Phase 5: iteration & polish (visualizations, goal history, smart tagging)

Final roadmap phase. No DB migration — it reads data already captured. - Goal-evolution history: GET /discovery/profile/me/goal-history derives a per-goal timeline from the profile_revision snapshots (pure aggregator in app/services/profile_history.py). - Smart tagging: POST /discovery/integration/suggest-foundation suggests which foundation a task builds toward + rationale/confidence (FoundationTagger, app/services/tagging.py). Suggestion only; the person confirms by posting the task mapping. - Deeper goal-refinement: the reflect loop accepts an optional focus ("goals") that steers the coach toward sharpening goals — still mirror, not compass. - Visualizations: visuals.html renders an Ikigai Venn and an Enneagram diagram (plain-language callouts, not the raw type number) plus the goal-evolution timeline; linked from profile.html. Tests: 99 passing (added pure goal-history tests, goal-history + suggest endpoint tests, reflect-focus passthrough; run in-container). README updated. This completes the ImpactFlow Vision roadmap (Phases 1-5) on the Discovery side. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-08-27 05:10:36 +00:00 · 2026-06-16 21:37:33 -05:00
parent c4fc1cccd7
commit b9d7b0e22b
14 changed files with 798 additions and 9 deletions
@@ -50,6 +50,10 @@ If you need to explain this app in detail, use this mental model:
    profile foundation via `POST /discovery/integration/task-mappings`. The
    rolled-up work patterns drive `/static/dashboard.html` and are fed into the
    coaching check-ins so they can reflect where time has actually gone.
+15. (Phase 5) `/static/visuals.html` shows the Ikigai/Enneagram visuals and the
+    goal-evolution timeline; the tracker can call `suggest-foundation` for
+    smart tagging, and the reflect loop accepts a `focus` for deeper
+    goal-refinement.

 Machine-to-machine callers (e.g. the MCP server) skip the OAuth dance and
 authenticate with `X-API-Key: $IMPACTFLOW_API_KEY` instead. That header
@@ -152,13 +156,16 @@ Important files:
 | `app/static/reflect.html` | Phase 2 AI-coach reflection chat (mirror loop, applies revisions, affirm) |
 | `app/static/coaching.html` | Phase 3 coaching preferences form + check-in feed |
 | `app/static/dashboard.html` | Phase 4 goal dashboard: time per foundation + neglected ones |
+| `app/static/visuals.html` | Phase 5 visualizations: Ikigai Venn, Enneagram diagram, goal-evolution timeline |
 | `app/static/auth.js` | Shared `authedFetch` helper: sends session cookies, silently refreshes on `401`, redirects to login |
 | `app/static/style.css` | Shared UI styling |
 | `app/services/reflector.py` | `ReflectionCoach`: Anthropic-backed mirror loop, JSON parsing, revision filtering |
 | `app/services/coaching.py` | Deterministic preference generator + `CheckinCoach` (Anthropic check-in text) |
 | `app/services/foundations.py` | The six foundations + pure work-pattern aggregator (Phase 4) |
+| `app/services/profile_history.py` | Pure goal-evolution aggregator over profile revisions (Phase 5) |
+| `app/services/tagging.py` | `FoundationTagger`: Anthropic smart-tagging suggestion (Phase 5) |
 | `app/routers/coaching.py` | Phase 3 coaching routes: preferences, check-ins, weekly batch `/run` |
-| `app/routers/integration.py` | Phase 4 task-to-goal integration: foundations, task-mappings, work-patterns |
+| `app/routers/integration.py` | Phase 4/5 integration: foundations, task-mappings, work-patterns, suggest-foundation |
 | `alembic/versions/001_initial.py` | Initial database schema migration |
 | `alembic/versions/002_add_auth.py` | Adds `users`, `refresh_tokens`, and `activity_log` tables |
 | `alembic/versions/003_add_goals.py` | Adds the goal columns to `discovery_conversation` and `discovery_profile` |
@@ -172,6 +179,8 @@ Important files:
 | `tests/test_coaching.py` | Tests for coaching endpoints (preferences, check-ins, due-logic batch, work-pattern wiring) |
 | `tests/test_foundations.py` | Unit tests for the pure work-pattern aggregator |
 | `tests/test_integration.py` | Tests for the task-to-goal integration endpoints |
+| `tests/test_profile_history.py` | Unit tests for the pure goal-evolution aggregator |
+| `tests/test_phase5.py` | Tests for goal-history and smart-tagging endpoints |
 | `tests/test_auth.py` | Tests for the dual-auth dependency (JWT + cookie + API key), token refresh/logout, admin enforcement, and domain allow-list |
 | `tests/test_profile_edit.py` | Tests for `PATCH /discovery/profile/me` (edit, partial update, lock/`409`) |
 | `tests/test_reflection.py` | Tests for the reflection endpoints (turns, applied revisions, lock/`409`, history) |
@@ -461,6 +470,23 @@ the user (forwarded session/JWT) or service-to-service with `X-API-Key`.
  time has and hasn't gone — as an observation to check against the person's
  own words, never a verdict (mirror, not compass).

