"""Demo dataset for the FablePool inspection CLI.

Seeds a node with a small but realistic personal memory graph for the
fictional user "Avery": calendar events, notes, and mock photo metadata as
raw evidence; a three-level derivation graph of claims on top of it.

The graph is intentionally shaped so the canonical demo flow works:

* refuting ``yoga_habit`` cascades through ``exercise_pattern`` into
  ``wellness_profile`` and ``packing_suggestion`` (three levels deep);
* correcting ``training_for`` cascades through ``trip_purpose`` into
  ``packing_suggestion``.

All timestamps are fixed so output is stable for tests and documentation.
"""

from __future__ import annotations

from dataclasses import dataclass, field
from typing import Any, Dict

from fablepool.errors import FablePoolError


@dataclass
class SeedResult:
    """Identifiers of everything the seed created, keyed by logical name."""

    evidence: Dict[str, str] = field(default_factory=dict)
    claims: Dict[str, str] = field(default_factory=dict)

    @property
    def evidence_count(self) -> int:
        return len(self.evidence)

    @property
    def claim_count(self) -> int:
        return len(self.claims)


#: Derivation depth of each seeded claim, used by docs and the seed command.
CLAIM_LEVELS: Dict[str, int] = {
    "yoga_habit": 1,
    "runs_mornings": 1,
    "dentist": 1,
    "home_city": 1,
    "work_schedule": 1,
    "training_for": 1,
    "buys_oat_milk": 1,
    "peanut_caution": 1,
    "upcoming_trip": 1,
    "exercise_pattern": 2,
    "diet_leaning": 2,
    "trip_purpose": 2,
    "wellness_profile": 3,
    "packing_suggestion": 3,
}


def seed_demo(node: Any) -> SeedResult:
    """Populate ``node`` with the Avery demo dataset. Refuses to double-seed."""
    for op in node.operations():
        if getattr(op, "kind", None) == "claim":
            raise FablePoolError(
                "this node already contains claims; seed onto a fresh node "
                "(fablepool init <new-dir> && fablepool --node <new-dir> seed)"
            )

    result = SeedResult()
    ev = result.evidence
    cl = result.claims

    def _id(obj: Any) -> str:
        for attr in ("id", "claim_id", "evidence_id", "op_id"):
            value = getattr(obj, attr, None)
            if value:
                return str(value)
        raise FablePoolError(f"seeded object {obj!r} has no identifier")

    # ------------------------------------------------------------------
    # Raw evidence: calendar
    # ------------------------------------------------------------------
    for name, day in (("yoga1", "02"), ("yoga2", "09"), ("yoga3", "16")):
        ev[name] = _id(
            node.record_evidence(
                source="calendar",
                media_type="calendar/event",
                content={
                    "summary": "Yoga class",
                    "location": "Harbor Yoga, Alameda",
                    "start": f"2025-09-{day}T18:00:00Z",
                    "end": f"2025-09-{day}T19:00:00Z",
                    "recurrence_hint": "weekly Tuesday",
                },
                observed_at=f"2025-09-{day}T18:00:00Z",
            )
        )

    for name, day in (("run1", "04"), ("run2", "11")):
        ev[name] = _id(
            node.record_evidence(
                source="calendar",
                media_type="calendar/event",
                content={
                    "summary": "Morning run — half marathon plan wk 6",
                    "start": f"2025-09-{day}T06:45:00Z",
                    "end": f"2025-09-{day}T07:30:00Z",
                },
                observed_at=f"2025-09-{day}T06:45:00Z",
            )
        )

    for name, date in (("dentist1", "2025-03-14"), ("dentist2", "2025-09-12")):
        ev[name] = _id(
            node.record_evidence(
                source="calendar",
                media_type="calendar/event",
                content={
                    "summary": "Dentist — Dr. Park",
                    "location": "Alameda Dental, 1200 Park St",
                    "start": f"{date}T10:00:00Z",
                    "end": f"{date}T10:45:00Z",
                },
                observed_at=f"{date}T10:00:00Z",
            )
        )

