app/api/v1/endpoints/screener.py

@@ -22,8 +22,7 @@ logger = logging.getLogger("app.api.v1.screener")
     ttl=300,
     key_params=[
         "market_cap_min", "market_cap_max", "exchange", "min_avg_volume",
-        "min_dollar_volume", "exclude_types", "sector", "pe_min", "pe_max",
+        "exclude_types", "sector", "pe_min", "pe_max", "price_min", "price_max",
-        "price_min", "price_max",
         "page", "page_size", "sort_by", "sort_ascending",
     ],
 )
@@ -41,18 +40,7 @@ async def screen_stocks(
                     "Omit for all US exchanges.",
     ),
     min_avg_volume: Optional[int] = Query(
-        None, ge=0, description="Minimum 3-month average daily volume in SHARES (e.g. 500000)"
+        None, ge=0, description="Minimum 3-month average daily volume (e.g. 500000)"
-    ),
-    min_dollar_volume: Optional[float] = Query(
-        None,
-        ge=0,
-        description="Opt-in. Minimum 3-month average daily DOLLAR volume "
-                    "(regularMarketPrice × averageDailyVolume3Month), e.g. "
-                    "100000000 for $100M/day. When set, the full match set is "
-                    "fetched and post-filtered, and the share-count "
-                    "min_avg_volume gate is NOT applied (it would exclude "
-                    "high-priced low-share-volume names like BLK/KLAC at the "
-                    "source). When omitted, screener behaviour is unchanged.",
     ),
     exclude_types: Optional[str] = Query(
         None,
@@ -111,7 +99,6 @@ async def screen_stocks(
             market_cap_max=market_cap_max,
             exchange=exchange,
             min_avg_volume=min_avg_volume,
-            min_dollar_volume=min_dollar_volume,
             exclude_types=exclude_types,
             sector=sector,
             pe_min=pe_min,

app/services/screener_service.py

@@ -46,23 +46,6 @@ class ScreenerService:
         "price_to_book": "pricebook",
     }
 
-    # Maps API sort_by names to keys on the parsed-stock dict. Used only in
-    # min_dollar_volume mode, where results are assembled across multiple
-    # Yahoo pages and must be re-sorted server-side.
-    POST_SORT_KEY = {
-        "market_cap": "market_cap",
-        "volume": "volume",
-        "avg_volume": "avg_volume_3m",
-        "price": "price",
-        "pe_ratio": "pe_ratio",
-        "change_percent": "change_percent",
-        "name": "name",
-        "eps": "eps_ttm",
-        "dividend_yield": "dividend_yield",
-        "forward_pe": "forward_pe",
-        "price_to_book": "price_to_book",
-    }
-
     def _build_query(
         self,
         market_cap_min: Optional[float],
@@ -177,47 +160,6 @@ class ScreenerService:
         import yfinance_plus as yf
         return yf.screen(preset, offset=offset, count=count)
 
-    # Yahoo's screen() max page is 250; cap total internal fetch depth so a
-    # broad query (e.g. low market_cap_min) can't issue unbounded requests.
-    _COLLECT_PAGE = 250
-    _COLLECT_MAX_PAGES = 24  # up to 6000 rows
-
-    async def _collect_all_quotes(self, query, sort_field: str, sort_asc: bool):
-        """Page through Yahoo screen() until the full match set is collected.
-
-        Returns (quotes, yahoo_total, truncated). Used only by the
-        min_dollar_volume path, which must post-filter the complete set.
-        """
-        loop = asyncio.get_event_loop()
-        seen: dict = {}
-        offset = 0
-        yahoo_total = None
-        pages = 0
-        while pages < self._COLLECT_MAX_PAGES:
-            raw = await loop.run_in_executor(
-                None, self._screen_sync, query, offset,
-                self._COLLECT_PAGE, sort_field, sort_asc,
-            )
-            qs = raw.get('quotes', []) or []
-            if yahoo_total is None:
-                yahoo_total = raw.get('total') or raw.get('count') or 0
-            for q in qs:
-                sym = q.get('symbol')
-                if sym and sym not in seen:
-                    seen[sym] = q
-            pages += 1
-            offset += self._COLLECT_PAGE
-            if not qs or len(qs) < self._COLLECT_PAGE:
-                break
-            if yahoo_total and offset >= yahoo_total:
-                break
-        truncated = bool(
-            pages >= self._COLLECT_MAX_PAGES
-            and yahoo_total
-            and offset < yahoo_total
-        )
-        return list(seen.values()), (yahoo_total or len(seen)), truncated
-
     async def screen_preset(self, preset: str, page: int = 1, page_size: int = 25) -> dict:
         """Fetch a Yahoo Finance predefined screener (e.g. day_gainers)."""
         start_time = time.time()
@@ -231,10 +173,7 @@ class ScreenerService:
         )
 
