</th><th>Lines</th><th>Files</th>{{if and (gt (len .TopCommits) 0) (index .TopCommits 0).Message}}<th>Message</th>{{end}}</tr>
@@ -398,7 +398,7 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
     <span>{{printf "%.1f" .Pace}} commits/active day</span>
 
     <span style="color:#656d76;">Collaboration</span>
-    <span>{{if .Collaborators}}{{range $i, $c := .Collaborators}}{{if $i}}, {{end}}{{$c.Email}} ({{thousands $c.SharedFiles}} files, {{thousands $c.SharedLines}} lines){{end}}{{else}}solo contributor{{end}}</span>
+    <span>{{if .Collaborators}}{{range $i, $c := .Collaborators}}{{if $i}}, {{end}}{{$c.Email}} ({{thousands $c.SharedFiles}} files, {{thousands $c.SharedLines}} lines){{end}}{{if gt .CollaboratorsHidden 0}} <span style="color:#656d76; font-style:italic;">(+{{.CollaboratorsHidden}} more)</span>{{end}}{{else}}solo contributor{{end}}</span>
 
     <span style="color:#656d76;">Weekend</span>
     <span>{{printf "%.1f" .WeekendPct}}%</span>
@@ -414,6 +414,9 @@ footer { margin-top: 40px; padding-top: 16px; border-top: 1px solid #d0d7de; col
       <span style="color:#656d76;">{{thousands .Churn}} churn</span>
     </div>
     {{end}}
+    {{if gt .TopFilesHidden 0}}
+    <div style="color:#656d76; font-style:italic; margin-top:2px;">+{{.TopFilesHidden}} more files not shown</div>
+    {{end}}
   </div>
   {{end}}
 </div>

internal/stats/extension_test.go

@@ -1,6 +1,7 @@
 package stats
 
 import (
+	"fmt"
 	"strings"
 	"testing"
 	"time"
@@ -56,6 +57,62 @@ func TestExtractExtensionPolicy(t *testing.T) {
 	}
 }
 
+// DirectoryCount and ExtensionCount back the "N of M" header badges
+// in the main report. They must match what DirectoryStats and
+// ExtensionStats would produce pre-truncation — otherwise "showing 20
+// of 127" lies when the user expands to --top 0 and finds a
+// different number.
+func TestDirectoryCountAndExtensionCount(t *testing.T) {
+	ds := &Dataset{
+		files: map[string]*fileEntry{
+			"cmd/main.go":   {devLines: map[string]int64{"a": 1}},
+			"cmd/util.go":   {devLines: map[string]int64{"a": 1}},
+			"internal/a.go": {devLines: map[string]int64{"a": 1}},
+			"internal/b.go": {devLines: map[string]int64{"a": 1}},
+			"docs/x.md":     {devLines: map[string]int64{"a": 1}},
+			"README.md":     {devLines: map[string]int64{"a": 1}}, // "." bucket
+			"Makefile":      {devLines: map[string]int64{"a": 1}}, // "." bucket, (none) ext
+		},
+	}
+	// Distinct dirs: "cmd", "internal", "docs", "." → 4
+	if n := DirectoryCount(ds); n != 4 {
+		t.Errorf("DirectoryCount = %d, want 4", n)
+	}
+	// Distinct exts: ".go", ".md", "(none)" → 3
+	if n := ExtensionCount(ds); n != 3 {
+		t.Errorf("ExtensionCount = %d, want 3", n)
+	}
+	// Consistency invariant: must match len of the stats function's
+	// full output. If they ever drift, the "N of M" header lies.
+	if got, want := DirectoryCount(ds), len(DirectoryStats(ds, 0)); got != want {
+		t.Errorf("DirectoryCount (%d) != len(DirectoryStats(_, 0)) (%d)", got, want)
+	}
+	if got, want := ExtensionCount(ds), len(ExtensionStats(ds, 0)); got != want {
+		t.Errorf("ExtensionCount (%d) != len(ExtensionStats(_, 0)) (%d)", got, want)
+	}
+}
+
+// BusFactorCount must exclude pure-rename files (empty devLines) —
+// those are skipped by BusFactor itself and so cannot be part of its
+// denominator. Using Summary.TotalFiles would over-count here. Build
+// a dataset with one file that carries dev lines and one that
+// doesn't; assert BusFactorCount == 1 and matches the real output.
+func TestBusFactorCountExcludesEmptyDevLines(t *testing.T) {
+	ds := &Dataset{
+		UniqueFileCount: 2, // Summary total; includes both
+		files: map[string]*fileEntry{
+			"src/authored.go":       {devLines: map[string]int64{"alice@x": 10}},
+			"src/pure-rename-only":  {devLines: map[string]int64{}}, // no authored lines
+		},
+	}
+	if got := BusFactorCount(ds); got != 1 {
+		t.Errorf("BusFactorCount = %d, want 1 (pure-rename file must be excluded)", got)
+	}
+	if got, want := BusFactorCount(ds), len(BusFactor(ds, 0)); got != want {
+		t.Errorf("BusFactorCount (%d) != len(BusFactor(_, 0)) (%d) — header would lie", got, want)
+	}
+}
+
 func TestExtensionStatsAggregation(t *testing.T) {
 	// Hand-built dataset so aggregation is inspectable: two .go files
 	// with distinct devs, one .yaml shared by both, and a Makefile
@@ -667,6 +724,62 @@ func TestDevProfileHiddenCounters(t *testing.T) {
 	}
 }
 