    for name, day in (("standup1", "08"), ("standup2", "15")):
        ev[name] = _id(
            node.record_evidence(
                source="calendar",
                media_type="calendar/event",
                content={
                    "summary": "Team standup",
                    "start": f"2025-09-{day}T09:30:00Z",
                    "end": f"2025-09-{day}T09:45:00Z",
                    "recurrence_hint": "weekdays",
                },
                observed_at=f"2025-09-{day}T09:30:00Z",
            )
        )

    ev["flight"] = _id(
        node.record_evidence(
            source="calendar",
            media_type="calendar/event",
            content={
                "summary": "Flight UA512 SFO → PDX",
                "start": "2025-10-10T16:05:00Z",
                "end": "2025-10-10T17:55:00Z",
                "location": "San Francisco International Airport",
                "notes": "return UA517 Oct 13",
            },
            observed_at="2025-09-20T12:00:00Z",
        )
    )

    # ------------------------------------------------------------------
    # Raw evidence: notes
    # ------------------------------------------------------------------
    ev["note_training"] = _id(
        node.record_evidence(
            source="notes",
            media_type="text/markdown",
            content={
                "title": "Portland Half Marathon — training plan",
                "text": (
                    "Race day Oct 12. Week 6 of 10. Long run Sundays, tempo "
                    "Thursdays. Goal: under 1:55. Booked hotel near Waterfront Park."
                ),
            },
            observed_at="2025-09-07T20:15:00Z",
        )
    )

    ev["note_grocery"] = _id(
        node.record_evidence(
            source="notes",
            media_type="text/markdown",
            content={
                "title": "Groceries",
                "text": "oat milk x2, bananas, lentils, sourdough, peanut-free granola, tofu",
            },
            observed_at="2025-09-13T09:10:00Z",
        )
    )

    ev["note_allergy"] = _id(
        node.record_evidence(
            source="notes",
            media_type="text/markdown",
            content={
                "title": "After dinner",
                "text": "Mild itchiness after pad thai again — peanuts? Watch this, maybe ask Dr. Park for a referral.",
            },
            observed_at="2025-08-29T21:40:00Z",
        )
    )

    # ------------------------------------------------------------------
    # Raw evidence: mock photo metadata
    # ------------------------------------------------------------------
    for name, (day, hour) in (
        ("photo_home1", ("05", "08")),
        ("photo_home2", ("12", "19")),
        ("photo_home3", ("19", "07")),
    ):
        ev[name] = _id(
            node.record_evidence(
                source="photos",
                media_type="image/metadata",
                content={
                    "filename": f"IMG_2025-09-{day}_{hour}00.jpg",
                    "taken_at": f"2025-09-{day}T{hour}:00:00Z",
                    "gps": {"lat": 37.7652, "lon": -122.2416},
                    "camera": "Pixel 8",
                },
                observed_at=f"2025-09-{day}T{hour}:05:00Z",
            )
        )

    ev["photo_trail"] = _id(
        node.record_evidence(
            source="photos",
            media_type="image/metadata",
            content={
                "filename": "IMG_2025-09-14_0712.jpg",
                "taken_at": "2025-09-14T07:12:00Z",
                "gps": {"lat": 37.8324, "lon": -122.1986},
                "camera": "Pixel 8",
                "scene_hint": "outdoor trail",
            },
            observed_at="2025-09-14T07:20:00Z",
        )
    )

    # ------------------------------------------------------------------
    # Level 1 claims (derived directly from evidence)
    # ------------------------------------------------------------------
    cl["yoga_habit"] = _id(
        node.record_claim(
            predicate="attends_weekly_yoga",
            value="Tuesdays 18:00 at Harbor Yoga",
            topic="fitness",
            confidence=0.86,
            sources=[ev["yoga1"], ev["yoga2"], ev["yoga3"]],
            rule="recurring_event_pattern",
        )
    )

    cl["runs_mornings"] = _id(
        node.record_claim(
            predicate="runs_in_the_morning",
            value="roughly twice a week before 08:00",
            topic="fitness",
            confidence=0.74,
            sources=[ev["run1"], ev["run2"], ev["photo_trail"]],
            rule="recurring_event_pattern",
        )
    )