+### 11. Iteration & Polish (Phase 5)
+
+- **Goal-evolution history:** `GET /discovery/profile/me/goal-history` derives a
+  per-goal timeline (near- and long-term) from the `profile_revision` snapshots
+  captured since Phase 2 — one entry per actual change, with `source` and `at`.
+  No new storage; it reads existing revisions.
+- **Smart tagging:** `POST /discovery/integration/suggest-foundation`
+  (`{task_label}`) returns a suggested `foundation` + `rationale` + `confidence`.
+  It only suggests — the person confirms by posting the task mapping. The core
+  tracker uses this to pre-fill "which goal does this build toward?".
+- **Deeper goal-refinement:** the reflect loop (Phase 2) accepts an optional
+  `focus` (e.g. `"goals"`) that steers the coach toward sharpening the
+  near/long-term goals — still a mirror.
+- **Visualizations:** `visuals.html` renders an Ikigai four-circle Venn and an
+  Enneagram diagram from the profile (plain-language callouts — it does not
+  headline the raw type number), plus the goal-evolution timeline.
+
 ## API Reference

 All `/discovery/*`, `/api/me*`, `/api/activity*`, and `/api/admin/*` routes
@@ -507,6 +533,8 @@ clients.) `/api/auth/login`, `/api/auth/callback`, `/health`, `/`, and
 | `POST` | `/discovery/integration/task-mappings` | yes | Record a logged unit of work mapped to a foundation |
 | `GET` | `/discovery/integration/task-mappings` | yes | List the user's task mappings in a window |
 | `GET` | `/discovery/integration/work-patterns` | yes | Per-foundation work-pattern rollup over a window |
+| `POST` | `/discovery/integration/suggest-foundation` | yes | Smart tagging: suggest a foundation for a task (person confirms) |
+| `GET` | `/discovery/profile/me/goal-history` | yes | Timeline of how the person's goals evolved |

 ## Data Model

@@ -780,6 +808,14 @@ callback), so the pages hold no tokens of their own.
  the person wants their energy — it observes, it does not prescribe
 - linked from `profile.html` ("Where your time goes")

+`visuals.html` (Phase 5 visualizations):
+
+- reads `GET /discovery/profile/me` and renders an Ikigai four-circle Venn and
+  an Enneagram diagram in SVG (callouts stay in plain language)
+- reads `GET /discovery/profile/me/goal-history` and shows how the near- and
+  long-term goals have changed over time
+- linked from `profile.html` ("See your profile visualized")
+
 ## Configuration

 Populate `.env` with at minimum the Anthropic key and the auth-related
@@ -30,6 +30,7 @@ from app.services.reflector import (
    ReflectionCoach,
    ReflectionError,
 )
+from app.services.profile_history import goal_history

 router = APIRouter(prefix="/discovery", tags=["discovery"])

@@ -383,7 +384,11 @@ async def reflect_on_profile(
    model = os.getenv("ANTHROPIC_MODEL", "claude-sonnet-4-6")
    try:
        coach = ReflectionCoach(api_key=api_key, model=model)
-        result = await coach.reflect(_profile_fields(profile) | {"triad": profile.triad}, coach_history)
+        result = await coach.reflect(
+            _profile_fields(profile) | {"triad": profile.triad},
+            coach_history,
+            focus=payload.focus,
+        )
    except ReflectionError as exc:
        raise HTTPException(status_code=502, detail=str(exc)) from exc

@@ -459,6 +464,44 @@ async def get_profile_revisions(
    return out


+@router.get(
+    "/profile/me/goal-history", response_model=schemas.GoalHistoryOut
+)
+async def get_goal_history(
+    db: AsyncSession = Depends(get_db),
+    user: User = Depends(get_current_user),
+):
+    """How the person's near/long-term goals evolved over time, derived from
+    the profile revision snapshots (Phase 5)."""
+    profile = await _latest_profile(db, user.id)
+    if profile is None:
+        raise HTTPException(status_code=404, detail="No profile for this user")
+    rev_stmt = (
+        select(ProfileRevision)
+        .where(ProfileRevision.profile_id == profile.id)
+        .order_by(ProfileRevision.created_at)
+    )
+    rows = (await db.execute(rev_stmt)).scalars().all()
+    revisions = []
+    for r in rows:
+        try:
+            fields = json.loads(r.fields_json)
+        except (json.JSONDecodeError, TypeError):
+            fields = {}
+        revisions.append(
+            {"fields": fields, "source": r.source, "created_at": r.created_at}
+        )
+    timelines = goal_history(revisions)
+    return schemas.GoalHistoryOut(
+        short_term_goals=[
+            schemas.GoalHistoryEntry(**e) for e in timelines["short_term_goals"]
+        ],
+        long_term_goals=[
+            schemas.GoalHistoryEntry(**e) for e in timelines["long_term_goals"]
+        ],
+    )
+
+
@router.get(
    "/conversation/{conversation_id}",
    response_model=schemas.ConversationResponse,
@@ -9,6 +9,7 @@ up to feed the coaching reminder engine and the goal dashboard.
 All routes are user-scoped via the dual-auth dependency; the tracker calls as
 the user (forwarded session/JWT) or, service-to-service, with ``X-API-Key``.
 """
+import os
 import uuid
 from datetime import datetime, timedelta, timezone

