feat: 以图搜图返回 business_data 包含 name/spec_model/url，支持详情页跳转
This commit is contained in:
@ -1,6 +1,7 @@
|
||||
# -*- coding: utf-8 -*-
|
||||
"""
|
||||
以图搜图 API - CLIP Vision Embedding + pgvector 余弦距离检索
|
||||
数据源：image_embeddings 表（统一向量存储）
|
||||
"""
|
||||
|
||||
import os
|
||||
@ -10,6 +11,10 @@ from flask import Blueprint, request, jsonify
|
||||
from sqlalchemy import text
|
||||
from app.extensions import db
|
||||
from app.utils.ai_vision import load_clip_model, get_image_embedding
|
||||
from app.models.inbound.buy import StockBuy
|
||||
from app.models.inbound.semi import StockSemi
|
||||
from app.models.inbound.product import StockProduct
|
||||
from app.models.base import MaterialBase
|
||||
|
||||
# 注册蓝图
|
||||
image_search_bp = Blueprint('image_search', __name__)
|
||||
@ -71,88 +76,158 @@ def image_search():
|
||||
print(f"⚠️ [ImageSearch] 临时文件删除失败: {e}")
|
||||
|
||||
# ---------------------------------------------------------
|
||||
# 5. pgvector 余弦相似度检索（跨表联合检索）
|
||||
# 5. pgvector 余弦相似度检索（统一查 image_embeddings 表）
|
||||
# ---------------------------------------------------------
|
||||
try:
|
||||
query_vector_str = '[' + ','.join(str(v) for v in embedding) + ']'
|
||||
|
||||
sql = text("""
|
||||
SELECT id, name, spec_model, image_url,
|
||||
(1 - (vec <=> :query_vector)) AS similarity
|
||||
FROM (
|
||||
-- 1. 基础物料表
|
||||
SELECT id, name, spec_model, product_image AS image_url, img_embedding AS vec
|
||||
FROM material_base
|
||||
WHERE img_embedding IS NOT NULL
|
||||
|
||||
UNION ALL
|
||||
|
||||
-- 2. 采购入库表 (通过 base_id 关联拿真实物料信息)
|
||||
SELECT mb.id, mb.name, mb.spec_model, sb.arrival_photo AS image_url, sb.arrival_image_embedding AS vec
|
||||
FROM stock_buy sb
|
||||
JOIN material_base mb ON sb.base_id = mb.id
|
||||
WHERE sb.arrival_image_embedding IS NOT NULL
|
||||
|
||||
UNION ALL
|
||||
|
||||
-- 3. 半成品入库表 (通过 base_id 关联拿真实物料信息)
|
||||
SELECT mb.id, mb.name, mb.spec_model, ss.arrival_photo AS image_url, ss.arrival_image_embedding AS vec
|
||||
FROM stock_semi ss
|
||||
JOIN material_base mb ON ss.base_id = mb.id
|
||||
WHERE ss.arrival_image_embedding IS NOT NULL
|
||||
|
||||
UNION ALL
|
||||
|
||||
-- 4. 成品入库表 (通过 base_id 关联拿真实物料信息)
|
||||
SELECT mb.id, mb.name, mb.spec_model, sp.product_photo AS image_url, sp.arrival_image_embedding AS vec
|
||||
FROM stock_product sp
|
||||
JOIN material_base mb ON sp.base_id = mb.id
|
||||
WHERE sp.arrival_image_embedding IS NOT NULL
|
||||
) AS combined
|
||||
|
||||
-- 核心：计算余弦距离并排序，取最接近的前 50 个！
|
||||
ORDER BY vec <=> :query_vector LIMIT 50
|
||||
SELECT
|
||||
ie.id AS embedding_id,
|
||||
ie.module_name,
|
||||
ie.target_id,
|
||||
ie.image_url,
|
||||
(1 - (ie.embedding <=> :query_vector)) AS similarity
|
||||
FROM image_embeddings ie
|
||||
WHERE ie.embedding IS NOT NULL
|
||||
ORDER BY ie.embedding <=> :query_vector
|
||||
LIMIT 200
|
||||
""")
|
||||
|
||||
# 执行查询
|
||||
records = db.session.execute(sql, {"query_vector": query_vector_str}).fetchall()
|
||||
raw_records = db.session.execute(sql, {"query_vector": query_vector_str}).fetchall()
|
||||
if not raw_records:
|
||||
return jsonify({"code": 200, "data": []})
|
||||
|
||||
# 按 (module_name, target_id) 去重，每业务记录只保留最相似的那张图
|
||||
seen = {}
|
||||
for row in raw_records:
|
||||
key = (row.module_name, row.target_id)
|
||||
if key not in seen:
|
||||
seen[key] = row
|
||||
|
||||
# 批量回填业务数据
|
||||
target_ids_by_module = {}
|
||||
for row in seen.values():
|
||||
target_ids_by_module.setdefault(row.module_name, []).append(row.target_id)
|
||||
|
||||
business_map = {}
|
||||
|
||||
# 回填 StockBuy
|
||||
if 'stock_buy' in target_ids_by_module:
|
||||
ids = target_ids_by_module['stock_buy']
|
||||
records = (
|
||||
db.session.query(StockBuy)
|
||||
.filter(StockBuy.id.in_(ids))
|
||||
.outerjoin(MaterialBase, StockBuy.base_id == MaterialBase.id)
|
||||
.all()
|
||||
)
|
||||
for r in records:
|
||||
business_map[('stock_buy', r.id)] = {
|
||||
'record_id': r.id,
|
||||
'name': r.base.name if r.base else None,
|
||||
'spec_model': r.base.spec_model if r.base else None,
|
||||
'sku': r.sku,
|
||||
'barcode': r.barcode,
|
||||
'serial_number': r.serial_number,
|
||||
'batch_number': r.batch_number,
|
||||
'status': r.status,
