AI UPDATE

2026-03-22 03:08:27 +03:30
parent 3ee14ca977
commit d977a583c6
37 changed files with 3525 additions and 263 deletions
@@ -37,11 +37,11 @@ def get_embedding_client(config: RAGConfig | None = None) -> OpenAI:
 def get_chat_client(config: RAGConfig | None = None) -> OpenAI:
    """
    ساخت کلاینت OpenAI برای Chat/LLM بر اساس provider فعال.
-    provider از config.embedding.provider خوانده می‌شود (مشترک بین embedding و chat).
+    provider از config.llm.provider خوانده می‌شود.
    """
    cfg = config or load_rag_config()
    llm = cfg.llm
-    provider = cfg.embedding.provider
+    provider = llm.provider or cfg.embedding.provider

    
    logger.info(provider)
@@ -4,7 +4,7 @@
 import logging
 from pathlib import Path

-from .config import load_rag_config, RAGConfig
+from .config import load_rag_config, RAGConfig, get_service_config, ServiceConfig
 from .api_provider import get_chat_client
 from .retrieve import search_with_query
 from .user_data import build_user_soil_text, build_user_weather_text
@@ -43,6 +43,16 @@ def _load_kb_tone(kb_name: str, config: RAGConfig | None = None) -> str:
    return ""


+def _load_service_tone(service: ServiceConfig, config: RAGConfig | None = None) -> str:
+    cfg = config or load_rag_config()
+    if service.tone_file:
+        base = Path(__file__).resolve().parent.parent
+        tone_path = base / service.tone_file
+        if tone_path.exists():
+            return tone_path.read_text(encoding="utf-8").strip()
+    return _load_kb_tone(service.knowledge_base, cfg)
+
+
 def _detect_kb_intent(query: str) -> str:
    """تشخیص ساده نوع پایگاه دانش مورد نیاز از روی متن سوال."""
    q = query.lower()
@@ -59,10 +69,11 @@ def _detect_kb_intent(query: str) -> str:

 def build_rag_context(
    query: str,
-    sensor_uuid: str,
+    sensor_uuid: str | None = None,
    config: RAGConfig | None = None,
    limit: int = 8,
    kb_name: str | None = None,
+    service_id: str | None = None,
 ) -> str:
    """
    ساخت context برای LLM: دیتای فعلی خاک کاربر + متن‌های مرتبط از RAG.
@@ -76,24 +87,34 @@ def build_rag_context(
        len(query or ""),
    )
    parts: list[str] = []
+    cfg = config or load_rag_config()
+    service = get_service_config(service_id, cfg) if service_id else None
+    include_user_embeddings = service.use_user_embeddings if service else True
+    resolved_kb_name = kb_name or (service.knowledge_base if service else None)

-    user_soil = build_user_soil_text(sensor_uuid)
-    if user_soil and user_soil.strip():
-        parts.append("[داده‌های فعلی خاک شما]\n" + user_soil.strip())
-        logger.debug("Included user soil section sensor_uuid=%s", sensor_uuid)
-    else:
-        logger.info("No user soil data found sensor_uuid=%s", sensor_uuid)
+    if include_user_embeddings and sensor_uuid:
+        user_soil = build_user_soil_text(sensor_uuid)
+        if user_soil and user_soil.strip():
+            parts.append("[داده‌های فعلی خاک شما]\n" + user_soil.strip())
+            logger.debug("Included user soil section sensor_uuid=%s", sensor_uuid)
+        else:
+            logger.info("No user soil data found sensor_uuid=%s", sensor_uuid)

-    weather_text = build_user_weather_text(sensor_uuid)
-    if weather_text and weather_text.strip():
-        parts.append("[پیش‌بینی هواشناسی]\n" + weather_text.strip())
-        logger.debug("Included weather section sensor_uuid=%s", sensor_uuid)
-    else:
-        logger.info("No weather data found sensor_uuid=%s", sensor_uuid)
+        weather_text = build_user_weather_text(sensor_uuid)
+        if weather_text and weather_text.strip():
+            parts.append("[پیش‌بینی هواشناسی]\n" + weather_text.strip())
+            logger.debug("Included weather section sensor_uuid=%s", sensor_uuid)
+        else:
+            logger.info("No weather data found sensor_uuid=%s", sensor_uuid)

