feat(dataset): 库内浏览支持跳转页、排序与类目价格筛选

新增 JdLeafCategoryNorm 与 price 解析字段；数据集 API 支持 sort/order、category_norm_id、price 区间与类目路径模糊；摘要返回类目选项；前端 JobDatasetModal 增加筛选与分页跳转。

Made-with: Cursor

backend/pipeline/dataset_api.py

@@ -0,0 +1,169 @@
+"""库内数据浏览 API：排序、价格与类目筛选（查询参数解析与 QuerySet 变换）。"""
+from __future__ import annotations
+
+from typing import Any
+
+from django.db.models import F, QuerySet
+from django.db.models.expressions import OrderBy
+from rest_framework.request import Request
+
+SEARCH_SORT_FIELDS = frozenset({"row_index", "price", "sku_id", "title", "leaf_category"})
+DETAIL_SORT_FIELDS = frozenset(
+    {"row_index", "price", "sku_id", "detail_category_path", "detail_brand"}
+)
+MERGED_SORT_FIELDS = frozenset(
+    {"row_index", "price", "sku_id", "title", "leaf_category", "detail_category_path"}
+)
+
+
+def _parse_opt_float(val: str | None) -> float | None:
+    if val is None or not str(val).strip():
+        return None
+    try:
+        return float(str(val).strip())
+    except ValueError:
+        return None
+
+
+def parse_sort_meta(request: Request) -> tuple[str, bool]:
+    sort = (request.query_params.get("sort") or "row_index").strip()
+    order = (request.query_params.get("order") or "asc").strip().lower()
+    desc = order == "desc"
+    return sort, desc
+
+
+def price_bounds_from_request(request: Request) -> tuple[float | None, float | None]:
+    return (
+        _parse_opt_float(request.query_params.get("price_min")),
+        _parse_opt_float(request.query_params.get("price_max")),
+    )
+
+
+def category_norm_id_from_request(request: Request) -> int | None:
+    raw = (request.query_params.get("category_norm_id") or "").strip()
+    if raw.isdigit():
+        return int(raw)
+    return None
+
+
+def detail_category_q_from_request(request: Request) -> str:
+    return (request.query_params.get("detail_category_q") or "").strip()
+
+
+def filter_echo(
+    *,
+    category_norm_id: int | None,
+    price_min: float | None,
+    price_max: float | None,
+    detail_category_q: str,
+    sort: str,
+    desc: bool,
+) -> dict[str, Any]:
+    return {
+        "category_norm_id": category_norm_id,
+        "price_min": price_min,
+        "price_max": price_max,
+        "detail_category_q": detail_category_q or None,
+        "sort": sort,
+        "order": "desc" if desc else "asc",
+    }
+
+
+def apply_search_filters(qs: QuerySet, request: Request) -> QuerySet:
+    cid = category_norm_id_from_request(request)
+    if cid is not None:
+        qs = qs.filter(leaf_category_norm_id=cid)
+    pmin, pmax = price_bounds_from_request(request)
+    if pmin is not None:
+        qs = qs.filter(price_value__gte=pmin)
+    if pmax is not None:
+        qs = qs.filter(price_value__lte=pmax)
+    return qs
+
+
+def apply_search_order(qs: QuerySet, sort: str, desc: bool) -> QuerySet:
+    sort = sort if sort in SEARCH_SORT_FIELDS else "row_index"
+    if sort == "price":
+        return qs.order_by(
+            OrderBy(F("price_value"), descending=desc, nulls_last=True),
+            "row_index",
+        )
+    if sort == "row_index":
+        return qs.order_by(OrderBy(F("row_index"), descending=desc))
+    field = {
+        "sku_id": "sku_id",
+        "title": "title",
+        "leaf_category": "leaf_category",
+    }[sort]
+    return qs.order_by(
+        OrderBy(F(field), descending=desc, nulls_last=True),
+        "row_index",
+    )
+
+
+def apply_detail_filters(qs: QuerySet, request: Request) -> QuerySet:
+    q = detail_category_q_from_request(request)
+    if q:
+        qs = qs.filter(detail_category_path__icontains=q)
+    pmin, pmax = price_bounds_from_request(request)
+    if pmin is not None:
+        qs = qs.filter(detail_price_value__gte=pmin)
+    if pmax is not None:
+        qs = qs.filter(detail_price_value__lte=pmax)
+    return qs
+
+
+def apply_detail_order(qs: QuerySet, sort: str, desc: bool) -> QuerySet:
+    sort = sort if sort in DETAIL_SORT_FIELDS else "row_index"
+    if sort == "price":
+        return qs.order_by(
+            OrderBy(F("detail_price_value"), descending=desc, nulls_last=True),
+            "row_index",
+        )
+    if sort == "row_index":
+        return qs.order_by(OrderBy(F("row_index"), descending=desc))
+    field = {
+        "sku_id": "sku_id",
+        "detail_category_path": "detail_category_path",
+        "detail_brand": "detail_brand",
+    }[sort]
+    return qs.order_by(
+        OrderBy(F(field), descending=desc, nulls_last=True),
+        "row_index",
+    )
+
+
+def apply_merged_filters(qs: QuerySet, request: Request) -> QuerySet:
+    cid = category_norm_id_from_request(request)
+    if cid is not None:
+        qs = qs.filter(leaf_category_norm_id=cid)
+    q = detail_category_q_from_request(request)
+    if q:
+        qs = qs.filter(detail_category_path__icontains=q)
+    pmin, pmax = price_bounds_from_request(request)
+    if pmin is not None:
+        qs = qs.filter(price_value__gte=pmin)
+    if pmax is not None:
+        qs = qs.filter(price_value__lte=pmax)
+    return qs
+
+
+def apply_merged_order(qs: QuerySet, sort: str, desc: bool) -> QuerySet:
+    sort = sort if sort in MERGED_SORT_FIELDS else "row_index"
+    if sort == "price":
+        return qs.order_by(
+            OrderBy(F("price_value"), descending=desc, nulls_last=True),
+            "row_index",
+        )
+    if sort == "row_index":
+        return qs.order_by(OrderBy(F("row_index"), descending=desc))
+    field = {
+        "sku_id": "sku_id",
+        "title": "title",
+        "leaf_category": "leaf_category",
+        "detail_category_path": "detail_category_path",
+    }[sort]
+    return qs.order_by(
+        OrderBy(F(field), descending=desc, nulls_last=True),
+        "row_index",
+    )

