Rebase seeded time-series data to now, so the demo stops going quiet

ShiftRecord is generated against now (shifts.go); everything else stayed on the
fixed calendar in seed.js while the calendar moved on. Twenty-three days after
that file was written the activity agent truthfully reported zero events in the
last seven days, and applications were sixteen days stale. Nothing was broken —
the data had simply aged out of every window the product reports over, and it
gets worse every day nobody reseeds.

RebaseToNow moves a set of records so the newest sits at now, keeping every gap
exactly as authored. The SHAPE is what every reader of this data is looking at:
three hires on one day, a screening the day after, a quiet fortnight before it.
Shifting the whole set by one delta preserves all of it. Scaling into a window
or scattering events across recent days would invent a rhythm nobody wrote.

Applied per entity — UserActivity, JobApplication, AIInterview — because each
anchors on its own newest record. One shared anchor would drag the quieter
entities by another entity's delta and invent relationships between them.
Reference data is untouched: a course's date is a fact about the course, not a
position in a window.

Rebasing rather than generating, so seed.json stays the single authored source,
still deterministic and still comparable byte-for-byte by the drift check. The
alternative — excluding these from the fixture the way ShiftRecord is — means a
second generator to keep in step with the frontend's copy.

The existing TestSeedPreservesSourceValues caught a real bug in the first
attempt, and its comment is why: "Applications carry updated_date in the
source, and the gap from created_date is what buildHires reads as
time-to-hire." I had shifted created_date alone, which turned five-day hires
into three-week ones. EVERY timestamp on a record now moves by the same delta,
and there is a test on that specifically.

That test now asserts the GAP rather than the absolute dates, because for a
rebased entity the dates are deliberately different — which is the one reason
it is supposed to allow. Its real subject was always the interval.

Verified against the database: before, activity was 23 days old with 0 events
in the last 7; after, all three entities are current, with 24 applications
spread across the last 30 days and 15 activity events inside 30.

Not fixed here: the Hiring activity chart on Control Center still renders
empty, and it was equally empty before this change. Its bucketing is correct —
replaying it in the browser against live data matched all 24 applications into
the right days — so the fault is further down in that component, not in the
data. Naming it rather than leaving it implied by a chart that still looks
wrong.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJvibeSc1JYXjatankqM1g

go-api/internal/seeder/activity.go

@@ -0,0 +1,122 @@
+package seeder
+
+import (
+	"strings"
+	"time"
+)
+
+// RebaseToNow moves a fixture's records so the most recent one sits at `now`,
+// keeping every gap between them exactly as authored.
+//
+// The problem this solves is that the demo goes quiet. ShiftRecord is generated
+// against now (see shifts.go); everything else stayed on the fixed calendar in
+// seed.js while the calendar moved on. Twenty-three days after that file was
+// written the activity agent truthfully reported zero events in the last seven
+// days, and the hiring chart on Control Center drew one bar in a thirty-day
+// window. Nothing was broken; the data had simply aged out of every window the
+// product asks about.
+//
+// Applied per entity rather than globally, because each is anchored on its own
+// newest record. Rebasing them together against one shared anchor would drag
+// the quieter entities forward or back by another entity's delta and invent
+// relationships between them that nobody authored.
+//
+// Rebasing rather than generating, deliberately. ShiftRecord is excluded from
+// seed.json because it is built fresh each run; doing the same here would mean
+// a second generator to keep in step with the frontend's copy, and a fixture
+// that no longer describes what a seeded workspace contains. Rebasing keeps
+// seed.json the single authored source — still deterministic, still comparable
+// byte-for-byte by the drift check — and moves the window instead of the data.
+//
+// The SHAPE is what matters to every reader of this data: three hires on one
+// day, a screening the day after, a quiet fortnight before it. Shifting the
+// whole set by one delta preserves all of that. Scaling it into a window, or
+// scattering events across recent days, would invent a rhythm nobody authored.
+//
+// EVERY timestamp on a record moves by the same delta, not just the anchor.
+// The gaps WITHIN a record are as load-bearing as the gaps between them: an
+// application's created_date and updated_date are what buildHires subtracts to
+// get time-to-hire, so shifting one and not the other invents a hire that took
+// three weeks or minus one. The seeder's own TestSeedPreservesSourceValues
+// caught exactly that, which is why it says what it says about the gap.
+//
+// Idempotent: the delta is recomputed from the fixture every run, so seeding
+// twice rebases the same authored dates twice rather than compounding.
+// Returns the records unchanged when there are none, or when none of them
+// carry a usable anchor.
+func RebaseToNow(records []map[string]any, field string, now time.Time) []map[string]any {
+	if len(records) == 0 {
+		return records
+	}
+
+	var newest time.Time
+	for _, rec := range records {
+		if at, ok := recordTime(rec, field); ok && at.After(newest) {
+			newest = at
+		}
+	}
+	if newest.IsZero() {
+		// Nothing parseable to anchor on. Leaving the fixture alone is the
+		// honest outcome: a wrong guess about these dates is worse than dates
+		// that are visibly old.
+		return records
+	}
+
+	delta := now.Sub(newest)
+	out := make([]map[string]any, 0, len(records))
+	for _, rec := range records {
+		if _, ok := recordTime(rec, field); !ok {
+			out = append(out, rec)
+			continue
+		}
+		// Copied, not mutated: the fixture is read once and seeded into
+		// possibly several organizations, and rewriting it in place would make
+		// the second one depend on the first.
+		shifted := make(map[string]any, len(rec))
+		for k, v := range rec {
+			shifted[k] = v
+			// Every timestamp moves together. A field that is not a timestamp,
+			// or a null one, is copied across untouched.
+			if at, isTime := recordTime(rec, k); isTime {
+				shifted[k] = at.Add(delta).UTC().Format(dateLayoutOf(rec[k]))
+			}
+		}
+		out = append(out, shifted)
+	}
+	return out
+}
+
+// recordTime reads one timestamp field, whatever shape the fixture used.
+func recordTime(rec map[string]any, field string) (time.Time, bool) {
+	raw, ok := rec[field]
+	if !ok {
+		return time.Time{}, false
+	}
+	switch v := raw.(type) {
+	case time.Time:
+		return v, true
+	case string:
+		for _, layout := range []string{time.RFC3339Nano, time.RFC3339, "2006-01-02"} {
+			if at, err := time.Parse(layout, v); err == nil {
+				return at, true
+			}
+		}
+	}
+	return time.Time{}, false
+}
+
+// dateLayoutOf keeps a rewritten timestamp in the shape the fixture used, so a
+// date-only field stays a date and a full timestamp keeps its precision.
+// Anything else the seeder reads back would differ from the source for reasons
+// that have nothing to do with the shift.
+func dateLayoutOf(raw any) string {
+	if s, ok := raw.(string); ok {
+		if len(s) == len("2006-01-02") {
+			return "2006-01-02"
+		}
+		if strings.Contains(s, ".") {
+			return "2006-01-02T15:04:05.000Z"
+		}
+	}
+	return time.RFC3339
+}

