"""Diff engine — catalog × canonical products → preview records (SD-0002 §6.5.2). Classifies each canonical product against the storefront's current catalog as add / update / unchanged / error, producing JSON-ready preview records (stored as draft JSONB, served verbatim to the SPA) plus a summary and a deterministic fingerprint that detects catalog drift between preview and confirm (INV-11). Deliberately DB-free: the catalog snapshot dataclasses are defined here and the repo layer builds them. Only fields present in the file's canonical fields{} participate in a comparison — an absent column is untouched, never a change (§6.5.1); a present-but-empty cell resolves to the field's CLEAR_DEFAULTS entry. Catalog variants/images absent from the file are likewise untouched (INV-10); file variants match catalog variants by their option-value combination (INV-13). """ from __future__ import annotations import hashlib import json from dataclasses import dataclass from decimal import Decimal from .models import CLEAR_DEFAULTS, CanonicalProduct, CanonicalVariant @dataclass class CatalogVariant: id: int options: tuple[str | None, str | None, str | None] position: int # sku, barcode, price (Decimal|None), cost, weight, weight_unit, volume, # volume_unit, tax_id_1, tax_id_2, inventory_tracker, inventory_qty, # variant_image (linked image source_url or None) fields: dict[str, object] @dataclass class CatalogImage: id: int source_url: str position: int alt_text: str | None @dataclass class CatalogProduct: id: int handle: str title: str option_names: tuple[str | None, str | None, str | None] # title, description_html, vendor, product_type, google_product_category, # tags (list[str]), status, published (bool) fields: dict[str, object] variants: list[CatalogVariant] images: list[CatalogImage] @dataclass(frozen=True) class DiffResult: records: list[dict] summary: dict fingerprint: str # Option-name presence markers in fields{} (see validate.py); the values are # compared via the option_names attribute, not through the generic field loop. _OPTION_NAME_FIELDS = ("option1_name", "option2_name", "option3_name") _SUMMARY_KEY = {"add": "adds", "update": "updates", "unchanged": "unchanged", "error": "errors"} def compute_diff(catalog: dict[str, CatalogProduct], products: list[CanonicalProduct]) -> DiffResult: records: list[dict] = [] summary = {"adds": 0, "updates": 0, "unchanged": 0, "errors": 0} for product in products: current = catalog.get(product.handle) if product.errors: kind, detail = "error", {"errors": [e.as_json() for e in product.errors]} elif current is None: kind, detail = "add", _add_detail(product) else: detail = _update_detail(product, current) kind = "update" if detail else "unchanged" records.append( { "handle": product.handle, "title": product.title or (current.title if current else ""), "kind": kind, "variant_count": len(current.variants) if kind == "unchanged" else len(product.variants), "detail": detail, } ) summary[_SUMMARY_KEY[kind]] += 1 fingerprint = hashlib.sha256( json.dumps(records, sort_keys=True, separators=(",", ":")).encode() ).hexdigest() return DiffResult(records=records, summary=summary, fingerprint=fingerprint) def _add_detail(product: CanonicalProduct) -> dict: # On add, absent fields fall back to their defaults for display where one # exists — the detail shows what will actually be set. set_fields = _resolved_product_fields(product) for field_name, default in CLEAR_DEFAULTS.items(): set_fields.setdefault(field_name, default) return { "set": {f: _json_safe(v) for f, v in set_fields.items()}, "option_names": list(product.option_names), "variants": [ {"options": list(v.options), "set": _resolved_variant_fields(v)} for v in product.variants ], "images": [ {"src": i.source_url, "position": i.position, "alt_text": i.alt_text} for i in product.images ], } def _update_detail(product: CanonicalProduct, current: CatalogProduct) -> dict: changes: list[dict] = [] # Title is always file-present (required header column); "" means the block # already carries an error and never reaches here. if product.title and product.title != current.title: changes.append(_change("title", current.title, product.title)) for field_name, resolved in _resolved_product_fields(product).items(): before = current.fields.get(field_name) if resolved != before: changes.append(_change(field_name, before, resolved)) for slot in (1, 2, 3): if f"option{slot}_name" not in product.fields: continue before, after = current.option_names[slot - 1], product.option_names[slot - 1] if after != before: changes.append(_change(f"option{slot}_name", before, after)) variant_entries: list[dict] = [] by_options = {v.options: v for v in current.variants} for variant in product.variants: match = by_options.get(variant.options) if match is None: variant_entries.append( {"options": list(variant.options), "kind": "add", "set": _resolved_variant_fields(variant)} ) continue variant_changes = [] for f, resolved in _resolved_fields(variant.fields).items(): # position lives on the catalog variant as an attribute, not in # fields{} — compare it explicitly (as images do for theirs). before = match.position if f == "position" else match.fields.get(f) if resolved != before: variant_changes.append(_change(f, before, resolved)) if variant_changes: variant_entries.append( {"options": list(variant.options), "kind": "update", "changes": variant_changes} ) image_entries: list[dict] = [] by_src = {i.source_url: i for i in current.images} for image in product.images: match = by_src.get(image.source_url) if match is None: image_entries.append( {"src": image.source_url, "kind": "add", "position": image.position, "alt_text": image.alt_text} ) continue image_changes = [ _change(f, before, after) for f, before, after in ( ("position", match.position, image.position), ("alt_text", match.alt_text, image.alt_text), ) if after != before ] if image_changes: image_entries.append({"src": image.source_url, "kind": "update", "changes": image_changes}) detail: dict = {} if changes: detail["changes"] = changes if variant_entries: detail["variants"] = variant_entries if image_entries: detail["images"] = image_entries return detail def _resolved_fields(fields: dict[str, object]) -> dict[str, object]: """File-present fields with None (an explicit clear) resolved to the default.""" return {f: (v if v is not None else CLEAR_DEFAULTS.get(f)) for f, v in fields.items()} def _resolved_product_fields(product: CanonicalProduct) -> dict[str, object]: return { f: v for f, v in _resolved_fields(product.fields).items() if f not in _OPTION_NAME_FIELDS } def _resolved_variant_fields(variant: CanonicalVariant) -> dict[str, object]: return {f: _json_safe(v) for f, v in _resolved_fields(variant.fields).items()} def _change(field_name: str, before: object, after: object) -> dict: return {"field": field_name, "before": _json_safe(before), "after": _json_safe(after)} def _json_safe(value: object) -> object: if isinstance(value, Decimal): return str(value) if isinstance(value, (tuple, list)): return [_json_safe(v) for v in value] return value