    results = search_with_query(
-        query, sensor_uuid=sensor_uuid, limit=limit, config=config,
-        kb_name=kb_name,
+        query,
+        sensor_uuid=sensor_uuid,
+        limit=limit,
+        config=cfg,
+        kb_name=resolved_kb_name,
+        service_id=service_id,
+        use_user_embeddings=include_user_embeddings,
    )
    if results:
        logger.info("Retrieved RAG results count=%s sensor_uuid=%s", len(results), sensor_uuid)
@@ -109,11 +130,12 @@ def build_rag_context(

 def chat_rag_stream(
    query: str,
-    sensor_uuid: str,
+    sensor_uuid: str | None = None,
    config: RAGConfig | None = None,
    limit: int = 5,
    system_override: str | None = None,
    kb_name: str | None = None,
+    service_id: str | None = None,
 ):
    logger.info(
        "chat_rag_stream started sensor_uuid=%s kb_name=%s limit=%s query_len=%s",
@@ -137,24 +159,43 @@ def chat_rag_stream(
        تک‌تک deltaهای content به‌صورت رشته
    """
    cfg = config or load_rag_config()
-    client = get_chat_client(cfg)
-    model = cfg.llm.model
-    logger.debug("Loaded RAG config with model=%s", model)
+    resolved_service_id = service_id or kb_name or _detect_kb_intent(query)
+    service = get_service_config(resolved_service_id, cfg)
+    service_llm_config = service.llm
+    service_cfg = RAGConfig(
+        embedding=cfg.embedding,
+        qdrant=cfg.qdrant,
+        chunking=cfg.chunking,
+        llm=service_llm_config,
+        knowledge_bases=cfg.knowledge_bases,
+        services=cfg.services,
+        chromadb=cfg.chromadb,
+    )
+    client = get_chat_client(service_cfg)
+    model = service_llm_config.model
+    logger.debug("Loaded service config service_id=%s model=%s", resolved_service_id, model)

-    detected_kb = kb_name or _detect_kb_intent(query)
-    logger.info("Using knowledge base=%s", detected_kb)
+    detected_kb = kb_name or service.knowledge_base
+    logger.info("Using knowledge base=%s for service_id=%s", detected_kb, resolved_service_id)
    context = build_rag_context(
-        query, sensor_uuid, config=cfg, limit=limit, kb_name=detected_kb,
+        query,
+        sensor_uuid,
+        config=cfg,
+        limit=limit,
+        kb_name=detected_kb,
+        service_id=resolved_service_id,
    )
    logger.debug("Built context length=%s", len(context))

    if system_override is not None:
        system_content = system_override
    else:
-        tone = _load_kb_tone(detected_kb, cfg)
+        tone = _load_service_tone(service, cfg)
        if not tone:
            tone = _load_tone(cfg)
        system_parts = [tone] if tone else []
+        if service.system_prompt:
+            system_parts.append(service.system_prompt)
        system_parts.append(
            "با استفاده از بخش «داده‌های فعلی خاک شما» و «متن‌های مرجع» زیر به سوال کاربر پاسخ بده. "
            "برای سوالاتی درباره خاک کاربر (مثل pH، رطوبت، NPK) حتماً از داده‌های فعلی استفاده کن. "
@@ -169,7 +210,7 @@ def chat_rag_stream(
        {"role": "system", "content": system_content},
        {"role": "user", "content": query},
    ]
-    logger.info("Prepared messages for model=%s  message=%s", model,messages)
+    logger.info("Prepared messages for model=%s service_id=%s", model, resolved_service_id)

    stream = client.chat.completions.create(
        model=model,
@@ -1,5 +1,6 @@
 """
-بارگذاری تنظیمات RAG از rag_config.yaml — با پشتیبانی از چند provider و چند پایگاه دانش
+بارگذاری تنظیمات RAG از rag_config.yaml — با پشتیبانی از چند provider،
+چند پایگاه دانش و چند سرویس.
 """
 import os
 from dataclasses import dataclass, field
@@ -36,6 +37,7 @@ class ChunkingConfig:

@dataclass
 class LLMConfig:
+    provider: str = "gapgpt"
    model: str = "gpt-4o"
    base_url: str | None = None
    api_key_env: str | None = None
@@ -50,6 +52,17 @@ class KnowledgeBaseConfig:
    description: str = ""