go-api/internal/seeder/activity_test.go

@@ -0,0 +1,173 @@
+package seeder_test
+
+import (
+	"testing"
+	"time"
+
+	"github.com/krow/krow-backend/go-api/internal/seeder"
+)
+
+func activityFixture() []map[string]any {
+	// The authored shape: a burst, then a gap, then a later cluster.
+	return []map[string]any{
+		{"id": "act_1", "event_type": "create_position", "created_date": "2026-07-18T09:00:00.000Z"},
+		{"id": "act_2", "event_type": "apply_job", "created_date": "2026-07-24T09:00:00.000Z"},
+		{"id": "act_3", "event_type": "hire_candidate", "created_date": "2026-07-25T09:00:00.000Z"},
+		{"id": "act_4", "event_type": "login", "created_date": "2026-08-06T09:00:00.000Z"},
+	}
+}
+
+func mustTime(t *testing.T, rec map[string]any) time.Time {
+	t.Helper()
+	at, err := time.Parse(time.RFC3339, rec["created_date"].(string))
+	if err != nil {
+		t.Fatalf("created_date %v does not parse: %v", rec["created_date"], err)
+	}
+	return at
+}
+
+// The newest event lands on now, so the demo always has something recent.
+func TestRebaseActivityAnchorsTheNewestEventToNow(t *testing.T) {
+	now := time.Date(2027, 3, 14, 12, 0, 0, 0, time.UTC)
+	out := seeder.RebaseToNow(activityFixture(), "created_date", now)
+
+	if len(out) != 4 {
+		t.Fatalf("got %d records, want 4", len(out))
+	}
+	newest := mustTime(t, out[3])
+	if !newest.Equal(now) {
+		t.Errorf("newest event at %s, want %s", newest, now)
+	}
+	for _, rec := range out {
+		if at := mustTime(t, rec); at.After(now) {
+			t.Errorf("%v is in the future at %s", rec["id"], at)
+		}
+	}
+}
+
+// Every gap is preserved. The shape of this data — three hires on one day, a
+// quiet fortnight before it — is what every reader of it is looking at.
+func TestRebaseActivityPreservesTheGaps(t *testing.T) {
+	in := activityFixture()
+	now := time.Date(2027, 3, 14, 12, 0, 0, 0, time.UTC)
+	out := seeder.RebaseToNow(in, "created_date", now)
+
+	for i := 1; i < len(in); i++ {
+		before := mustTime(t, in[i]).Sub(mustTime(t, in[i-1]))
+		after := mustTime(t, out[i]).Sub(mustTime(t, out[i-1]))
+		if before != after {
+			t.Errorf("gap %d changed from %s to %s", i, before, after)
+		}
+	}
+}
+
+// The whole point: events land inside the windows the activity tools ask about.
+func TestRebaseActivityLandsInsideTheReportingWindows(t *testing.T) {
+	now := time.Date(2027, 3, 14, 12, 0, 0, 0, time.UTC)
+	out := seeder.RebaseToNow(activityFixture(), "created_date", now)
+
+	var last7, last30 int
+	for _, rec := range out {
+		age := now.Sub(mustTime(t, rec))
+		if age < 7*24*time.Hour {
+			last7++
+		}
+		if age < 30*24*time.Hour {
+			last30++
+		}
+	}
+	if last7 == 0 {
+		t.Error("no events in the last 7 days — the demo still looks dead, " +
+			"which is the condition this function exists to prevent")
+	}
+	if last30 != 4 {
+		t.Errorf("%d of 4 events in the last 30 days, want all of them", last30)
+	}
+}
+
+// Seeding twice must not compound the shift.
+func TestRebaseActivityIsIdempotent(t *testing.T) {
+	now := time.Date(2027, 3, 14, 12, 0, 0, 0, time.UTC)
+	first := seeder.RebaseToNow(activityFixture(), "created_date", now)
+	second := seeder.RebaseToNow(activityFixture(), "created_date", now)
+
+	for i := range first {