|
||||
'warehouse_location': r.warehouse_location,
|
||||
'stock_quantity': r.stock_quantity,
|
||||
'module_name': 'stock_buy',
|
||||
'url': '/inventory/buy',
|
||||
}
|
||||
|
||||
# 回填 StockSemi
|
||||
if 'stock_semi' in target_ids_by_module:
|
||||
ids = target_ids_by_module['stock_semi']
|
||||
records = (
|
||||
db.session.query(StockSemi)
|
||||
.filter(StockSemi.id.in_(ids))
|
||||
.outerjoin(MaterialBase, StockSemi.base_id == MaterialBase.id)
|
||||
.all()
|
||||
)
|
||||
for r in records:
|
||||
business_map[('stock_semi', r.id)] = {
|
||||
'record_id': r.id,
|
||||
'name': r.base.name if r.base else None,
|
||||
'spec_model': r.base.spec_model if r.base else None,
|
||||
'sku': r.sku,
|
||||
'barcode': r.barcode,
|
||||
'serial_number': r.serial_number,
|
||||
'batch_number': r.batch_number,
|
||||
'status': r.status,
|
||||
'warehouse_location': r.warehouse_location,
|
||||
'stock_quantity': r.stock_quantity,
|
||||
'module_name': 'stock_semi',
|
||||
'url': '/inventory/semi',
|
||||
}
|
||||
|
||||
# 回填 StockProduct
|
||||
if 'stock_product' in target_ids_by_module:
|
||||
ids = target_ids_by_module['stock_product']
|
||||
records = (
|
||||
db.session.query(StockProduct)
|
||||
.filter(StockProduct.id.in_(ids))
|
||||
.outerjoin(MaterialBase, StockProduct.base_id == MaterialBase.id)
|
||||
.all()
|
||||
)
|
||||
for r in records:
|
||||
business_map[('stock_product', r.id)] = {
|
||||
'record_id': r.id,
|
||||
'name': r.base.name if r.base else None,
|
||||
'spec_model': r.base.spec_model if r.base else None,
|
||||
'sku': r.sku,
|
||||
'barcode': r.barcode,
|
||||
'serial_number': r.serial_number,
|
||||
'batch_number': r.batch_number,
|
||||
'status': r.status,
|
||||
'warehouse_location': r.warehouse_location,
|
||||
'stock_quantity': r.stock_quantity,
|
||||
'sale_price': r.sale_price,
|
||||
'module_name': 'stock_product',
|
||||
'url': '/inventory/product',
|
||||
}
|
||||
|
||||
# 回填 MaterialBase
|
||||
if 'material_base' in target_ids_by_module:
|
||||
ids = target_ids_by_module['material_base']
|
||||
records = MaterialBase.query.filter(MaterialBase.id.in_(ids)).all()
|
||||
for r in records:
|
||||
business_map[('material_base', r.id)] = {
|
||||
'record_id': r.id,
|
||||
'name': r.name,
|
||||
'spec_model': r.spec_model,
|
||||
'common_name': r.common_name,
|
||||
'category': r.category,
|
||||
'material_type': r.material_type,
|
||||
'unit': r.unit,
|
||||
'module_name': 'material_base',
|
||||
'url': '/material/index',
|
||||
}
|
||||
|
||||
# 组装最终返回
|
||||
results = []
|
||||
seen_product_ids = set() # 【新增】用来记录已经添加过的物料 ID
|
||||
|
||||
for row in records:
|
||||
# 【新增】如果这个物料已经在这个列表里了，直接跳过它
|
||||
if row.id in seen_product_ids:
|
||||
continue
|
||||
|
||||
# 记录这个物料 ID，保证下次不会再重复添加
|
||||
seen_product_ids.add(row.id)
|
||||
|
||||
# 1. 提取原始 URL
|
||||
raw_url = row.image_url
|
||||
clean_url = ""
|
||||
|
||||
if raw_url:
|
||||
if raw_url.startswith('[') and raw_url.endswith(']'):
|
||||
import json
|
||||
try:
|
||||
url_list = json.loads(raw_url)
|
||||
clean_url = url_list[0] if url_list else ""
|
||||
except:
|
||||
clean_url = raw_url
|
||||
else:
|
||||
clean_url = raw_url
|
||||
|
||||
# 2. 组装返回结果
|
||||
for row in seen.values():
|
||||
key = (row.module_name, row.target_id)
|
||||
biz = business_map.get(key, {})
|
||||
raw_url = row.image_url or ''
|
||||
clean_url = raw_url
|
||||
if raw_url.startswith('['):
|
||||
try:
|
||||
url_list = json.loads(raw_url)
|
||||
clean_url = url_list[0] if url_list else ''
|
||||
except:
|
||||
pass
|
||||
results.append({
|
||||
"product_id": row.id,
|
||||
"product_name": row.name,
|
||||
"spec_model": row.spec_model,
|
||||
"module_name": row.module_name,
|
||||
"target_id": row.target_id,
|
||||
"image_url": clean_url,
|
||||
"similarity": round(float(row.similarity), 4)
|
||||
"similarity": round(float(row.similarity), 4),
|
||||
"product_name": biz.get('name') or biz.get('material_name') or '未命名物料',
|
||||
"product_id": row.target_id,
|
||||
"spec_model": biz.get('spec_model') or '',
|
||||
"business_data": biz,
|
||||
})
|
||||
|
||||
# 修改后：只要凑够了 10 个完全不同的物料，就立刻结束循环
|
||||
if len(results) >= 10:
|
||||
break
|
||||
|
||||
|
||||
Reference in New Issue
Block a user