pyr0ball
98695b00f0
Core trust scoring: - Five metadata signals (account age, feedback count/ratio, price vs market, category history), composited 0–100 - CV-based price signal suppression for heterogeneous search results (e.g. mixed laptop generations won't false-positive suspicious_price) - Expanded scratch/dent title detection: evasive redirects, functional problem phrases, DIY/repair indicators - Hard filters: new_account, established_bad_actor - Soft flags: low_feedback, suspicious_price, duplicate_photo, scratch_dent, long_on_market, significant_price_drop Search & filtering: - Browse API adapter (up to 200 items/page) + Playwright scraper fallback - OR-group query expansion for comprehensive variant coverage - Must-include (AND/ANY/groups), must-exclude, category, price range filters - Saved searches with full filter round-trip via URL params Seller enrichment: - Background BTF /itm/ scraping for account age (Kasada-safe headed Chromium) - On-demand enrichment: POST /api/enrich + ListingCard ↻ button - Category history derived from Browse API categories field (free, no extra calls) - Shopping API GetUserProfile inline enrichment for API adapter Market comps: - eBay Marketplace Insights API with Browse API fallback (catches 403 + 404) - Comps prioritised in ThreadPoolExecutor (submitted first) Infrastructure: - Staging DB fields: times_seen, first_seen_at, price_at_first_seen, category_name - Migrations 004 (staging tracking) + 005 (listing category) - eBay webhook handler stub - Cloud compose stack (compose.cloud.yml) - Vue frontend: search store, saved searches store, ListingCard, filter sidebar Docs: - README fully rewritten to reflect MVP status + full feature documentation - Roadmap table linked to all 13 Forgejo issues
60 lines
2.3 KiB
Python
60 lines
2.3 KiB
Python
from .metadata import MetadataScorer
|
|
from .photo import PhotoScorer
|
|
from .aggregator import Aggregator
|
|
from app.db.models import Seller, Listing, TrustScore
|
|
from app.db.store import Store
|
|
import hashlib
|
|
import math
|
|
|
|
|
|
class TrustScorer:
|
|
"""Orchestrates metadata + photo scoring for a batch of listings."""
|
|
|
|
def __init__(self, store: Store):
|
|
self._store = store
|
|
self._meta = MetadataScorer()
|
|
self._photo = PhotoScorer()
|
|
self._agg = Aggregator()
|
|
|
|
def score_batch(
|
|
self,
|
|
listings: list[Listing],
|
|
query: str,
|
|
) -> list[TrustScore]:
|
|
query_hash = hashlib.md5(query.encode()).hexdigest()
|
|
comp = self._store.get_market_comp("ebay", query_hash)
|
|
market_median = comp.median_price if comp else None
|
|
|
|
# Coefficient of variation: stddev/mean across batch prices.
|
|
# None when fewer than 2 priced listings (can't compute variance).
|
|
_prices = [l.price for l in listings if l.price > 0]
|
|
if len(_prices) >= 2:
|
|
_mean = sum(_prices) / len(_prices)
|
|
_stddev = math.sqrt(sum((p - _mean) ** 2 for p in _prices) / len(_prices))
|
|
price_cv: float | None = _stddev / _mean if _mean > 0 else None
|
|
else:
|
|
price_cv = None
|
|
|
|
photo_url_sets = [l.photo_urls for l in listings]
|
|
duplicates = self._photo.check_duplicates(photo_url_sets)
|
|
|
|
scores = []
|
|
for listing, is_dup in zip(listings, duplicates):
|
|
seller = self._store.get_seller("ebay", listing.seller_platform_id)
|
|
if seller:
|
|
signal_scores = self._meta.score(seller, market_median, listing.price, price_cv)
|
|
else:
|
|
signal_scores = {k: None for k in
|
|
["account_age", "feedback_count", "feedback_ratio",
|
|
"price_vs_market", "category_history"]}
|
|
trust = self._agg.aggregate(
|
|
signal_scores, is_dup, seller,
|
|
listing_id=listing.id or 0,
|
|
listing_title=listing.title,
|
|
times_seen=listing.times_seen,
|
|
first_seen_at=listing.first_seen_at,
|
|
price=listing.price,
|
|
price_at_first_seen=listing.price_at_first_seen,
|
|
)
|
|
scores.append(trust)
|
|
return scores
|