+		if first[i]["created_date"] != second[i]["created_date"] {
+			t.Errorf("record %d differs between runs: %v vs %v",
+				i, first[i]["created_date"], second[i]["created_date"])
+		}
+	}
+}
+
+// The fixture is read once and may be seeded into several organizations, so
+// rebasing must not rewrite it in place.
+func TestRebaseActivityDoesNotMutateTheFixture(t *testing.T) {
+	in := activityFixture()
+	original := in[0]["created_date"]
+	seeder.RebaseToNow(in, "created_date", time.Date(2027, 3, 14, 12, 0, 0, 0, time.UTC))
+	if in[0]["created_date"] != original {
+		t.Errorf("the fixture was rewritten in place: %v became %v",
+			original, in[0]["created_date"])
+	}
+}
+
+func TestRebaseActivityHandlesNothingToDo(t *testing.T) {
+	now := time.Now()
+	if got := seeder.RebaseToNow(nil, "created_date", now); got != nil {
+		t.Errorf("nil in, %v out", got)
+	}
+	if got := seeder.RebaseToNow([]map[string]any{}, "created_date", now); len(got) != 0 {
+		t.Errorf("empty in, %d out", len(got))
+	}
+	// Unparseable dates are left alone rather than guessed at.
+	junk := []map[string]any{{"id": "x", "created_date": "not a date"}}
+	out := seeder.RebaseToNow(junk, "created_date", now)
+	if out[0]["created_date"] != "not a date" {
+		t.Errorf("an unreadable date was rewritten to %v", out[0]["created_date"])
+	}
+}
+
+// The gap WITHIN a record matters as much as the gaps between records.
+// buildHires subtracts an application's created_date from its updated_date to
+// get time-to-hire, so shifting one and not the other invents a hire that took
+// three weeks or minus one. The seeder's TestSeedPreservesSourceValues caught
+// this the first time; this pins it where the shifting happens.
+func TestRebaseToNowShiftsEveryTimestampTogether(t *testing.T) {
+	in := []map[string]any{
+		{
+			"id":           "app_1",
+			"created_date": "2026-07-20T09:00:00.000Z",
+			"updated_date": "2026-07-25T09:00:00.000Z", // five days later
+			"status":       "hired",
+			"ai_score":     91,
+		},
+		{
+			"id":           "app_2",
+			"created_date": "2026-08-06T09:00:00.000Z",
+			"updated_date": "2026-08-06T09:00:00.000Z",
+			"status":       "applied",
+		},
+	}
+	now := time.Date(2027, 3, 14, 12, 0, 0, 0, time.UTC)
+	out := seeder.RebaseToNow(in, "created_date", now)
+
+	created := mustTime(t, out[0])
+	updated, err := time.Parse(time.RFC3339, out[0]["updated_date"].(string))
+	if err != nil {
+		t.Fatalf("updated_date did not survive as a timestamp: %v", out[0]["updated_date"])
+	}
+	if gap := updated.Sub(created); gap != 5*24*time.Hour {
+		t.Errorf("time-to-hire became %s, want 120h — the two timestamps moved "+
+			"by different amounts", gap)
+	}
+
+	// Non-date fields are untouched.
+	if out[0]["status"] != "hired" || out[0]["ai_score"] != 91 {
+		t.Errorf("a non-date field was rewritten: %+v", out[0])
+	}
+	// And a record whose two dates were equal still has them equal.
+	if out[1]["created_date"] != out[1]["updated_date"] {
+		t.Errorf("equal timestamps diverged: %v vs %v",
+			out[1]["created_date"], out[1]["updated_date"])
+	}
+}

go-api/internal/seeder/seeder.go

@@ -127,6 +127,19 @@ func Load(path string) (*Fixture, error) {
 	return &f, nil
 }
 