backend/pipeline/ingest.py

@@ -34,10 +34,12 @@ from .models import (
     JdJobDetailRow,
     JdJobMergedRow,
     JdJobSearchRow,
+    JdLeafCategoryNorm,
     JdProduct,
     JdProductSnapshot,
     PipelineJob,
 )
+from .price_parse import effective_list_price_value, float_price_from_cell
 
 logger = logging.getLogger(__name__)
 
@@ -171,11 +173,45 @@ def ingest_job_dataset_rows(job: PipelineJob) -> dict[str, Any]:
     search_rows = _read_csv_rows(search_path)
     if not search_rows and search_path.is_file() is False:
         pass
-    s_objs: list[JdJobSearchRow] = []
+    search_kw_list: list[tuple[int, dict[str, str]]] = []
+    leaf_labels: set[str] = set()
     for i, row in enumerate(search_rows):
         _normalize_search_csv_total_sales(row)
         kw = _search_row_kwargs(row)
-        s_objs.append(JdJobSearchRow(job=job, row_index=i, **kw))
+        search_kw_list.append((i, kw))
+        lc = (kw.get("leaf_category") or "").strip()[:512]
+        if lc:
+            leaf_labels.add(lc)
+    norm_map: dict[str, JdLeafCategoryNorm] = {}
+    if leaf_labels:
+        have = set(
+            JdLeafCategoryNorm.objects.filter(label__in=leaf_labels).values_list(
+                "label", flat=True
+            )
+        )
+        missing = [JdLeafCategoryNorm(label=l) for l in leaf_labels if l not in have]
+        if missing:
+            JdLeafCategoryNorm.objects.bulk_create(missing, ignore_conflicts=True)
+        norm_map = {
+            n.label: n
+            for n in JdLeafCategoryNorm.objects.filter(label__in=leaf_labels)
+        }
+    s_objs: list[JdJobSearchRow] = []
+    for i, kw in search_kw_list:
+        lc = (kw.get("leaf_category") or "").strip()[:512]
+        norm = norm_map.get(lc) if lc else None
+        pv = effective_list_price_value(
+            kw.get("coupon_price"), kw.get("price"), kw.get("original_price")
+        )
+        s_objs.append(
+            JdJobSearchRow(
+                job=job,
+                row_index=i,
+                leaf_category_norm=norm,
+                price_value=pv,
+                **kw,
+            )
+        )
     _bulk_create_in_chunks(JdJobSearchRow, s_objs)
     stats["search_rows"] = len(s_objs)
 
@@ -184,7 +220,12 @@ def ingest_job_dataset_rows(job: PipelineJob) -> dict[str, Any]:
     d_objs: list[JdJobDetailRow] = []
     for i, row in enumerate(detail_rows):
         kw = _detail_row_kwargs(row)
-        d_objs.append(JdJobDetailRow(job=job, row_index=i, **kw))
+        dpv = float_price_from_cell(kw.get("detail_price_final"))
+        d_objs.append(
+            JdJobDetailRow(
+                job=job, row_index=i, detail_price_value=dpv, **kw
+            )
+        )
     _bulk_create_in_chunks(JdJobDetailRow, d_objs)
     stats["detail_rows"] = len(d_objs)
 