+@dataclass
+class ServiceConfig:
+    service_id: str
+    knowledge_base: str
+    llm: LLMConfig = field(default_factory=LLMConfig)
+    tone_file: str | None = None
+    system_prompt: str | None = None
+    use_user_embeddings: bool = True
+    description: str = ""
+
+
@dataclass
 class RAGConfig:
    embedding: EmbeddingConfig
@@ -57,9 +70,31 @@ class RAGConfig:
    chunking: ChunkingConfig
    llm: LLMConfig = field(default_factory=LLMConfig)
    knowledge_bases: dict[str, KnowledgeBaseConfig] = field(default_factory=dict)
+    services: dict[str, ServiceConfig] = field(default_factory=dict)
    chromadb: dict[str, Any] = field(default_factory=dict)


+def _build_llm_config(data: dict[str, Any] | None, default: LLMConfig | None = None) -> LLMConfig:
+    llm_data = data or {}
+    fallback = default or LLMConfig()
+    return LLMConfig(
+        provider=llm_data.get("provider", fallback.provider),
+        model=llm_data.get("model", fallback.model),
+        base_url=llm_data.get("base_url", fallback.base_url),
+        api_key_env=llm_data.get("api_key_env", fallback.api_key_env),
+        avalai_base_url=llm_data.get("avalai_base_url", fallback.avalai_base_url),
+        avalai_api_key_env=llm_data.get("avalai_api_key_env", fallback.avalai_api_key_env),
+    )
+
+
+def get_service_config(service_id: str, config: RAGConfig | None = None) -> ServiceConfig:
+    cfg = config or load_rag_config()
+    service = cfg.services.get(service_id)
+    if service is None:
+        raise KeyError(f"Unknown service_id: {service_id}")
+    return service
+
+
 def load_rag_config(config_path: str | Path | None = None) -> RAGConfig:
    """
    بارگذاری تنظیمات از YAML و env.
@@ -101,14 +136,7 @@ def load_rag_config(config_path: str | Path | None = None) -> RAGConfig:
        overlap_tokens=ch.get("overlap_tokens", 50),
    )

-    llm_data = data.get("llm", {})
-    llm = LLMConfig(
-        model=llm_data.get("model", "gpt-4o"),
-        base_url=llm_data.get("base_url"),
-        api_key_env=llm_data.get("api_key_env"),
-        avalai_base_url=llm_data.get("avalai_base_url"),
-        avalai_api_key_env=llm_data.get("avalai_api_key_env"),
-    )
+    llm = _build_llm_config(data.get("llm", {}))

    kb_data = data.get("knowledge_bases", {})
    knowledge_bases: dict[str, KnowledgeBaseConfig] = {}
@@ -119,11 +147,38 @@ def load_rag_config(config_path: str | Path | None = None) -> RAGConfig:
            description=kb_conf.get("description", ""),
        )

+    services_data = data.get("services", {})
+    services: dict[str, ServiceConfig] = {}
+    for service_id, service_conf in services_data.items():
+        kb_name = service_conf.get("knowledge_base", service_id)
+        kb_conf = knowledge_bases.get(kb_name)
+        services[service_id] = ServiceConfig(
+            service_id=service_id,
+            knowledge_base=kb_name,
+            llm=_build_llm_config(service_conf.get("llm"), default=llm),
+            tone_file=service_conf.get("tone_file") or (kb_conf.tone_file if kb_conf else None),
+            system_prompt=service_conf.get("system_prompt"),
+            use_user_embeddings=service_conf.get("use_user_embeddings", True),
+            description=service_conf.get("description", ""),
+        )
+
+    if not services:
+        for kb_name, kb_conf in knowledge_bases.items():
+            services[kb_name] = ServiceConfig(
+                service_id=kb_name,
+                knowledge_base=kb_name,
+                llm=llm,
+                tone_file=kb_conf.tone_file,
+                use_user_embeddings=True,
+                description=kb_conf.description,
+            )
+
    return RAGConfig(
        embedding=embedding,
        qdrant=qdrant,
        chunking=chunking,
        llm=llm,
        knowledge_bases=knowledge_bases,
+        services=services,
        chromadb=data.get("chromadb", {}),
    )
@@ -1,18 +1,20 @@
 """
 بازیابی RAG: embed کوئری و جستجو در vector store
 """
-from .config import load_rag_config, RAGConfig
+from .config import load_rag_config, RAGConfig, get_service_config
 from .embedding import embed_single
 from .vector_store import QdrantVectorStore