@@ -25,6 +26,7 @@ from app.services.foundations import (
    FOUNDATIONS,
    rollup,
 )
+from app.services.tagging import FoundationTagger, TaggingError

 router = APIRouter(prefix="/discovery/integration", tags=["integration"])

@@ -141,6 +143,44 @@ async def work_patterns_for(
    return rollup(entries)


+@router.post(
+    "/suggest-foundation", response_model=schemas.SuggestFoundationOut
+)
+async def suggest_foundation(
+    payload: schemas.SuggestFoundationRequest,
+    db: AsyncSession = Depends(get_db),
+    user: User = Depends(get_current_user),
+):
+    """Phase 5 smart tagging: suggest which foundation a task builds toward.
+    The person confirms by posting the task mapping. Suggestion only — nothing
+    is stored here."""
+    label = payload.task_label.strip()
+    if not label:
+        raise HTTPException(status_code=400, detail="task_label is required")
+    profile = await _latest_profile(db, user.id)
+    if profile is None:
+        raise HTTPException(status_code=404, detail="No profile for this user")
+
+    foundations_text = {
+        key: getattr(profile, field, None)
+        for key, field in FOUNDATION_TO_FIELD.items()
+    }
+    api_key = os.getenv("ANTHROPIC_API_KEY")
+    model = os.getenv("ANTHROPIC_MODEL", "claude-sonnet-4-6")
+    try:
+        tagger = FoundationTagger(api_key=api_key, model=model)
+        result = await tagger.suggest(label, foundations_text)
+    except TaggingError as exc:
+        raise HTTPException(status_code=502, detail=str(exc)) from exc
+
+    return schemas.SuggestFoundationOut(
+        foundation=result["foundation"],
+        label=FOUNDATIONS[result["foundation"]],
+        rationale=result["rationale"],
+        confidence=result["confidence"],
+    )
+
+
@router.get("/work-patterns", response_model=schemas.WorkPatternsOut)
 async def get_work_patterns(
    days: int = Query(30, ge=1, le=365),
@@ -77,6 +77,9 @@ class ConfirmResponse(BaseModel):
 class ReflectRequest(BaseModel):
    # Empty/omitted starts the loop (the coach's opening reflection).
    message: str = ""
+    # Phase 5: optional refinement focus, e.g. "goals" to steer the coach
+    # toward sharpening the near/long-term goals.
+    focus: str = ""


 class ReflectionMessageOut(BaseModel):
@@ -194,6 +197,31 @@ class WorkPatternsOut(BaseModel):
    neglected: list[str]


+# -- Phase 5: iteration & polish ---------------------------------------------
+
+
+class GoalHistoryEntry(BaseModel):
+    value: Optional[str] = None
+    source: Optional[str] = None
+    at: Optional[datetime] = None
+
+
+class GoalHistoryOut(BaseModel):
+    short_term_goals: list[GoalHistoryEntry]
+    long_term_goals: list[GoalHistoryEntry]
+
+
+class SuggestFoundationRequest(BaseModel):
+    task_label: str
+
+
+class SuggestFoundationOut(BaseModel):
+    foundation: str
+    label: str
+    rationale: str = ""
+    confidence: str = "low"
+
+
 class ConversationResponse(BaseModel):
    id: str
    user_id: str
@@ -0,0 +1,50 @@
+"""Phase 5: derive how a person's goals evolved from the profile_revision
+snapshots already captured since Phase 2.
+
+Pure functions over revision rows — no DB access, fully testable.
+"""
+from typing import Any, Dict, List
+
+# The goal fields we track an evolution timeline for.
+GOAL_FIELDS = ("short_term_goals", "long_term_goals")
+
+
+def goal_history(revisions: List[Dict[str, Any]]) -> Dict[str, List[dict]]:
+    """Build a per-goal timeline of distinct values over time.
+
+    Args:
+        revisions: each ``{"fields": {<field>: value, ...}, "source": str,
+            "created_at": datetime}``, in ascending chronological order.
+
+    Returns:
+        ``{"short_term_goals": [{"value", "source", "at"}...],
+           "long_term_goals": [...]}`` — one entry per *change*; consecutive
+        identical values are collapsed so the timeline shows only when a goal
+        actually moved.
+    """
+    timelines: Dict[str, List[dict]] = {f: [] for f in GOAL_FIELDS}
+    last: Dict[str, Any] = {f: _SENTINEL for f in GOAL_FIELDS}
+
+    for rev in revisions:
+        fields = rev.get("fields") or {}
+        for field in GOAL_FIELDS:
+            value = fields.get(field)
+            if value == last[field]:
+                continue
+            last[field] = value
+            timelines[field].append(
+                {
+                    "value": value,
+                    "source": rev.get("source"),
+                    "at": rev.get("created_at"),
+                }
+            )
+    return timelines
+
+
+class _Sentinel:
+    pass
+
+
+# Distinct from None so a first snapshot whose goal is None still emits once.
+_SENTINEL = _Sentinel()
@@ -124,20 +124,36 @@ class ReflectionCoach:
        return messages

    async def _call_model(
-        self, profile: Dict[str, Any], messages: List[Dict[str, str]]
+        self,
+        profile: Dict[str, Any],
+        messages: List[Dict[str, str]],
+        system_suffix: str = "",
    ) -> str:
        response = await self.client.messages.create(
            model=self.model,
            max_tokens=MAX_TOKENS,
            system=SYSTEM_PROMPT.format(
                profile=self.profile_context(profile)
-            ),
+            )
+            + system_suffix,
            messages=messages,
        )
        return response.content[0].text