@@ -199,11 +240,47 @@ def ingest_job_dataset_rows(job: PipelineJob) -> dict[str, Any]:
 
     merged_path = run_dir / FILE_MERGED_CSV
     merged_rows = _read_csv_rows(merged_path) if merged_path.is_file() else []
-    m_objs: list[JdJobMergedRow] = []
+    merged_kw_list: list[tuple[int, dict[str, str]]] = []
+    leaf_labels_m: set[str] = set()
     for i, row in enumerate(merged_rows):
         _normalize_merged_csv_total_sales(row)
         kw = _merged_row_kwargs(row)
-        m_objs.append(JdJobMergedRow(job=job, row_index=i, **kw))
+        merged_kw_list.append((i, kw))
+        lc = (kw.get("leaf_category") or "").strip()[:512]
+        if lc:
+            leaf_labels_m.add(lc)
+    norm_map_m: dict[str, JdLeafCategoryNorm] = {}
+    if leaf_labels_m:
+        have_m = set(
+            JdLeafCategoryNorm.objects.filter(label__in=leaf_labels_m).values_list(
+                "label", flat=True
+            )
+        )
+        missing_m = [
+            JdLeafCategoryNorm(label=l) for l in leaf_labels_m if l not in have_m
+        ]
+        if missing_m:
+            JdLeafCategoryNorm.objects.bulk_create(missing_m, ignore_conflicts=True)
+        norm_map_m = {
+            n.label: n
+            for n in JdLeafCategoryNorm.objects.filter(label__in=leaf_labels_m)
+        }
+    m_objs: list[JdJobMergedRow] = []
+    for i, kw in merged_kw_list:
+        lc = (kw.get("leaf_category") or "").strip()[:512]
+        norm = norm_map_m.get(lc) if lc else None
+        pv = effective_list_price_value(
+            kw.get("coupon_price"), kw.get("price"), kw.get("original_price")
+        )
+        m_objs.append(
+            JdJobMergedRow(
+                job=job,
+                row_index=i,
+                leaf_category_norm=norm,
+                price_value=pv,
+                **kw,
+            )
+        )
     _bulk_create_in_chunks(JdJobMergedRow, m_objs)
     stats["merged_table_rows"] = len(m_objs)
 

backend/pipeline/migrations/0017_dataset_browse_filters.py

@@ -0,0 +1,167 @@
+# Generated by Django 5.2.1 on 2026-04-16 01:49
+
+import re
+
+import django.db.models.deletion
+from django.db import migrations, models
+
+
+def _float_price_from_cell(s):
+    t = (s or "").strip().replace(",", "").replace("，", "")
+    if not t:
+        return None
+    m = re.search(r"(\d+(?:\.\d+)?)", t)
+    if not m:
+        return None
+    try:
+        v = float(m.group(1))
+    except ValueError:
+        return None
+    if 0 < v < 1_000_000:
+        return v
+    return None
+
+
+def _effective_list_price_value(coupon, price, original):
+    for s in (coupon, price, original):
+        v = _float_price_from_cell(s)
+        if v is not None:
+            return v
+    return None
+
+
+def backfill_dataset_browse_fields(apps, schema_editor):
+    JdLeafCategoryNorm = apps.get_model("pipeline", "JdLeafCategoryNorm")
+    JdJobSearchRow = apps.get_model("pipeline", "JdJobSearchRow")
+    JdJobMergedRow = apps.get_model("pipeline", "JdJobMergedRow")
+    JdJobDetailRow = apps.get_model("pipeline", "JdJobDetailRow")
+
+    raw_labels = set(
+        JdJobSearchRow.objects.exclude(leaf_category="").values_list(
+            "leaf_category", flat=True
+        )
+    )
+    raw_labels |= set(
+        JdJobMergedRow.objects.exclude(leaf_category="").values_list(
+            "leaf_category", flat=True
+        )
+    )
+    labels = sorted({(x or "").strip()[:512] for x in raw_labels if (x or "").strip()})
+    if labels:
+        JdLeafCategoryNorm.objects.bulk_create(
+            [JdLeafCategoryNorm(label=lbl) for lbl in labels],
+            ignore_conflicts=True,
+        )
+    norm_by_label = {n.label: n.id for n in JdLeafCategoryNorm.objects.all()}
+
+    chunk: list = []
+    for r in JdJobSearchRow.objects.all().iterator(chunk_size=400):
+        lbl = (r.leaf_category or "").strip()[:512]
+        r.leaf_category_norm_id = norm_by_label.get(lbl) if lbl else None
+        r.price_value = _effective_list_price_value(
+            r.coupon_price, r.price, r.original_price
+        )
+        chunk.append(r)
+        if len(chunk) >= 400:
+            JdJobSearchRow.objects.bulk_update(
+                chunk, ["leaf_category_norm_id", "price_value"]
+            )
+            chunk.clear()
+    if chunk:
+        JdJobSearchRow.objects.bulk_update(
+            chunk, ["leaf_category_norm_id", "price_value"]
+        )
+
+    chunk = []
+    for r in JdJobMergedRow.objects.all().iterator(chunk_size=400):
+        lbl = (r.leaf_category or "").strip()[:512]
+        r.leaf_category_norm_id = norm_by_label.get(lbl) if lbl else None
+        r.price_value = _effective_list_price_value(
+            r.coupon_price, r.price, r.original_price
+        )
+        chunk.append(r)
+        if len(chunk) >= 400:
+            JdJobMergedRow.objects.bulk_update(
+                chunk, ["leaf_category_norm_id", "price_value"]
+            )
+            chunk.clear()
+    if chunk:
+        JdJobMergedRow.objects.bulk_update(
+            chunk, ["leaf_category_norm_id", "price_value"]
+        )
+
+    chunk = []
+    for r in JdJobDetailRow.objects.all().iterator(chunk_size=400):
+        r.detail_price_value = _float_price_from_cell(r.detail_price_final)
+        chunk.append(r)
+        if len(chunk) >= 400:
+            JdJobDetailRow.objects.bulk_update(chunk, ["detail_price_value"])
+            chunk.clear()
+    if chunk:
+        JdJobDetailRow.objects.bulk_update(chunk, ["detail_price_value"])
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('pipeline', '0016_buyer_offer_csv_columns'),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name='JdLeafCategoryNorm',
+            fields=[
+                ('id', models.BigAutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('label', models.CharField(db_index=True, max_length=512, unique=True)),
+            ],
+            options={
+                'ordering': ['label'],
+            },
+        ),
+        migrations.AddField(
+            model_name='jdjobdetailrow',
+            name='detail_price_value',
+            field=models.FloatField(blank=True, db_index=True, null=True),
+        ),
+        migrations.AddField(
+            model_name='jdjobmergedrow',
+            name='price_value',
+            field=models.FloatField(blank=True, db_index=True, null=True),
+        ),
+        migrations.AddField(
+            model_name='jdjobsearchrow',
+            name='price_value',
+            field=models.FloatField(blank=True, db_index=True, null=True),
+        ),
+        migrations.AddIndex(
+            model_name='jdjobdetailrow',
+            index=models.Index(fields=['job', 'detail_price_value'], name='pipeline_jd_job_id_528890_idx'),
+        ),
+        migrations.AddField(
+            model_name='jdjobmergedrow',
+            name='leaf_category_norm',
+            field=models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.SET_NULL, related_name='merged_rows', to='pipeline.jdleafcategorynorm'),
+        ),
+        migrations.AddField(
+            model_name='jdjobsearchrow',
+            name='leaf_category_norm',
+            field=models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.SET_NULL, related_name='search_rows', to='pipeline.jdleafcategorynorm'),
+        ),
+        migrations.AddIndex(
+            model_name='jdjobmergedrow',
+            index=models.Index(fields=['job', 'leaf_category_norm'], name='pipeline_jd_job_id_fe7b92_idx'),
+        ),
+        migrations.AddIndex(
+            model_name='jdjobmergedrow',
+            index=models.Index(fields=['job', 'price_value'], name='pipeline_jd_job_id_49f026_idx'),
+        ),
+        migrations.AddIndex(
+            model_name='jdjobsearchrow',
+            index=models.Index(fields=['job', 'leaf_category_norm'], name='pipeline_jd_job_id_d77fae_idx'),
+        ),
+        migrations.AddIndex(
+            model_name='jdjobsearchrow',
+            index=models.Index(fields=['job', 'price_value'], name='pipeline_jd_job_id_b76707_idx'),
+        ),
+        migrations.RunPython(backfill_dataset_browse_fields, migrations.RunPython.noop),
+    ]