+// rebasedEntities names the entities whose dates are moved to sit against now,
+// and the field to move them by.
+//
+// ShiftRecord is absent because it is GENERATED against now rather than
+// rebased (see shifts.go). Everything not listed here is reference data —
+// courses, role categories, badges — where a date is a fact about the record
+// rather than a position in a window, and moving it would be a lie.
+var rebasedEntities = map[string]string{
+	"UserActivity":   "created_date",
+	"JobApplication": "created_date",
+	"AIInterview":    "created_date",
+}
+
 // New builds a seeder. `now` anchors the generated shift records.
 func New(pool *pgxpool.Pool, fixture *Fixture, now time.Time) *Seeder {
 	return &Seeder{pool: pool, fixture: fixture, now: now}
@@ -158,6 +171,15 @@ func (s *Seeder) Run(ctx context.Context) (*Result, error) {
 		if entity == "ShiftRecord" {
 			records = BuildShifts(s.now)
 		}
+		// Time-series entities are authored on a fixed calendar and would
+		// otherwise age out of every window the product asks about — the
+		// activity tools, and the hiring chart on Control Center. See
+		// RebaseToNow. Each is anchored on its OWN newest record, so the gaps
+		// within an entity are preserved without inventing a relationship
+		// between entities.
+		if field, ok := rebasedEntities[entity]; ok {
+			records = RebaseToNow(records, field, s.now)
+		}
 		count, err := s.upsertEntity(ctx, tx, orgID, entity, records)
 		if err != nil {
 			return nil, fmt.Errorf("seed %s: %w", entity, err)

go-api/internal/seeder/seeder_test.go

@@ -3,6 +3,7 @@ package seeder_test
 import (
 	"context"
 	"encoding/json"
+	"fmt"
 	"os"
 	"path/filepath"
 	"strings"
@@ -153,11 +154,23 @@ func TestSeedPreservesSourceValues(t *testing.T) {
 		if status != want["status"] {
 			t.Errorf("%s status = %q, want %q", legacy, status, want["status"])
 		}
-		if created != want["created_date"] {
+		// Applications are REBASED (see RebaseToNow), so their absolute dates
-			t.Errorf("%s created_date = %q, want %q", legacy, created, want["created_date"])
+		// deliberately differ from the fixture — otherwise the demo ages out of
+		// every window the product reports over. What must survive is the GAP,
+		// because buildHires subtracts these two to get time-to-hire. Asserting
+		// the gap keeps this test's real subject and stops it failing for the
+		// one reason it is supposed to allow.
+		wantGap, err := fixtureGap(want["created_date"], want["updated_date"])
+		if err != nil {
+			t.Fatalf("%s: fixture dates: %v", legacy, err)
 		}
-		if updated != want["updated_date"] {
+		gotGap, err := fixtureGap(created, updated)
-			t.Errorf("%s updated_date = %q, want %q", legacy, updated, want["updated_date"])
+		if err != nil {
+			t.Fatalf("%s: seeded dates: %v", legacy, err)
+		}
+		if gotGap != wantGap {
+			t.Errorf("%s time-to-hire = %s, want %s (created %s, updated %s)",
+				legacy, gotGap, wantGap, created, updated)
 		}
 		if v, ok := want["ai_score"].(float64); ok && score != int(v) {
 			t.Errorf("%s ai_score = %d, want %d", legacy, score, int(v))
@@ -414,3 +427,17 @@ func TestFixtureIsGeneratedNotHandWritten(t *testing.T) {
 		t.Errorf("`_generated` does not name its source: %q", head.Generated)
 	}
 }
+
+// fixtureGap is the interval between two fixture timestamps.
+func fixtureGap(from, to any) (time.Duration, error) {
+	const layout = "2006-01-02T15:04:05.000Z"
+	a, err := time.Parse(layout, fmt.Sprint(from))
+	if err != nil {
+		return 0, fmt.Errorf("parse %v: %w", from, err)
+	}
+	b, err := time.Parse(layout, fmt.Sprint(to))
+	if err != nil {
+		return 0, fmt.Errorf("parse %v: %w", to, err)
+	}
+	return b.Sub(a), nil
+}