+    # Optional focus steers (Phase 5). Appended to the system prompt.
+    FOCUS_STEERS = {
+        "goals": (
+            "\n\nFOCUS: Concentrate this turn on helping the person sharpen "
+            "their near-term and long-term goals — make them concrete and "
+            "theirs. Still a mirror: clarify what THEY said, never assign goals."
+        )
+    }
+
    async def reflect(
-        self, profile: Dict[str, Any], history: List[Dict[str, str]]
+        self,
+        profile: Dict[str, Any],
+        history: List[Dict[str, str]],
+        focus: str = "",
    ) -> Dict[str, Any]:
        """Produce one coach turn.

@@ -156,9 +172,10 @@ class ReflectionCoach:
            ReflectionError on API failure or repeated parse failure.
        """
        messages = self._build_messages(history)
+        steer = self.FOCUS_STEERS.get(focus, "")

        try:
-            raw = await self._call_model(profile, messages)
+            raw = await self._call_model(profile, messages, system_suffix=steer)
        except Exception as exc:  # noqa: BLE001 - surface any SDK/transport error
            raise ReflectionError(f"Anthropic API call failed: {exc}") from exc

@@ -170,7 +187,9 @@ class ReflectionCoach:
                {"role": "user", "content": RETRY_REMINDER},
            ]
            try:
-                raw_retry = await self._call_model(profile, retry)
+                raw_retry = await self._call_model(
+                    profile, retry, system_suffix=steer
+                )
            except Exception as exc:  # noqa: BLE001
                raise ReflectionError(
                    f"Anthropic API call failed on retry: {exc}"
@@ -0,0 +1,137 @@
+"""Phase 5 smart tagging: suggest which foundation a logged task builds toward.
+
+The suggestion is exactly that — a suggestion the person confirms (by posting a
+task mapping). The tagger reads the person's own foundation text and the task
+label and proposes the best-fit foundation with a short rationale.
+"""
+import json
+from typing import Any, Dict
+
+from anthropic import AsyncAnthropic
+
+from app.services.foundations import FOUNDATIONS
+
+DEFAULT_MODEL = "claude-sonnet-4-6"
+MAX_TOKENS = 400
+
+SYSTEM_PROMPT = """You help a person tag a unit of work to one of their own self-discovery "foundations". You SUGGEST; the person confirms — never decide for them.
+
+THEIR FOUNDATIONS (their own words):
+{foundations}
+
+Given a task, pick the ONE foundation it most plausibly builds toward. If it is genuinely unclear, pick the closest and mark confidence low. Do not invent foundations.
+
+Valid foundation keys: love, strength, mission, vocation, short_term, long_term.
+
+Respond ONLY with valid JSON, no markdown fences:
+{{
+  "foundation": "<one of the valid keys>",
+  "rationale": "one short sentence, addressed to the person (you/your)",
+  "confidence": "high | medium | low"
+}}"""
+
+RETRY_REMINDER = (
+    "Your previous response could not be parsed. Respond ONLY with the single "
+    "valid JSON object described — no preamble, no markdown fences."
+)
+
+
+class TaggingError(Exception):
+    """Raised when tagging fails (API error or unparseable/invalid output)."""
+
+
+class FoundationTagger:
+    """Suggests a foundation for a task label."""
+
+    def __init__(self, api_key: str, model: str = DEFAULT_MODEL):
+        if not api_key:
+            raise TaggingError(
+                "ANTHROPIC_API_KEY is not set; cannot suggest a foundation."
+            )
+        self.model = model
+        self.client = AsyncAnthropic(api_key=api_key)
+
+    @staticmethod
+    def _foundations_block(foundations: Dict[str, str]) -> str:
+        return "\n".join(
+            f"- {key} ({FOUNDATIONS[key]}): {foundations.get(key) or '(none)'}"
+            for key in FOUNDATIONS
+        )
+
+    async def _call(self, system: str, messages: list) -> str:
+        resp = await self.client.messages.create(
+            model=self.model,
+            max_tokens=MAX_TOKENS,
+            system=system,
+            messages=messages,
+        )
+        return resp.content[0].text
+
+    async def suggest(
+        self, task_label: str, foundations: Dict[str, str]
+    ) -> Dict[str, Any]:
+        """Return ``{"foundation", "rationale", "confidence"}``. The foundation
+        is guaranteed to be a valid key. Raises TaggingError on failure."""
+        system = SYSTEM_PROMPT.format(
+            foundations=self._foundations_block(foundations)
+        )
+        messages = [
+            {"role": "user", "content": f"Task: {task_label}"}
+        ]
+        try:
+            raw = await self._call(system, messages)
+        except Exception as exc:  # noqa: BLE001
+            raise TaggingError(f"Anthropic API call failed: {exc}") from exc
+
+        try:
+            return self._parse(raw)
+        except (json.JSONDecodeError, ValueError):
+            retry = messages + [
+                {"role": "assistant", "content": raw},
+                {"role": "user", "content": RETRY_REMINDER},
+            ]
+            try:
+                raw_retry = await self._call(system, retry)
+            except Exception as exc:  # noqa: BLE001
+                raise TaggingError(
+                    f"Anthropic API call failed on retry: {exc}"
+                ) from exc
+            try:
+                return self._parse(raw_retry)
+            except (json.JSONDecodeError, ValueError) as exc:
+                raise TaggingError(
+                    f"Model did not return a valid suggestion: {exc}"
+                ) from exc
+
+    @staticmethod
+    def _strip_fences(text: str) -> str:
+        s = (text or "").strip()
+        if s.startswith("```"):
+            nl = s.find("\n")
+            if nl != -1:
+                s = s[nl + 1 :]
+            if s.rstrip().endswith("```"):
+                s = s.rstrip()[:-3]
+        return s.strip()
+
+    @classmethod
+    def _parse(cls, raw: str) -> Dict[str, Any]:
+        if not raw or not raw.strip():
+            raise ValueError("empty response")
+        data = json.loads(cls._strip_fences(raw))
+        if not isinstance(data, dict):
+            raise ValueError("not an object")
+        foundation = data.get("foundation")
+        if foundation not in FOUNDATIONS:
+            raise ValueError(f"invalid foundation: {foundation!r}")
+        confidence = data.get("confidence")
+        if confidence not in ("high", "medium", "low"):
+            confidence = "low"
+        rationale = data.get("rationale")
+        if not isinstance(rationale, str):
+            rationale = ""
+        return {
+            "foundation": foundation,
+            "rationale": rationale.strip(),
+            "confidence": confidence,
+        }
@@ -148,6 +148,9 @@
         </p>
         <p class="back-link" style="text-align:center;margin-top:10px">
           <a href="/static/dashboard.html">Where your time goes →</a>
+         </p>
+         <p class="back-link" style="text-align:center;margin-top:10px">
+           <a href="/static/visuals.html">See your profile visualized →</a>
         </p>`;