backend/pipeline/models.py

@@ -142,6 +142,18 @@ class JdProductSnapshot(models.Model):
         return f"{self.product_id} @ job {self.job_id}"
 
 
+class JdLeafCategoryNorm(models.Model):
+    """叶类目归一：与导出 ``leaf_category`` 原文一致（去空格），用于任务内类目筛选索引。"""
+
+    label = models.CharField(max_length=512, unique=True, db_index=True)
+
+    class Meta:
+        ordering = ["label"]
+
+    def __str__(self) -> str:
+        return self.label[:80]
+
+
 class JdJobSearchRow(models.Model):
     """单次任务下 PC 搜索导出表一行，字段与 ``pc_search_export.csv`` 列一一对应（内部英文属性名）。"""
 
@@ -171,6 +183,14 @@ class JdJobSearchRow(models.Model):
     seckill_info = models.TextField(blank=True, default="")
     attributes = models.TextField(blank=True, default="")
     leaf_category = models.TextField(blank=True, default="")
+    leaf_category_norm = models.ForeignKey(
+        JdLeafCategoryNorm,
+        null=True,
+        blank=True,
+        on_delete=models.SET_NULL,
+        related_name="search_rows",
+    )
+    price_value = models.FloatField(null=True, blank=True, db_index=True)
     platform = models.TextField(blank=True, default="")
     keyword = models.TextField(blank=True, default="")
     page = models.TextField(blank=True, default="")
@@ -185,6 +205,8 @@ class JdJobSearchRow(models.Model):
         ]
         indexes = [
             models.Index(fields=["job", "sku_id"]),
+            models.Index(fields=["job", "leaf_category_norm"]),
+            models.Index(fields=["job", "price_value"]),
         ]
 
     def __str__(self) -> str:
@@ -209,6 +231,7 @@ class JdJobDetailRow(models.Model):
     detail_body_ingredients = models.TextField(blank=True, default="")
     buyer_ranking_line = models.TextField(blank=True, default="")
     buyer_promo_text = models.TextField(blank=True, default="")
+    detail_price_value = models.FloatField(null=True, blank=True, db_index=True)
 
     class Meta:
         ordering = ["row_index"]
@@ -220,6 +243,7 @@ class JdJobDetailRow(models.Model):
         ]
         indexes = [
             models.Index(fields=["job", "sku_id"]),
+            models.Index(fields=["job", "detail_price_value"]),
         ]
 
