Initial commit: TradeMate 外贸小助手 MVP

项目结构: - backend/ Python FastAPI 后端 - uni-app/ uni-app跨端前端 - docs/ 设计文档 - docker-compose.yml Docker编排 - nginx/scripts/systemd 运维配置已完成功能: - 用户认证 (JWT) - 智能翻译 + 回复建议 - 营销素材生成 - 客户管理 + 沉默检测 - 报价单管理 - 产品库管理 - 汇率换算 - 推送通知 (uni-push) - WhatsApp Webhook框架 - Celery定时任务
2026-05-08 18:17:12 +08:00
commit c6206787da
121 changed files with 11743 additions and 0 deletions
@@ -0,0 +1,115 @@
+from typing import Dict, Any, Optional, List
+from app.ai.router import get_ai_router
+from app.ai.trade_corpus import TradeCorpus
+import logging
+
+logger = logging.getLogger(__name__)
+
+
+class TranslationService:
+    def __init__(self):
+        self.ai = get_ai_router()
+        self.corpus = TradeCorpus()
+
+    async def translate(
+        self, text: str, target_lang: str, source_lang: Optional[str] = None,
+        context: Optional[str] = None, user_id: Optional[str] = None,
+    ) -> Dict[str, Any]:
+        similar = await self.corpus.find_similar(text, "translate")
+        if similar:
+            best = similar[0]
+            if len(best["source"]) > 20 and self._similarity_ratio(text, best["source"]) > 0.85:
+                return {
+                    "translated_text": best["target"],
+                    "source_lang": source_lang or "auto",
+                    "provider_used": "corpus_cache",
+                    "from_cache": True,
+                }
+
+        result = await self.ai.translate(text, target_lang, source_lang, context)
+        translated = result.get("translated_text", "")
+        provider = result.get("provider_used", "unknown")
+
+        await self.corpus.record(
+            source_text=text,
+            target_text=translated,
+            task_type="translate",
+            provider=provider,
+            source_lang=source_lang,
+            target_lang=target_lang,
+            metadata={"user_id": user_id} if user_id else None,
+        )
+
+        result["source_lang"] = result.get("detected_source_lang", source_lang or "auto")
+        result["from_cache"] = False
+        return result
+
+    async def generate_reply(
+        self, inquiry: str, context: Optional[Dict[str, Any]] = None,
+        tone: str = "professional", count: int = 3,
+    ) -> List[Dict[str, Any]]:
+        similar = await self.corpus.find_similar(inquiry, "reply")
+        if similar and count > 1:
+            pass
+
+        results = []
+        tones = self._get_tones(tone, count)
+
+        for t in tones:
+            try:
+                result = await self.ai.reply(inquiry, context, t)
+                results.append({
+                    "reply": result.get("reply", ""),
+                    "tone": t,
+                    "provider": result.get("provider_used", "unknown"),
+                })
+            except Exception as e:
+                logger.warning(f"Reply generation failed for tone '{t}': {e}")
+                results.append({"reply": "", "tone": t, "error": str(e)})
+
+        return results
+
+    async def extract_info(self, text: str, extract_type: str = "auto") -> Dict[str, Any]:
+        schemas = {
+            "product": {
+                "type": "object",
+                "properties": {
+                    "product_name": {"type": "string"},
+                    "quantity": {"type": "string"},
+                    "price": {"type": "string"},
+                    "currency": {"type": "string"},
+                    "delivery_terms": {"type": "string"},
+                    "target_country": {"type": "string"},
+                },
+            },
+            "inquiry": {
+                "type": "object",
+                "properties": {
+                    "intent": {"type": "string"},
+                    "product_interest": {"type": "string"},
+                    "quantity": {"type": "string"},
+                    "budget": {"type": "string"},
+                    "urgency": {"type": "string"},
+                    "contact_info": {"type": "string"},
+                },
+            },
+        }
+
+        schema = schemas.get(extract_type, schemas["inquiry"])
+        result = await self.ai.extract(text, schema)
+        return result.get("data", {})
+
+    def _get_tones(self, base: str, count: int) -> List[str]:
+        tones = ["professional", "friendly", "formal"]
+        if base in tones:
+            tones.remove(base)
+            tones.insert(0, base)
+        return tones[:count]
+
+    def _similarity_ratio(self, a: str, b: str) -> float:
+        if not a or not b:
+            return 0.0
+        set_a, set_b = set(a.lower().split()), set(b.lower().split())
+        if not set_a or not set_b:
+            return 0.0
+        return len(set_a & set_b) / len(set_a | set_b)