 def search_with_query(
    query: str,
-    sensor_uuid: str,
+    sensor_uuid: str | None = None,
    limit: int = 5,
    score_threshold: float | None = None,
    config: RAGConfig | None = None,
    kb_name: str | None = None,
+    service_id: str | None = None,
+    use_user_embeddings: bool | None = None,
 ) -> list[dict]:
    """
    کوئری را embed می‌کند و در vector store جستجو می‌کند.
@@ -27,12 +29,28 @@ def search_with_query(
        لیست نتایج با id, score, text, metadata
    """
    cfg = config or load_rag_config()
+    service = get_service_config(service_id, cfg) if service_id else None
+    resolved_kb_name = kb_name or (service.knowledge_base if service else None)
+    include_user_embeddings = (
+        use_user_embeddings
+        if use_user_embeddings is not None
+        else (service.use_user_embeddings if service else True)
+    )
+
+    sensor_filters = ["__global__"]
+    if include_user_embeddings and sensor_uuid:
+        sensor_filters.insert(0, sensor_uuid)
+
+    kb_filters = [resolved_kb_name] if resolved_kb_name else []
+    if include_user_embeddings:
+        kb_filters.append("__all__")
+
    query_vector = embed_single(query, config=cfg)
    store = QdrantVectorStore(config=cfg)
    return store.search(
        query_vector=query_vector,
        limit=limit,
        score_threshold=score_threshold,
-        sensor_uuid=sensor_uuid,
-        kb_name=kb_name,
+        sensor_uuids=sensor_filters,
+        kb_names=kb_filters,
    )
@@ -6,13 +6,14 @@ import json
 import logging

 from rag.api_provider import get_chat_client
-from rag.chat import build_rag_context, _load_kb_tone
-from rag.config import load_rag_config, RAGConfig
+from rag.chat import build_rag_context, _load_service_tone
+from rag.config import load_rag_config, RAGConfig, get_service_config
 from rag.user_data import build_plant_text

 logger = logging.getLogger(__name__)

 KB_NAME = "fertilization"
+SERVICE_ID = "fertilization"

 DEFAULT_FERTILIZATION_PROMPT = (
    "بر اساس داده‌های خاک (NPK، pH)، مشخصات گیاه، مرحله رشد و پایگاه دانش کودهی، "
@@ -56,13 +57,23 @@ def get_fertilization_recommendation(
        dict با کلیدهای fertilizer_needed, fertilizer_type, amount_kg_per_hectare, reason, npk_status, raw_response
    """
    cfg = config or load_rag_config()
-    client = get_chat_client(cfg)
-    model = cfg.llm.model
+    service = get_service_config(SERVICE_ID, cfg)
+    service_cfg = RAGConfig(
+        embedding=cfg.embedding,
+        qdrant=cfg.qdrant,
+        chunking=cfg.chunking,
+        llm=service.llm,
+        knowledge_bases=cfg.knowledge_bases,
+        services=cfg.services,
+        chromadb=cfg.chromadb,
+    )
+    client = get_chat_client(service_cfg)
+    model = service.llm.model

    user_query = query or "توصیه کودهی برای مزرعه من چیست؟"

    context = build_rag_context(
-        user_query, sensor_uuid, config=cfg, limit=limit, kb_name=KB_NAME,
+        user_query, sensor_uuid, config=cfg, limit=limit, kb_name=KB_NAME, service_id=SERVICE_ID,
    )

    extra_parts: list[str] = []
@@ -73,8 +84,10 @@ def get_fertilization_recommendation(
    if extra_parts:
        context = "\n\n---\n\n".join(extra_parts) + ("\n\n---\n\n" + context if context else "")

-    tone = _load_kb_tone(KB_NAME, cfg)
+    tone = _load_service_tone(service, cfg)
    system_parts = [tone] if tone else []
+    if service.system_prompt:
+        system_parts.append(service.system_prompt)
    system_parts.append(DEFAULT_FERTILIZATION_PROMPT)
    if context:
        system_parts.append("\n\n" + context)
@@ -5,18 +5,22 @@
 import json
 import logging