     def __str__(self) -> str:
@@ -286,6 +310,14 @@ class JdJobMergedRow(models.Model):
     image = models.TextField(blank=True, default="")
     attributes = models.TextField(blank=True, default="")
     leaf_category = models.TextField(blank=True, default="")
+    leaf_category_norm = models.ForeignKey(
+        JdLeafCategoryNorm,
+        null=True,
+        blank=True,
+        on_delete=models.SET_NULL,
+        related_name="merged_rows",
+    )
+    price_value = models.FloatField(null=True, blank=True, db_index=True)
     keyword = models.TextField(blank=True, default="")
     page = models.TextField(blank=True, default="")
     detail_brand = models.TextField(blank=True, default="")
@@ -309,6 +341,8 @@ class JdJobMergedRow(models.Model):
         ]
         indexes = [
             models.Index(fields=["job", "sku_id"]),
+            models.Index(fields=["job", "leaf_category_norm"]),
+            models.Index(fields=["job", "price_value"]),
         ]
 
     def __str__(self) -> str:

backend/pipeline/price_parse.py

@@ -0,0 +1,29 @@
+"""从爬虫导出单元格解析数值价格（供入库索引与数据集筛选）。"""
+from __future__ import annotations
+
+import re
+
+
+def float_price_from_cell(s: str | None) -> float | None:
+    t = (s or "").strip().replace(",", "").replace("，", "")
+    if not t:
+        return None
+    m = re.search(r"(\d+(?:\.\d+)?)", t)
+    if not m:
+        return None
+    try:
+        v = float(m.group(1))
+    except ValueError:
+        return None
+    if 0 < v < 1_000_000:
+        return v
+    return None
+
+
+def effective_list_price_value(coupon: str, price: str, original: str) -> float | None:
+    """优先券后价，其次标价，再次原价（与列表侧展示习惯一致）。"""
+    for s in (coupon, price, original):
+        v = float_price_from_cell(s)
+        if v is not None:
+            return v
+    return None

backend/pipeline/views.py

@@ -17,6 +17,22 @@ from rest_framework import status
 from rest_framework.response import Response
 from rest_framework.views import APIView
 
+from .dataset_api import (
+    DETAIL_SORT_FIELDS,
+    MERGED_SORT_FIELDS,
+    SEARCH_SORT_FIELDS,
+    apply_detail_filters,
+    apply_detail_order,
+    apply_merged_filters,
+    apply_merged_order,
+    apply_search_filters,
+    apply_search_order,
+    category_norm_id_from_request,
+    detail_category_q_from_request,
+    filter_echo,
+    parse_sort_meta,
+    price_bounds_from_request,
+)
 from .dataset_nonempty import (
     comment_columns_for_api,
     detail_columns_for_api,
@@ -50,6 +66,7 @@ from .models import (
     JdJobDetailRow,
     JdJobMergedRow,
     JdJobSearchRow,
+    JdLeafCategoryNorm,
     JdProduct,
     JdProductSnapshot,
     JobStatus,
@@ -739,6 +756,29 @@ def _read_page_params(request) -> tuple[int, int]:
     return page, page_size
 
 
+def _category_norm_options_for_job(job: PipelineJob, RowModel: type) -> list[dict[str, Any]]:
+    ids = (
+        RowModel.objects.filter(job=job, leaf_category_norm_id__isnull=False)
+        .values_list("leaf_category_norm_id", flat=True)
+        .distinct()
+    )
+    return list(
+        JdLeafCategoryNorm.objects.filter(id__in=ids)
+        .order_by("label")
+        .values("id", "label")
+    )
+
+
+def _detail_category_path_options(job: PipelineJob) -> list[str]:
+    return list(
+        JdJobDetailRow.objects.filter(job=job)
+        .exclude(detail_category_path="")
+        .values_list("detail_category_path", flat=True)
+        .distinct()
+        .order_by("detail_category_path")[:400]
+    )
+
+
 class JobDatasetSummaryView(APIView):
     """任务在库中的搜索/详情/评价行数（入库后可用）。"""
 
@@ -757,6 +797,19 @@ class JobDatasetSummaryView(APIView):
                 "detail_columns": detail_columns_for_api(job),
                 "comment_columns": comment_columns_for_api(job),
                 "merged_columns": merged_columns_for_api(job),
+                "search_category_options": _category_norm_options_for_job(
+                    job, JdJobSearchRow
+                ),
+                "merged_category_options": _category_norm_options_for_job(
+                    job, JdJobMergedRow
+                ),
+                "detail_category_path_options": _detail_category_path_options(job),
+                "dataset_sort_help": {
+                    "search": sorted(SEARCH_SORT_FIELDS),
+                    "detail": sorted(DETAIL_SORT_FIELDS),
+                    "merged": sorted(MERGED_SORT_FIELDS),
+                    "comments": ["row_index"],
+                },
             }
         )
 
