backend/tests/conftest.py

"""Shared test fixtures.

Tests run fully offline: the FAISS store is backed by a deterministic fake
embedder and no OpenAI key is configured, so mapping/grounding take their
deterministic fallback paths.
"""

from __future__ import annotations

import re
from pathlib import Path

import pytest
from docx import Document

from backend.config import settings
from backend.core import embeddings, notes_parser, pii_scrubber, rag_store
from backend.core.rics_canonical_l3 import build_section_anchor_text, iter_canonical_leaf_sections

_ANCHOR_OVERLAP_STOP = frozenset({
    "about", "elements", "other", "property", "section", "inside", "outside",
    "survey", "rics", "level", "typical", "matters", "engagement", "terms",
    "looked", "overall", "fine", "noted", "specific", "defects", "general",
    "inspection", "opinion", "declaration", "diagram", "recommendations",
    "the", "and", "not", "with", "for", "are", "was", "from", "this", "that",
    "have", "has", "were", "its", "any", "all", "may", "can", "your", "our",
})


class AnchorSemanticFakeEmbedder(embeddings.FakeEmbedder):
    """Deterministic embedder that approximates anchor similarity for offline tests."""

    def __init__(self, dim: int = 384) -> None:
        super().__init__(dim)
        self._anchor_texts: dict[str, str] = {}
        self._anchor_vecs: dict[str, list[float]] = {}

    def register_anchors(self, anchors: dict[str, str]) -> None:
        self._anchor_texts = dict(anchors)
        self._anchor_vecs = {sid: super()._vec(text) for sid, text in anchors.items()}

    @staticmethod
    def _token_set(text: str) -> set[str]:
        return {
            w
            for w in re.findall(r"[a-z0-9]{3,}", (text or "").lower())
            if w not in _ANCHOR_OVERLAP_STOP
        }

    def _vec(self, text: str) -> list[float]:
        base = super()._vec(text)
        if not self._anchor_texts:
            return base
        words = self._token_set(text)
        best_sid: str | None = None
        best_overlap = 0
        for sid, anchor in self._anchor_texts.items():
            overlap = len(words & self._token_set(anchor))
            if overlap > best_overlap:
                best_overlap = overlap
                best_sid = sid
        if best_sid and best_overlap > 0:
            anchor_vec = self._anchor_vecs[best_sid]
            alpha = min(0.95, 0.55 + 0.12 * best_overlap)
            blended = [alpha * a + (1.0 - alpha) * b for a, b in zip(anchor_vec, base)]
            norm = sum(x * x for x in blended) ** 0.5 or 1.0
            return [x / norm for x in blended]
        return base


@pytest.fixture(autouse=True)
def isolate(tmp_path, monkeypatch):
    monkeypatch.setattr(settings, "data_dir", str(tmp_path / "data"))
    monkeypatch.setattr(settings, "openai_api_key", "")
    monkeypatch.setattr(settings, "pii_use_spacy", False)
    monkeypatch.setattr(settings, "reference_auto_ingest_enabled", False)

    fake = AnchorSemanticFakeEmbedder()
    anchors = {
        sid: build_section_anchor_text(sid)
        for sid, _, _, _ in iter_canonical_leaf_sections()
    }
    fake.register_anchors(anchors)
    monkeypatch.setattr(embeddings, "get_embedder", lambda: fake)
    monkeypatch.setattr(rag_store, "get_embedder", lambda: fake)

    embeddings.reset_embedder()
    rag_store.reset_rag_store()
    pii_scrubber.reset_nlp_cache()
    notes_parser.reset_section_anchors()
    notes_parser.initialize_section_anchors()
    yield
    notes_parser.reset_section_anchors()
    rag_store.reset_rag_store()


def _write_report_template(path: Path) -> Path:
    """Minimal report template: section structure only."""
    doc = Document()
    doc.add_paragraph("A About the inspection")
    doc.add_paragraph("E1 Chimney Stacks")
    doc.add_paragraph("E2 Roof Coverings")
    doc.add_paragraph("G4 Central Heating")
    doc.save(str(path))
    return path


def _write_standard_paragraphs(path: Path, *, with_rating: bool = False) -> Path:
    """Standard paragraphs Word master: headings + boilerplate body."""
    doc = Document()
    doc.add_paragraph("Section E1 - Chimney Stacks")
    body = "The chimney stacks were inspected from ground level using binoculars."
    if with_rating:
        body += " Condition Rating 2 where defects require attention."
    doc.add_paragraph(body)
    doc.add_paragraph("Section E2 - Roof Coverings")
    doc.add_paragraph("The roof covering comprises slate laid to pitched timber rafters.")
    doc.add_paragraph("Section G4 - Central Heating")
    doc.add_paragraph("The property is served by a gas-fired central heating system.")
    doc.save(str(path))
    return path


@pytest.fixture
def report_template_docx(tmp_path) -> Path:
    return _write_report_template(tmp_path / "report_template.docx")


@pytest.fixture
def standard_paragraphs_docx(tmp_path) -> Path:
    return _write_standard_paragraphs(tmp_path / "standard_paragraphs.docx")


@pytest.fixture
def standard_paragraphs_rated_docx(tmp_path) -> Path:
    return _write_standard_paragraphs(tmp_path / "standard_paragraphs_rated.docx", with_rating=True)


# Legacy alias used by template_discoverer unit tests.
@pytest.fixture
def master_docx(standard_paragraphs_docx) -> Path:
    return standard_paragraphs_docx


@pytest.fixture
def master_docx_rated(standard_paragraphs_rated_docx) -> Path:
    return standard_paragraphs_rated_docx


@pytest.fixture
def client(configured_master):
    from backend.main import app
    from fastapi.testclient import TestClient

    with TestClient(app) as c:
        yield c


def _write_past_report(path: Path) -> Path:
    """Simulated uploaded past report for REFERENCE-tier generation."""
    doc = Document()
    doc.add_paragraph("Section E1 - Chimney Stacks")
    doc.add_paragraph(
        "The chimney stacks were inspected from ground level using binoculars. "
        "Inspection was limited to visible external surfaces at the time of survey."
    )
    doc.add_paragraph("Section E2 - Roof Coverings")
    doc.add_paragraph(
        "The roof coverings were inspected from ground level using binoculars and, "
        "where accessible, from within the roof void. The main roof covering "
        "comprises natural slate tiles laid to pitched timber rafters with underfelt "
        "and counter battens."
    )
    doc.add_paragraph("Section G4 - Central Heating")
    doc.add_paragraph(
        "The property was served by a gas-fired central heating system with "
        "radiators to most rooms."
    )
    doc.save(str(path))
    return path


@pytest.fixture
def configured_master(monkeypatch, report_template_docx, standard_paragraphs_docx) -> Path:
    """Point config at a split report template + standard paragraphs bundle.

    This fixture models an operator that explicitly configures AND enables a shared
    master, so it opts into ``master_template_auto_ingest`` (default-off in prod).
    """
    folder = report_template_docx.parent
    _write_past_report(folder / "past_report_sample.docx")
    monkeypatch.setattr(settings, "master_template_dir", str(folder))
    monkeypatch.setattr(settings, "report_template_filename", report_template_docx.name)
    monkeypatch.setattr(settings, "standard_paragraphs_filename", standard_paragraphs_docx.name)
    monkeypatch.setattr(settings, "master_template_filename", standard_paragraphs_docx.name)
    monkeypatch.setattr(settings, "master_template_auto_ingest", True)
    monkeypatch.setattr(settings, "reference_auto_ingest_enabled", True)
    return standard_paragraphs_docx