      content.innerHTML = `
@@ -398,6 +398,30 @@ textarea.edit {
  }
 }

+.viz-wrap {
+  display: flex;
+  justify-content: center;
+  margin-bottom: 20px;
+}
+
+.viz {
+  width: 100%;
+  max-width: 420px;
+  height: auto;
+}
+
+.goal-cols {
+  display: grid;
+  grid-template-columns: 1fr 1fr;
+  gap: 18px;
+}
+
+@media (max-width: 560px) {
+  .goal-cols {
+    grid-template-columns: 1fr;
+  }
+}
+
 .bars {
  display: flex;
  flex-direction: column;
@@ -0,0 +1,190 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8" />
+  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+  <title>ImpactFlow — Your Profile, Visualized</title>
+  <link rel="preconnect" href="https://fonts.googleapis.com" />
+  <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
+  <link
+    href="https://fonts.googleapis.com/css2?family=DM+Sans:wght@400;500;600;700&family=Playfair+Display:wght@500;600&display=swap"
+    rel="stylesheet"
+  />
+  <link rel="stylesheet" href="/static/style.css" />
+  <script src="/static/auth.js"></script>
+</head>
+<body>
+  <div class="wrap">
+    <div class="brand">ImpactFlow · Your Profile, Visualized</div>
+    <div id="content"><p>Loading…</p></div>
+    <p class="back-link"><a href="/static/profile.html">← Back to your profile</a></p>
+  </div>
+
+  <script>
+    const SVGNS = "http://www.w3.org/2000/svg";
+
+    function escapeHtml(s) {
+      if (s == null) return "";
+      return String(s).replace(/&/g, "&amp;").replace(/</g, "&lt;").replace(/>/g, "&gt;");
+    }
+    function el(tag, attrs, text) {
+      const n = document.createElementNS(SVGNS, tag);
+      for (const k in attrs) n.setAttribute(k, attrs[k]);
+      if (text != null) n.textContent = text;
+      return n;
+    }
+
+    // ---- Ikigai: four overlapping circles ----
+    function ikigaiSvg(p) {
+      const W = 460, H = 460, cx = W / 2, cy = H / 2, r = 130, off = 78;
+      const svg = el("svg", { viewBox: `0 0 ${W} ${H}`, class: "viz" });
+      const circles = [
+        { dx: 0, dy: -off, fill: "#c9a84c", label: "What you love", key: "love_summary" },
+        { dx: off, dy: 0, fill: "#2e7d32", label: "Good at", key: "strength_summary" },
+        { dx: 0, dy: off, fill: "#1c3049", label: "World needs", key: "mission_summary" },
+        { dx: -off, dy: 0, fill: "#a0522d", label: "Paid for", key: "vocation_summary" },
+      ];
+      circles.forEach((c) => {
+        svg.appendChild(
+          el("circle", {
+            cx: cx + c.dx, cy: cy + c.dy, r,
+            fill: c.fill, "fill-opacity": "0.22",
+            stroke: c.fill, "stroke-opacity": "0.7", "stroke-width": "1.5",
+          })
+        );
+      });
+      circles.forEach((c) => {
+        const lx = cx + c.dx * 1.7;
+        const ly = cy + c.dy * 1.7 + (c.dy === 0 ? 4 : 0);
+        svg.appendChild(
+          el("text", {
+            x: lx, y: ly, "text-anchor": "middle",
+            "font-size": "14", "font-weight": "600", fill: "#0d1b2a",
+          }, c.label)
+        );
+      });
+      svg.appendChild(
+        el("text", { x: cx, y: cy + 4, "text-anchor": "middle", "font-size": "15", "font-weight": "700", fill: "#0d1b2a" }, "Ikigai")
+      );
+      return svg;
+    }
+
+    // ---- Enneagram: 9 points, inner triangle + hexad, type/wing highlighted.
+    // We keep callouts in plain language rather than headlining a type number.
+    function enneagramSvg(p) {
+      const W = 420, H = 420, cx = W / 2, cy = H / 2, R = 160;
+      const deg = { 9: 0, 1: 40, 2: 80, 3: 120, 4: 160, 5: 200, 6: 240, 7: 280, 8: 320 };
+      const pt = (n) => {
+        const t = (deg[n] * Math.PI) / 180;
+        return [cx + R * Math.sin(t), cy - R * Math.cos(t)];
+      };
+      const svg = el("svg", { viewBox: `0 0 ${W} ${H}`, class: "viz" });
+      svg.appendChild(el("circle", { cx, cy, r: R, fill: "none", stroke: "#1c3049", "stroke-opacity": "0.3", "stroke-width": "1.5" }));
+      const line = (a, b) => {
+        const [x1, y1] = pt(a), [x2, y2] = pt(b);
+        svg.appendChild(el("line", { x1, y1, x2, y2, stroke: "#1c3049", "stroke-opacity": "0.25", "stroke-width": "1.2" }));
+      };
+      [[9, 3], [3, 6], [6, 9]].forEach(([a, b]) => line(a, b));
+      [[1, 4], [4, 2], [2, 8], [8, 5], [5, 7], [7, 1]].forEach(([a, b]) => line(a, b));
+
+      const type = p.probable_type, wing = p.wing;
+      for (let n = 1; n <= 9; n++) {
+        const [x, y] = pt(n);
+        const isType = n === type, isWing = n === wing;
+        svg.appendChild(
+          el("circle", {
+            cx: x, cy: y, r: isType ? 17 : 12,
+            fill: isType ? "#c9a84c" : "#fff",
+            stroke: isWing ? "#c9a84c" : "#1c3049",
+            "stroke-width": isWing ? "3" : "1.5",
+          })
+        );
+        svg.appendChild(
+          el("text", {
+            x, y: y + 4, "text-anchor": "middle", "font-size": "13",
+            "font-weight": isType ? "700" : "500",