@@ -765,15 +818,30 @@ class JobDatasetSearchView(APIView):
     def get(self, request, pk: int):
         job = _dataset_job(pk)
         page, page_size = _read_page_params(request)
+        sort, desc = parse_sort_meta(request)
+        sort_eff = sort if sort in SEARCH_SORT_FIELDS else "row_index"
+        cid = category_norm_id_from_request(request)
+        pmin, pmax = price_bounds_from_request(request)
+        dcq = detail_category_q_from_request(request)
         qs = JdJobSearchRow.objects.filter(job=job)
+        qs = apply_search_filters(qs, request)
+        qs = apply_search_order(qs, sort_eff, desc)
         total = qs.count()
         start = (page - 1) * page_size
-        rows = qs.order_by("row_index")[start : start + page_size]
+        rows = qs[start : start + page_size]
         return Response(
             {
                 "total": total,
                 "page": page,
                 "page_size": page_size,
+                "filters": filter_echo(
+                    category_norm_id=cid,
+                    price_min=pmin,
+                    price_max=pmax,
+                    detail_category_q=dcq,
+                    sort=sort_eff,
+                    desc=desc,
+                ),
                 "results": [search_row_to_dict(r) for r in rows],
             }
         )
@@ -783,15 +851,30 @@ class JobDatasetDetailView(APIView):
     def get(self, request, pk: int):
         job = _dataset_job(pk)
         page, page_size = _read_page_params(request)
+        sort, desc = parse_sort_meta(request)
+        sort_eff = sort if sort in DETAIL_SORT_FIELDS else "row_index"
+        cid = category_norm_id_from_request(request)
+        pmin, pmax = price_bounds_from_request(request)
+        dcq = detail_category_q_from_request(request)
         qs = JdJobDetailRow.objects.filter(job=job)
+        qs = apply_detail_filters(qs, request)
+        qs = apply_detail_order(qs, sort_eff, desc)
         total = qs.count()
         start = (page - 1) * page_size
-        rows = qs.order_by("row_index")[start : start + page_size]
+        rows = qs[start : start + page_size]
         return Response(
             {
                 "total": total,
                 "page": page,
                 "page_size": page_size,
+                "filters": filter_echo(
+                    category_norm_id=cid,
+                    price_min=pmin,
+                    price_max=pmax,
+                    detail_category_q=dcq,
+                    sort=sort_eff,
+                    desc=desc,
+                ),
                 "results": [detail_row_to_dict(r) for r in rows],
             }
         )
@@ -823,15 +906,30 @@ class JobDatasetMergedView(APIView):
     def get(self, request, pk: int):
         job = _dataset_job(pk)
         page, page_size = _read_page_params(request)
+        sort, desc = parse_sort_meta(request)
+        sort_eff = sort if sort in MERGED_SORT_FIELDS else "row_index"
+        cid = category_norm_id_from_request(request)
+        pmin, pmax = price_bounds_from_request(request)
+        dcq = detail_category_q_from_request(request)
         qs = JdJobMergedRow.objects.filter(job=job)
+        qs = apply_merged_filters(qs, request)
+        qs = apply_merged_order(qs, sort_eff, desc)
         total = qs.count()
         start = (page - 1) * page_size
-        rows = qs.order_by("row_index")[start : start + page_size]
+        rows = qs[start : start + page_size]
         return Response(
             {
                 "total": total,
                 "page": page,
                 "page_size": page_size,
+                "filters": filter_echo(
+                    category_norm_id=cid,
+                    price_min=pmin,
+                    price_max=pmax,
+                    detail_category_q=dcq,
+                    sort=sort_eff,
+                    desc=desc,
+                ),
                 "results": [merged_row_to_dict(r) for r in rows],
             }
         )

frontend/src/components/JobDatasetModal.vue

@@ -18,14 +18,31 @@ const emit = defineEmits(['close'])
 
 const paneActive = computed(() => !!(props.job && (props.embedded || props.open)))
 
+const SORT_LABELS = {
+  row_index: '入库顺序',
+  price: '价格',
+  sku_id: 'SKU',
+  title: '标题',
+  leaf_category: '叶类目',
+  detail_category_path: '类目路径',
+  detail_brand: '品牌',
+}
+
 const tab = ref('search')
 const page = ref(1)
 const pageSize = ref(30)
+const pageJumpDraft = ref(1)
 const summary = ref(null)
 const list = ref({ results: [], total: 0, page: 1, page_size: 30 })
 const loading = ref(false)
 const err = ref('')
 const commentSkuFilter = ref('')
+const sortField = ref('row_index')
+const sortOrder = ref('asc')
+const categoryNormId = ref('')
+const priceMin = ref('')
+const priceMax = ref('')
+const detailCategoryQ = ref('')
 const exportPanelOpen = ref(false)
 const exportLoading = ref(false)
 const exportErr = ref('')
@@ -39,6 +56,25 @@ function handleBackdrop(e) {
   onBackdrop(e)
 }
 