         quotes = raw.get('quotes', [])
-        # Yahoo returns the full match count under 'total'; 'count' is only the
+        total_available = raw.get('count') or raw.get('total') or len(quotes)
-        # number of rows in THIS page (== size). Prefer 'total' so total_pages
-        # reflects the real result set, not a single page.
-        total_available = raw.get('total') or raw.get('count') or len(quotes)
         total_pages = max(1, (total_available + page_size - 1) // page_size)
 
         return {
@@ -257,7 +196,6 @@ class ScreenerService:
         market_cap_max: Optional[float] = None,
         exchange: Optional[str] = None,
         min_avg_volume: Optional[int] = None,
-        min_dollar_volume: Optional[float] = None,
         exclude_types: Optional[str] = None,
         sector: Optional[str] = None,
         pe_min: Optional[float] = None,
@@ -269,31 +207,17 @@ class ScreenerService:
         sort_by: str = "market_cap",
         sort_ascending: bool = False,
     ) -> dict:
-        """Screen stocks with the given filters and return paginated results.
+        """Screen stocks with the given filters and return paginated results."""
-
-        ``min_dollar_volume`` is opt-in. When it is None the behaviour is
-        byte-for-byte identical to before this parameter existed (single Yahoo
-        page, Yahoo applies ``min_avg_volume``). When set, the share-count
-        ``min_avg_volume`` filter is intentionally NOT pushed to Yahoo — that
-        would drop high-priced, low-share-volume names (BLK, KLAC, …) at the
-        source before dollar volume can be evaluated — and the complete match
-        set is fetched and post-filtered on price × averageDailyVolume3Month.
-        """
         start_time = time.time()
 
         page_size = max(1, min(page_size, 250))
         page = max(1, page)
 
-        dollar_mode = min_dollar_volume is not None
-        # In dollar-volume mode the share-count avg-volume gate is replaced by
-        # the dollar-volume gate, so it must not be sent to Yahoo.
-        effective_min_avg_volume = None if dollar_mode else min_avg_volume
-
         query = self._build_query(
             market_cap_min=market_cap_min,
             market_cap_max=market_cap_max,
             exchange=exchange,
-            min_avg_volume=effective_min_avg_volume,
+            min_avg_volume=min_avg_volume,
             sector=sector,
             pe_min=pe_min,
             pe_max=pe_max,
@@ -302,86 +226,35 @@ class ScreenerService:
         )
 
         sort_field = self.SORT_FIELD_MAP.get(sort_by, "intradaymarketcap")
+        offset = (page - 1) * page_size
+
+        loop = asyncio.get_event_loop()
+        raw = await loop.run_in_executor(
+            None,
+            self._screen_sync,
+            query,
+            offset,
+            page_size,
+            sort_field,
+            sort_ascending,
+        )
+
+        quotes = raw.get('quotes', [])
+        # yfinance may return total count under 'count' or 'total'
+        total_available = raw.get('count') or raw.get('total') or len(quotes)
 
         # Post-filter: remove non-equity types if requested
         exclude_type_set = set()
         if exclude_types:
             exclude_type_set = {t.strip().upper() for t in exclude_types.split(',')}
 