+from irrigation.evapotranspiration import calculate_forecast_water_needs, resolve_crop_profile, resolve_kc
+from sensor_data.models import SensorData
 from rag.api_provider import get_chat_client
-from rag.chat import build_rag_context, _load_kb_tone
-from rag.config import load_rag_config, RAGConfig
+from rag.chat import build_rag_context, _load_service_tone
+from rag.config import load_rag_config, RAGConfig, get_service_config
 from rag.user_data import build_plant_text, build_irrigation_method_text
+from weather.models import WeatherForecast

 logger = logging.getLogger(__name__)

 KB_NAME = "irrigation"
+SERVICE_ID = "irrigation"

 DEFAULT_IRRIGATION_PROMPT = (
-    "بر اساس داده‌های خاک، هواشناسی، مشخصات گیاه، روش آبیاری و پایگاه دانش آبیاری، "
-    "یک توصیه آبیاری دقیق بده. "
+    "بر اساس محاسبات نهایی تبخیر-تعرق و نیاز آبی که در ورودی آمده، "
+    "یک برنامه آبیاری قابل‌فهم برای کشاورز تولید کن. "
    "پاسخ حتماً به فرمت JSON با ساختار زیر باشد:\n"
    '{\n'
    '    "plan": {\n'
@@ -28,7 +32,7 @@ DEFAULT_IRRIGATION_PROMPT = (
    '    }\n'
    '}\n'
    "فقط JSON خروجی بده، بدون توضیح اضافی. "
-    "مقادیر عددی را بر اساس شرایط واقعی محاسبه کن."
+    "از انجام هرگونه محاسبه عددی جدید خودداری کن و فقط از داده‌های ساختاریافته ورودی استفاده کن."
 )


@@ -58,13 +62,52 @@ def get_irrigation_recommendation(
        dict با کلیدهای irrigation_needed, amount_mm, reason, next_check_date, raw_response
    """
    cfg = config or load_rag_config()
-    client = get_chat_client(cfg)
-    model = cfg.llm.model
+    service = get_service_config(SERVICE_ID, cfg)
+    service_cfg = RAGConfig(
+        embedding=cfg.embedding,
+        qdrant=cfg.qdrant,
+        chunking=cfg.chunking,
+        llm=service.llm,
+        knowledge_bases=cfg.knowledge_bases,
+        services=cfg.services,
+        chromadb=cfg.chromadb,
+    )
+    client = get_chat_client(service_cfg)
+    model = service.llm.model

    user_query = query or "توصیه آبیاری برای مزرعه من چیست؟"

+    sensor = SensorData.objects.select_related("location").prefetch_related("plants").filter(uuid_sensor=sensor_uuid).first()
+    plant = None
+    if sensor is not None and plant_name:
+        plant = sensor.plants.filter(name=plant_name).first()
+    elif sensor is not None:
+        plant = sensor.plants.first()
+    crop_profile = resolve_crop_profile(plant, growth_stage=growth_stage)
+    active_kc = resolve_kc(crop_profile, growth_stage=growth_stage)
+    forecasts = []
+    daily_water_needs = []
+    if sensor is not None:
+        forecasts = list(
+            WeatherForecast.objects.filter(location=sensor.location, forecast_date__isnull=False)
+            .order_by("forecast_date")[:7]
+        )
+        efficiency_percent = None
+        if irrigation_method_name:
+            from irrigation.models import IrrigationMethod
+
+            method = IrrigationMethod.objects.filter(name=irrigation_method_name).first()
+            efficiency_percent = getattr(method, "water_efficiency_percent", None) if method else None
+        daily_water_needs = calculate_forecast_water_needs(
+            forecasts=forecasts,
+            latitude_deg=float(sensor.location.latitude),
+            crop_profile=crop_profile,
+            growth_stage=growth_stage,
+            irrigation_efficiency_percent=efficiency_percent,
+        )
+
    context = build_rag_context(
-        user_query, sensor_uuid, config=cfg, limit=limit, kb_name=KB_NAME,
+        user_query, sensor_uuid, config=cfg, limit=limit, kb_name=KB_NAME, service_id=SERVICE_ID,
    )