+const sortOptions = computed(() => {
+  const h = summary.value?.dataset_sort_help
+  let keys = ['row_index']
+  if (h) {
+    if (tab.value === 'search') keys = h.search?.length ? h.search : keys
+    else if (tab.value === 'detail') keys = h.detail?.length ? h.detail : keys
+    else if (tab.value === 'merged') keys = h.merged?.length ? h.merged : keys
+    else keys = h.comments?.length ? h.comments : ['row_index']
+  }
+  return keys.map((k) => ({ value: k, label: SORT_LABELS[k] || k }))
+})
+
+const categoryNormOptions = computed(() => {
+  if (!summary.value) return []
+  if (tab.value === 'search') return summary.value.search_category_options || []
+  if (tab.value === 'merged') return summary.value.merged_category_options || []
+  return []
+})
+
 const displayColumns = computed(() => {
   const s = summary.value
   let cols = []
@@ -122,14 +158,31 @@ async function refreshList() {
   err.value = ''
   try {
     await loadSummary()
-    const sku = tab.value === 'comments' ? commentSkuFilter.value.trim() : ''
+    const opts =
-    const url = jobDatasetPageUrl(props.job.id, tab.value, page.value, pageSize.value, sku)
+      tab.value === 'comments'
+        ? { skuId: commentSkuFilter.value.trim() }
+        : {
+            sort: sortField.value,
+            order: sortOrder.value,
+            categoryNormId: categoryNormId.value.trim(),
+            priceMin: priceMin.value,
+            priceMax: priceMax.value,
+            detailCategoryQ: detailCategoryQ.value.trim(),
+          }
+    const url = jobDatasetPageUrl(
+      props.job.id,
+      tab.value,
+      page.value,
+      pageSize.value,
+      opts,
+    )
     const r = await api(url)
     if (!r.ok) {
       err.value = await r.text()
       return
     }
     list.value = await r.json()
+    pageJumpDraft.value = list.value.page || page.value
   } catch (e) {
     err.value = String(e)
   } finally {
@@ -143,6 +196,12 @@ watch(
     if (paneActive.value) {
       tab.value = 'search'
       page.value = 1
+      sortField.value = 'row_index'
+      sortOrder.value = 'asc'
+      categoryNormId.value = ''
+      priceMin.value = ''
+      priceMax.value = ''
+      detailCategoryQ.value = ''
       commentSkuFilter.value = ''
       err.value = ''
       summary.value = null
@@ -155,11 +214,46 @@ watch(
 watch(tab, () => {
   page.value = 1
   exportPanelOpen.value = false
+  sortField.value = 'row_index'
+  sortOrder.value = 'asc'
+  categoryNormId.value = ''
+  priceMin.value = ''
+  priceMax.value = ''
+  detailCategoryQ.value = ''
 })
 
-watch([paneActive, () => props.job?.id, tab, page, commentSkuFilter], () => {
+watch(
-  if (paneActive.value && props.job) refreshList()
+  [
-})
+    sortField,
+    sortOrder,
+    categoryNormId,
+    priceMin,
+    priceMax,
+    detailCategoryQ,
+  ],
+  () => {
+    if (paneActive.value && props.job && tab.value !== 'comments') page.value = 1
+  },
+)
+
+watch(
+  [
+    paneActive,
+    () => props.job?.id,
+    tab,
+    page,
+    commentSkuFilter,
+    sortField,
+    sortOrder,
+    categoryNormId,
+    priceMin,
+    priceMax,
+    detailCategoryQ,
+  ],
+  () => {
+    if (paneActive.value && props.job) refreshList()
+  },
+)
 
 const totalPages = () => {
   const t = list.value.total || 0
@@ -175,6 +269,14 @@ function nextPage() {
   if (page.value < totalPages()) page.value += 1
 }
 
+function goToPage() {
+  const tp = totalPages()
+  let n = Math.round(Number(pageJumpDraft.value))
+  if (!Number.isFinite(n)) n = 1
+  page.value = Math.min(Math.max(1, n), tp)
+  pageJumpDraft.value = page.value
+}
+
 const exportPanelTitle = computed(() => {
   const m = { search: '搜索', detail: '商详', comments: '评论', merged: '整合宽表' }
   return `当前表（${m[tab.value] || tab.value}）`
@@ -314,6 +416,60 @@ async function runExport(format) {
           </template>
         </div>
 