+            fill: isType ? "#0d1b2a" : "#1c3049",
+          }, String(n))
+        );
+      }
+      return svg;
+    }
+
+    const TRIAD_LABEL = {
+      gut: "You lead with instinct and will",
+      heart: "You lead with feeling and connection",
+      head: "You lead with thought and perception",
+    };
+
+    async function loadGoalHistory() {
+      try {
+        const res = await authedFetch("/discovery/profile/me/goal-history");
+        if (!res.ok) return null;
+        return await res.json();
+      } catch {
+        return null;
+      }
+    }
+
+    function timelineHtml(entries) {
+      if (!entries || !entries.length)
+        return `<p class="edit-help">No changes recorded yet.</p>`;
+      return entries
+        .map((e) => {
+          const when = e.at ? new Date(e.at).toLocaleDateString() : "";
+          return `<div class="card" style="margin-bottom:12px">
+            <p style="margin:0 0 6px"><span class="dot ${
+              e.source === "extraction" ? "medium" : "high"
+            }"></span> <strong>${escapeHtml(e.source || "")}</strong> · ${when}</p>
+            <p style="margin:0">${escapeHtml(e.value) || "—"}</p>
+          </div>`;
+        })
+        .join("");
+    }
+
+    async function load() {
+      const content = document.getElementById("content");
+      let p;
+      try {
+        const res = await authedFetch("/discovery/profile/me");
+        if (!res.ok) throw new Error("Profile not found");
+        p = await res.json();
+      } catch (e) {
+        content.innerHTML = `<div class="error-box">${escapeHtml(e.message)}</div>`;
+        return;
+      }
+
+      content.innerHTML = `
+        <p class="section-label">Your Ikigai</p>
+        <div id="ikigai" class="viz-wrap"></div>
+        <div class="ikigai-grid">
+          <div class="card"><h3>♥ What you love</h3><p>${escapeHtml(p.love_summary) || "—"}</p></div>
+          <div class="card"><h3>★ What you're good at</h3><p>${escapeHtml(p.strength_summary) || "—"}</p></div>
+          <div class="card"><h3>◆ What the world needs</h3><p>${escapeHtml(p.mission_summary) || "—"}</p></div>
+          <div class="card"><h3>$ What you can be paid for</h3><p>${escapeHtml(p.vocation_summary) || "—"}</p></div>
+        </div>
+
+        <p class="section-label" style="margin-top:30px">Your centre of gravity</p>
+        <div id="enneagram" class="viz-wrap"></div>
+        <div class="triad-block"><p>${escapeHtml(
+          TRIAD_LABEL[p.triad] || "Your pattern"
+        )}. This map highlights where your energy sits and the point it leans toward.</p></div>
+
+        <p class="section-label" style="margin-top:30px">How your goals have evolved</p>
+        <div class="goal-cols">
+          <div><p class="edit-help">Near-term (6–12 months)</p><div id="st"></div></div>
+          <div><p class="edit-help">Long-term (3–5 years)</p><div id="lt"></div></div>
+        </div>
+      `;
+      document.getElementById("ikigai").appendChild(ikigaiSvg(p));
+      document.getElementById("enneagram").appendChild(enneagramSvg(p));
+
+      const hist = await loadGoalHistory();
+      document.getElementById("st").innerHTML = timelineHtml(hist && hist.short_term_goals);
+      document.getElementById("lt").innerHTML = timelineHtml(hist && hist.long_term_goals);
+    }
+
+    load();
+  </script>
+</body>
+</html>
@@ -0,0 +1,145 @@
+"""Tests for Phase 5 endpoints: goal-history and smart-tagging suggestion."""
+import json
+import uuid
+from datetime import datetime, timedelta, timezone
+
+import pytest
+
+API_KEY = {"X-API-Key": "test-api-key"}
+
+
+async def _seed_profile() -> str:
+    from app.auth import API_KEY_ADMIN_ID, ensure_api_key_admin
+    from app.database import AsyncSessionLocal
+    from app.models import DiscoveryConversation, DiscoveryProfile
+
+    async with AsyncSessionLocal() as db:
+        await ensure_api_key_admin(db)
+        conv = DiscoveryConversation(
+            id=str(uuid.uuid4()),
+            user_id=API_KEY_ADMIN_ID,
+            started_at=datetime.now(timezone.utc),
+        )
+        db.add(conv)
+        await db.commit()
+        profile = DiscoveryProfile(
+            id=str(uuid.uuid4()),
+            user_id=API_KEY_ADMIN_ID,
+            conversation_id=conv.id,
+            generated_at=datetime.now(timezone.utc),
+            triad="gut",
+            short_term_goals="run a pilot",
+            long_term_goals="statewide outfit",
+        )
+        db.add(profile)
+        await db.commit()
+        return profile.id
+
+
+async def _add_revision(profile_id, short, source, day):
+    from app.auth import API_KEY_ADMIN_ID
+    from app.database import AsyncSessionLocal
+    from app.models import ProfileRevision
+
+    async with AsyncSessionLocal() as db:
+        db.add(
+            ProfileRevision(
+                id=str(uuid.uuid4()),
+                profile_id=profile_id,