+// Completes the Hidden-counter family: TopFiles truncates at 10,
+// Collaborators at 5. Both used to drop buckets silently — the "+N
+// more" surfaced for Scope/Extensions had no counterpart here, so a
+// dev with 25 touched files or 12 frequent collaborators looked like
+// they had exactly 10 / 5. Build a dev with 12 files and 7
+// collaborators; assert the counters report the true drop count.
+func TestDevProfileHiddenCountersTopFilesAndCollaborators(t *testing.T) {
+	files := map[string]*fileEntry{}
+	// 12 files authored by alice → TopFilesHidden = 2.
+	for i := 0; i < 12; i++ {
+		path := fmt.Sprintf("dir%d/file%d.go", i, i)
+		files[path] = &fileEntry{
+			devLines:   map[string]int64{"alice@x": int64(100 - i*5)},
+			devCommits: map[string]int{"alice@x": 1},
+		}
+	}
+	// Seed 6 shared files between alice and each of 6 other devs →
+	// alice has 6 collaborators total; top-5 truncation gives
+	// CollaboratorsHidden = 1.
+	for i := 0; i < 6; i++ {
+		path := fmt.Sprintf("shared/collab%d.go", i)
+		collab := fmt.Sprintf("bob%d@x", i)
+		files[path] = &fileEntry{
+			devLines:   map[string]int64{"alice@x": 50, collab: 50},
+			devCommits: map[string]int{"alice@x": 1, collab: 1},
+		}
+	}
+	contribs := map[string]*ContributorStat{
+		"alice@x": {Email: "alice@x", Commits: 18, FilesTouched: 18, ActiveDays: 1},
+	}
+	for i := 0; i < 6; i++ {
+		contribs[fmt.Sprintf("bob%d@x", i)] = &ContributorStat{
+			Email: fmt.Sprintf("bob%d@x", i), Commits: 1, FilesTouched: 1, ActiveDays: 1,
+		}
+	}
+	ds := &Dataset{
+		contributors: contribs,
+		files:        files,
+		commits:      map[string]*commitEntry{},
+		workGrid:     [7][24]int{},
+	}
+	p := DevProfiles(ds, "alice@x", 0)[0]
+	if len(p.TopFiles) != 10 {
+		t.Fatalf("TopFiles len = %d, want 10 (truncated from 18)", len(p.TopFiles))
+	}
+	if p.TopFilesHidden != 8 {
+		t.Errorf("TopFilesHidden = %d, want 8 (18 - 10)", p.TopFilesHidden)
+	}
+	if len(p.Collaborators) != 5 {
+		t.Fatalf("Collaborators len = %d, want 5 (truncated from 6)", len(p.Collaborators))
+	}
+	if p.CollaboratorsHidden != 1 {
+		t.Errorf("CollaboratorsHidden = %d, want 1 (6 - 5)", p.CollaboratorsHidden)
+	}
+}
+
 // Silent when nothing to hide — the counters must be zero so the
 // renderers don't emit "+0 more" (noise) for the common case.
 func TestDevProfileHiddenCountersZeroWhenFits(t *testing.T) {
@@ -687,6 +800,12 @@ func TestDevProfileHiddenCountersZeroWhenFits(t *testing.T) {
 		t.Errorf("Hidden counters: Scope=%d Ext=%d, want 0/0 (dev has ≤5 buckets each)",
 			p.ScopeHidden, p.ExtensionsHidden)
 	}
+	// TopFiles cap is 10, Collaborators cap is 5 — bob has 3 files
+	// and zero collaborators, so both must stay at 0.
+	if p.TopFilesHidden != 0 || p.CollaboratorsHidden != 0 {
+		t.Errorf("Hidden counters: TopFiles=%d Collab=%d, want 0/0",
+			p.TopFilesHidden, p.CollaboratorsHidden)
+	}
 }
 
 // Truncate to top-5 when a dev's extension set is larger. Under the

internal/stats/format.go

@@ -500,6 +500,9 @@ func (f *Formatter) PrintProfiles(profiles []DevProfile) error {
 					}
 					fmt.Fprintf(f.w, "%s (%d shared)", c.Email, c.SharedFiles)
 				}
+				if p.CollaboratorsHidden > 0 {
+					fmt.Fprintf(f.w, "  (+%d more)", p.CollaboratorsHidden)
+				}
 			} else {
 				fmt.Fprintf(f.w, "solo contributor")
 			}
@@ -512,6 +515,9 @@ func (f *Formatter) PrintProfiles(profiles []DevProfile) error {
 				for _, tf := range p.TopFiles {
 					fmt.Fprintf(f.w, "    %-50s %3d commits  %6d churn\n", tf.Path, tf.Commits, tf.Churn)
 				}
+				if p.TopFilesHidden > 0 {
+					fmt.Fprintf(f.w, "    ... (+%d more files not shown)\n", p.TopFilesHidden)
+				}
 			}
 
 			if len(p.MonthlyActivity) > 0 {