+        <div v-if="tab !== 'comments'" class="toolbar-filters">
+          <label class="filter-item">
+            排序
+            <select v-model="sortField" class="filter-select">
+              <option v-for="o in sortOptions" :key="o.value" :value="o.value">{{ o.label }}</option>
+            </select>
+          </label>
+          <label class="filter-item">
+            顺序
+            <select v-model="sortOrder" class="filter-select">
+              <option value="asc">升序</option>
+              <option value="desc">降序</option>
+            </select>
+          </label>
+          <template v-if="tab === 'search' || tab === 'merged'">
+            <label class="filter-item">
+              叶类目
+              <select v-model="categoryNormId" class="filter-select wide">
+                <option value="">全部</option>
+                <option v-for="c in categoryNormOptions" :key="c.id" :value="String(c.id)">
+                  {{ c.label }}
+                </option>
+              </select>
+            </label>
+          </template>
+          <template v-if="tab === 'detail' || tab === 'merged'">
+            <label class="filter-item">
+              类目路径包含
+              <input
+                v-model="detailCategoryQ"
+                type="search"
+                class="filter-input wide"
+                placeholder="模糊匹配商详类目路径"
+                list="detail-cat-dl"
+              />
+              <datalist id="detail-cat-dl">
+                <option
+                  v-for="p in summary?.detail_category_path_options || []"
+                  :key="p"
+                  :value="p"
+                />
+              </datalist>
+            </label>
+          </template>
+          <label class="filter-item">
+            价格 ≥
+            <input v-model="priceMin" type="number" step="any" class="filter-input narrow" placeholder="最低" />
+          </label>
+          <label class="filter-item">
+            价格 ≤
+            <input v-model="priceMax" type="number" step="any" class="filter-input narrow" placeholder="最高" />
+          </label>
+        </div>
+
         <div class="table-block">
           <div v-if="loading" class="state state-fill">加载中…</div>
           <p v-else-if="err" class="state err state-fill">{{ err }}</p>
@@ -371,6 +527,20 @@ async function runExport(format) {
           <span class="ma-muted"
             >第 {{ list.page || page }} / {{ totalPages() }} 页 · 共 {{ list.total ?? 0 }} 条</span
           >
+          <span class="pager-jump">
+            <label class="jump-label"
+              >跳转
+              <input
+                v-model.number="pageJumpDraft"
+                type="number"
+                :min="1"
+                :max="totalPages()"
+                class="jump-input"
+              />
+              页</label
+            >
+            <button type="button" class="ma-btn ma-btn-secondary" @click="goToPage">确定</button>
+          </span>
           <button type="button" class="ma-btn ma-btn-secondary" :disabled="page <= 1" @click="prevPage">
             上一页
           </button>
@@ -563,6 +733,40 @@ async function runExport(format) {
   width: 160px;
   font: inherit;
 }
+.toolbar-filters {
+  display: flex;
+  flex-wrap: wrap;
+  align-items: flex-end;
+  gap: 0.65rem 0.85rem;
+  padding: 0.55rem 1rem 0.65rem;
+  border-bottom: 1px solid #f1f5f9;
+  background: #fafafa;
+  flex-shrink: 0;
+}
+.filter-item {
+  display: flex;
+  flex-direction: column;
+  gap: 0.2rem;
+  font-size: 0.72rem;
+  color: #475569;
+}
+.filter-select,
+.filter-input {
+  font: inherit;
+  font-size: 0.8rem;
+  padding: 0.3rem 0.45rem;
+  border: 1px solid #d1d5db;
+  border-radius: 6px;
+  min-width: 0;
+}
+.filter-select.wide,
+.filter-input.wide {
+  min-width: 12rem;
+  max-width: 22rem;
+}
+.filter-input.narrow {
+  width: 5.5rem;
+}
 /* 高度封顶：数据再长也在表格内滚动，不把整块卡片无限撑高 */
 .table-block {
   flex: 1 1 auto;
@@ -689,6 +893,27 @@ async function runExport(format) {
   border-top: 1px solid #e5e7eb;
   flex-shrink: 0;
 }
+.pager-jump {
+  display: inline-flex;
+  flex-wrap: wrap;
+  align-items: center;
+  gap: 0.35rem;
+}
+.jump-label {
+  font-size: 0.78rem;
+  color: #475569;
+  display: inline-flex;
+  align-items: center;
+  gap: 0.25rem;
+}
+.jump-input {
+  width: 3.5rem;
+  font: inherit;
+  font-size: 0.8rem;
+  padding: 0.25rem 0.35rem;
+  border: 1px solid #d1d5db;
+  border-radius: 6px;
+}
 .ma-muted {
   color: #64748b;
 }

frontend/src/composables/useJobs.js

@@ -192,12 +192,31 @@ export function reportConfigDefaultsUrl() {
   return '/api/report-config-defaults/'
 }
 
-export function jobDatasetPageUrl(jobId, kind, page = 1, pageSize = 50, skuId = '') {
+/**
+ * @param {Record<string, string | number | undefined> | string} [opts] 筛选参数对象；兼容旧调用：传入字符串视为 comments 的 sku_id
+ */
+export function jobDatasetPageUrl(jobId, kind, page = 1, pageSize = 50, opts = {}) {
+  const o = typeof opts === 'string' ? { skuId: opts } : opts || {}
   const p = new URLSearchParams({
     page: String(page),
     page_size: String(pageSize),
   })
-  if (skuId) p.set('sku_id', skuId)
+  const sku = o.skuId ?? o.sku_id
+  if (sku) p.set('sku_id', String(sku))
+  if (o.sort) p.set('sort', String(o.sort))
+  if (o.order) p.set('order', String(o.order))
+  const cn = o.categoryNormId ?? o.category_norm_id
+  if (cn !== undefined && cn !== null && String(cn).trim() !== '')
+    p.set('category_norm_id', String(cn).trim())
+  const pmin = o.priceMin ?? o.price_min
+  if (pmin !== undefined && pmin !== null && String(pmin).trim() !== '')
+    p.set('price_min', String(pmin).trim())
+  const pmax = o.priceMax ?? o.price_max
+  if (pmax !== undefined && pmax !== null && String(pmax).trim() !== '')
+    p.set('price_max', String(pmax).trim())
+  const dcq = o.detailCategoryQ ?? o.detail_category_q
+  if (dcq !== undefined && dcq !== null && String(dcq).trim() !== '')
+    p.set('detail_category_q', String(dcq).trim())
   return `/api/jobs/${jobId}/dataset/${kind}/?${p.toString()}`
 }