+                user_id=API_KEY_ADMIN_ID,
+                source=source,
+                fields_json=json.dumps(
+                    {"short_term_goals": short, "long_term_goals": "statewide outfit"}
+                ),
+                note=None,
+                created_at=datetime(2026, 6, day, tzinfo=timezone.utc),
+            )
+        )
+        await db.commit()
+
+
+# ---- goal history ----------------------------------------------------------
+
+
+async def test_goal_history_timeline(app_client):
+    pid = await _seed_profile()
+    await _add_revision(pid, "run a pilot", "extraction", 1)
+    await _add_revision(pid, "run a pilot", "manual_edit", 2)  # unchanged
+    await _add_revision(pid, "launch fall cohort", "reflection", 3)
+
+    r = await app_client.get(
+        "/discovery/profile/me/goal-history", headers=API_KEY
+    )
+    assert r.status_code == 200
+    st = r.json()["short_term_goals"]
+    assert [e["value"] for e in st] == ["run a pilot", "launch fall cohort"]
+    assert st[-1]["source"] == "reflection"
+
+
+async def test_goal_history_without_profile_404(app_client):
+    r = await app_client.get(
+        "/discovery/profile/me/goal-history", headers=API_KEY
+    )
+    assert r.status_code == 404
+
+
+# ---- smart tagging ---------------------------------------------------------
+
+
+class FakeTagger:
+    result = {"foundation": "short_term", "rationale": "builds your pilot", "confidence": "high"}
+
+    def __init__(self, api_key=None, model=None):
+        pass
+
+    async def suggest(self, task_label, foundations):
+        FakeTagger.last_label = task_label
+        return dict(FakeTagger.result)
+
+
+@pytest.fixture(autouse=True)
+async def fake_tagger(monkeypatch, app_client):
+    monkeypatch.setattr("app.routers.integration.FoundationTagger", FakeTagger)
+    yield
+
+
+async def test_suggest_foundation(app_client):
+    await _seed_profile()
+    r = await app_client.post(
+        "/discovery/integration/suggest-foundation",
+        headers=API_KEY,
+        json={"task_label": "practice welding for the cohort"},
+    )
+    assert r.status_code == 200
+    body = r.json()
+    assert body["foundation"] == "short_term"
+    assert body["label"]  # human label filled in
+    assert body["confidence"] == "high"
+    assert FakeTagger.last_label == "practice welding for the cohort"
+
+
+async def test_suggest_foundation_empty_label_400(app_client):
+    await _seed_profile()
+    r = await app_client.post(
+        "/discovery/integration/suggest-foundation",
+        headers=API_KEY,
+        json={"task_label": "   "},
+    )
+    assert r.status_code == 400
+
+
+async def test_suggest_foundation_without_profile_404(app_client):
+    r = await app_client.post(
+        "/discovery/integration/suggest-foundation",
+        headers=API_KEY,
+        json={"task_label": "anything"},
+    )
+    assert r.status_code == 404
+
+
+async def test_suggest_requires_auth(app_client):
+    r = await app_client.post(
+        "/discovery/integration/suggest-foundation",
+        json={"task_label": "x"},
+    )
+    assert r.status_code == 401
@@ -0,0 +1,47 @@
+"""Unit tests for the pure goal-evolution aggregator."""
+from datetime import datetime, timezone
+
+from app.services.profile_history import goal_history
+
+
+def _rev(short, long_, source, day):
+    return {
+        "fields": {"short_term_goals": short, "long_term_goals": long_},
+        "source": source,
+        "created_at": datetime(2026, 6, day, tzinfo=timezone.utc),
+    }
+
+
+def test_collapses_consecutive_identical_values():
+    revs = [
+        _rev("a", "x", "extraction", 1),
+        _rev("a", "x", "manual_edit", 2),  # no change -> skipped
+        _rev("b", "x", "reflection", 3),   # short changed
+    ]
+    h = goal_history(revs)
+    assert [e["value"] for e in h["short_term_goals"]] == ["a", "b"]
+    assert [e["value"] for e in h["long_term_goals"]] == ["x"]
+    assert h["short_term_goals"][1]["source"] == "reflection"
+
+
+def test_each_field_tracked_independently():
+    revs = [
+        _rev("a", "x", "extraction", 1),
+        _rev("a", "y", "reflection", 2),  # only long changed
+    ]
+    h = goal_history(revs)
+    assert len(h["short_term_goals"]) == 1
+    assert [e["value"] for e in h["long_term_goals"]] == ["x", "y"]
+
+
+def test_empty_revisions_yield_empty_timelines():
+    h = goal_history([])
+    assert h == {"short_term_goals": [], "long_term_goals": []}
+
+
+def test_first_value_emitted_even_if_none():
+    revs = [_rev(None, None, "extraction", 1), _rev("a", None, "reflection", 2)]
+    h = goal_history(revs)
+    assert [e["value"] for e in h["short_term_goals"]] == [None, "a"]
+    # long stayed None throughout -> a single None entry
+    assert [e["value"] for e in h["long_term_goals"]] == [None]
@@ -55,8 +55,10 @@ class FakeCoach:
    def __init__(self, api_key=None, model=None):
        pass