-        extra_meta: dict = {}
+        stocks = []
-
+        for quote in quotes:
-        if not dollar_mode:
+            if exclude_type_set:
-            # ---- Unchanged legacy path (zero regression when opt-in is off) ----
+                qt = (quote.get('quoteType') or '').upper()
-            offset = (page - 1) * page_size
+                if qt in exclude_type_set:
-            loop = asyncio.get_event_loop()
-            raw = await loop.run_in_executor(
-                None,
-                self._screen_sync,
-                query,
-                offset,
-                page_size,
-                sort_field,
-                sort_ascending,
-            )
-
-            quotes = raw.get('quotes', [])
-            # Yahoo's 'total' is the full match count; 'count' is only this
-            # page's row count. Prefer 'total' so total_pages is correct and
-            # clients that paginate by total_pages don't stop after page 1.
-            total_available = raw.get('total') or raw.get('count') or len(quotes)
-
-            stocks = []
-            for quote in quotes:
-                if exclude_type_set:
-                    qt = (quote.get('quoteType') or '').upper()
-                    if qt in exclude_type_set:
-                        continue
-                stocks.append(self._parse_quote(quote))
-        else:
-            # ---- Opt-in dollar-volume path: fetch full set, post-filter ----
-            all_quotes, _yahoo_total, truncated = await self._collect_all_quotes(
-                query, sort_field, sort_ascending
-            )
-
-            filtered = []
-            for quote in all_quotes:
-                if exclude_type_set:
-                    qt = (quote.get('quoteType') or '').upper()
-                    if qt in exclude_type_set:
-                        continue
-                price = quote.get('regularMarketPrice')
-                avg_vol = quote.get('averageDailyVolume3Month')
-                if price is None or avg_vol is None:
                     continue
-                if price * avg_vol < min_dollar_volume:
+            stocks.append(self._parse_quote(quote))
-                    continue
-                filtered.append(self._parse_quote(quote))
-
-            sort_key = self.POST_SORT_KEY.get(sort_by, "market_cap")
-            # Partition so rows missing the sort field are always last,
-            # regardless of sort direction, and so str/num keys never mix.
-            present = [s for s in filtered if s.get(sort_key) is not None]
-            missing = [s for s in filtered if s.get(sort_key) is None]
-            present.sort(key=lambda s: s.get(sort_key), reverse=not sort_ascending)
-            filtered = present + missing
-
-            total_available = len(filtered)
-            start = (page - 1) * page_size
-            stocks = filtered[start:start + page_size]
-
-            extra_meta['dollar_volume_mode'] = True
-            extra_meta['fetched_universe'] = len(all_quotes)
-            if min_avg_volume is not None:
-                extra_meta['note_min_avg_volume'] = (
-                    'min_avg_volume ignored because min_dollar_volume is set '
-                    '(dollar volume replaces the share-count liquidity gate)'
-                )
-            if truncated:
-                extra_meta['truncated'] = True
-                extra_meta['note_truncated'] = (
-                    'Yahoo result set exceeded internal fetch cap; widen '
-                    'market_cap_min to narrow the universe for completeness'
-                )
 
         query_time = time.time() - start_time
         total_pages = max(1, (total_available + page_size - 1) // page_size)
@@ -393,10 +266,8 @@ class ScreenerService:
             filters_applied['market_cap_max'] = market_cap_max
         if exchange:
             filters_applied['exchange'] = exchange
-        if min_avg_volume is not None and not dollar_mode:
+        if min_avg_volume is not None:
             filters_applied['min_avg_volume'] = min_avg_volume
-        if min_dollar_volume is not None:
-            filters_applied['min_dollar_volume'] = min_dollar_volume
         if exclude_types:
             filters_applied['exclude_types'] = exclude_types
         if sector:
@@ -424,7 +295,6 @@ class ScreenerService:
                 'sort_ascending': sort_ascending,
                 'source': 'yfinance_screen',
                 'note': 'sector/industry not included in per-stock response (Yahoo API limitation)',
-                **extra_meta,
             },
         }
 

app/services/universe_service.py

@@ -411,7 +411,7 @@ class UniverseService:
                         continue
 
                     market_cap = shares * close
-                    if math.isnan(market_cap) or market_cap <= 0 or market_cap > 2e13:
+                    if math.isnan(market_cap) or market_cap <= 0 or market_cap > 5e12:
                         continue
 
                     snapshot_dt = datetime(
@@ -596,7 +596,7 @@ class UniverseService:
         snapshot_dt = datetime(target.year, target.month, 1, tzinfo=timezone.utc)
 
         # Always exclude clearly bad data (sanity cap: $5T max, historical record is ~$3.7T)
-        _MAX_MARKET_CAP = 2e13
+        _MAX_MARKET_CAP = 5e12
         conditions = [
             UniverseSnapshot.snapshot_date == snapshot_dt,
             UniverseSnapshot.market_cap <= _MAX_MARKET_CAP,