    cl["dentist"] = _id(
        node.record_claim(
            predicate="dentist_is",
            value="Dr. Park (Alameda Dental)",
            topic="health",
            confidence=0.90,
            sources=[ev["dentist1"], ev["dentist2"]],
            rule="repeated_provider_appointment",
        )
    )

    cl["home_city"] = _id(
        node.record_claim(
            predicate="home_city",
            value="Alameda, CA",
            topic="location",
            confidence=0.78,
            sources=[ev["photo_home1"], ev["photo_home2"], ev["photo_home3"]],
            rule="photo_gps_cluster",
        )
    )

    cl["work_schedule"] = _id(
        node.record_claim(
            predicate="works_weekday_mornings",
            value="daily standup at 09:30, Monday–Friday",
            topic="work",
            confidence=0.81,
            sources=[ev["standup1"], ev["standup2"]],
            rule="recurring_event_pattern",
        )
    )

    cl["training_for"] = _id(
        node.record_claim(
            predicate="training_for_race",
            value="Portland Half Marathon (Oct 12)",
            topic="fitness",
            confidence=0.83,
            sources=[ev["note_training"], ev["run1"], ev["run2"]],
            rule="note_goal_extraction",
        )
    )

    cl["buys_oat_milk"] = _id(
        node.record_claim(
            predicate="buys_oat_milk",
            value="recurring grocery item",
            topic="preferences",
            confidence=0.66,
            sources=[ev["note_grocery"]],
            rule="note_item_frequency",
        )
    )

    cl["peanut_caution"] = _id(
        node.record_claim(
            predicate="possible_peanut_sensitivity",
            value="suspected, unconfirmed",
            topic="health",
            confidence=0.52,
            sources=[ev["note_allergy"], ev["note_grocery"]],
            rule="note_health_mention",
        )
    )

    cl["upcoming_trip"] = _id(
        node.record_claim(
            predicate="upcoming_trip",
            value="Portland, OR — Oct 10 to Oct 13",
            topic="travel",
            confidence=0.88,
            sources=[ev["flight"]],
            rule="travel_event_detection",
        )
    )

    # ------------------------------------------------------------------
    # Level 2 claims (derived from level-1 claims)
    # ------------------------------------------------------------------
    cl["exercise_pattern"] = _id(
        node.record_claim(
            predicate="exercise_pattern",
            value="regularly active: weekly yoga plus morning runs",
            topic="fitness",
            confidence=0.80,
            sources=[cl["yoga_habit"], cl["runs_mornings"]],
            rule="habit_aggregation",
        )
    )

    cl["diet_leaning"] = _id(
        node.record_claim(
            predicate="diet_leaning",
            value="plant-leaning (non-dairy milk, tofu, lentils)",
            topic="preferences",
            confidence=0.58,
            sources=[cl["buys_oat_milk"]],
            rule="preference_inference",
        )
    )

    cl["trip_purpose"] = _id(
        node.record_claim(
            predicate="trip_purpose",
            value="race weekend — Portland Half Marathon",
            topic="travel",
            confidence=0.76,
            sources=[cl["upcoming_trip"], cl["training_for"]],
            rule="goal_event_correlation",
        )
    )

    # ------------------------------------------------------------------
    # Level 3 claims (derived from level-2 claims)
    # ------------------------------------------------------------------
    cl["wellness_profile"] = _id(
        node.record_claim(
            predicate="wellness_profile",
            value="active lifestyle with a plant-leaning diet",
            topic="health",
            confidence=0.62,
            sources=[cl["exercise_pattern"], cl["diet_leaning"]],
            rule="profile_synthesis",
        )
    )

    cl["packing_suggestion"] = _id(
        node.record_claim(
            predicate="suggests_packing",
            value="running gear and race kit for the Portland trip",
            topic="travel",
            confidence=0.70,
            sources=[cl["trip_purpose"], cl["exercise_pattern"]],
            rule="trip_preparation",
        )
    )

    return result