-    async def reflect(self, profile, history):
-        FakeCoach.calls.append({"profile": profile, "history": list(history)})
+    async def reflect(self, profile, history, focus=""):
+        FakeCoach.calls.append(
+            {"profile": profile, "history": list(history), "focus": focus}
+        )
        return dict(FakeCoach.response)


@@ -180,6 +182,17 @@ async def test_empty_message_after_opener_returns_400(app_client):
    assert r.status_code == 400


+async def test_reflect_passes_focus_to_coach(app_client):
+    """Phase 5: a goal-refinement focus reaches the coach."""
+    await _seed_profile()
+    await app_client.post(
+        "/discovery/profile/me/reflect",
+        headers=API_KEY,
+        json={"message": "let's sharpen my goals", "focus": "goals"},
+    )
+    assert FakeCoach.calls[-1]["focus"] == "goals"
+
+
 async def test_reflect_requires_auth(app_client):
    r = await app_client.post("/discovery/profile/me/reflect", json={})
    assert r.status_code == 401
@@ -75,6 +75,20 @@ def test_profile_page_links_to_dashboard():
    assert "/static/dashboard.html" in html


+def test_visuals_page_reads_profile_and_goal_history():
+    """Phase 5 visualizations read the profile and goal history via authedFetch."""
+    html = Path("app/static/visuals.html").read_text(encoding="utf-8")
+    assert "/discovery/profile/me" in html
+    assert "/discovery/profile/me/goal-history" in html
+    assert "authedFetch" in html
+    assert "user_id" not in html
+
+
+def test_profile_page_links_to_visuals():
+    html = Path("app/static/profile.html").read_text(encoding="utf-8")
+    assert "/static/visuals.html" in html
+
+
 def test_auth_helper_sends_credentials_and_refreshes():
    js = Path("app/static/auth.js").read_text(encoding="utf-8")