    extra_parts: list[str] = []
@@ -76,11 +119,27 @@ def get_irrigation_recommendation(
        method_text = build_irrigation_method_text(irrigation_method_name)
        if method_text:
            extra_parts.append("[روش آبیاری انتخابی]\n" + method_text)
+    if daily_water_needs:
+        total_mm = round(sum(item["gross_irrigation_mm"] for item in daily_water_needs), 2)
+        schedule_lines = [
+            f"- {item['forecast_date']}: ET0={item['et0_mm']} mm, ETc={item['etc_mm']} mm, "
+            f"بارش مؤثر={item['effective_rainfall_mm']} mm, نیاز آبی={item['gross_irrigation_mm']} mm, "
+            f"زمان پیشنهادی={item['irrigation_timing']}"
+            for item in daily_water_needs
+        ]
+        extra_parts.append(
+            "[خروجی قطعی محاسبات FAO-56]\n"
+            f"کل نیاز آبی ۷ روز آینده: {total_mm} mm\n"
+            f"Kc مورد استفاده: {active_kc}\n"
+            + "\n".join(schedule_lines)
+        )
    if extra_parts:
        context = "\n\n---\n\n".join(extra_parts) + ("\n\n---\n\n" + context if context else "")

-    tone = _load_kb_tone(KB_NAME, cfg)
+    tone = _load_service_tone(service, cfg)
    system_parts = [tone] if tone else []
+    if service.system_prompt:
+        system_parts.append(service.system_prompt)
    system_parts.append(DEFAULT_IRRIGATION_PROMPT)
    if context:
        system_parts.append("\n\n" + context)
@@ -120,4 +179,9 @@ def get_irrigation_recommendation(
        }

    result["raw_response"] = raw
+    result["water_balance"] = {
+        "daily": daily_water_needs,
+        "crop_profile": crop_profile,
+        "active_kc": active_kc,
+    }
    return result
@@ -97,6 +97,8 @@ class QdrantVectorStore:
        score_threshold: float | None = None,
        sensor_uuid: str | None = None,
        kb_name: str | None = None,
+        sensor_uuids: list[str] | None = None,
+        kb_names: list[str] | None = None,
    ) -> list[dict]:
        """
        جستجوی شباهت بر اساس query vector.
@@ -107,34 +109,30 @@ class QdrantVectorStore:
        """
        must_conditions = []

-        if sensor_uuid:
+        sensor_values = [value for value in (sensor_uuids or ([sensor_uuid] if sensor_uuid else [])) if value]
+        if sensor_values:
            must_conditions.append(
                qmodels.Filter(
                    should=[
                        qmodels.FieldCondition(
                            key="sensor_uuid",
-                            match=qmodels.MatchValue(value=sensor_uuid),
-                        ),
-                        qmodels.FieldCondition(
-                            key="sensor_uuid",
-                            match=qmodels.MatchValue(value="__global__"),
-                        ),
+                            match=qmodels.MatchValue(value=value),
+                        )
+                        for value in sensor_values
                    ]
                )
            )

-        if kb_name:
+        kb_values = [value for value in (kb_names or ([kb_name] if kb_name else [])) if value]
+        if kb_values:
            must_conditions.append(
                qmodels.Filter(
                    should=[
                        qmodels.FieldCondition(
                            key="kb_name",
-                            match=qmodels.MatchValue(value=kb_name),
-                        ),
-                        qmodels.FieldCondition(
-                            key="kb_name",
-                            match=qmodels.MatchValue(value="__all__"),
-                        ),
+                            match=qmodels.MatchValue(value=value),
+                        )
+                        for value in kb_values
                    ]
                )
            )
@@ -24,8 +24,8 @@ logger = logging.getLogger(__name__)
 class ChatView(APIView):
    """
    چت RAG با استریم.
-    POST با {"message": "متن سوال", "sensor_uuid": "uuid-سنسور"}
-    sensor_uuid اجباری — هر کاربر فقط به دیتای خودش دسترسی دارد.
+    POST با {"service_id": "...", "query": "متن سوال", "user_id": "شناسه کاربر"}
+    service_id اجباری است. user_id فقط برای سرویس‌هایی که user embeddings دارند اجباری می‌شود.
    """

    @extend_schema(
@@ -35,8 +35,11 @@ class ChatView(APIView):
        request=inline_serializer(
            name="ChatRequest",
            fields={
-                "message": drf_serializers.CharField(help_text="متن سوال کاربر"),
-                "sensor_uuid": drf_serializers.CharField(help_text="شناسه یکتای سنسور"),
+                "service_id": drf_serializers.CharField(help_text="شناسه سرویس"),
+                "query": drf_serializers.CharField(required=False, help_text="متن سوال کاربر"),
+                "message": drf_serializers.CharField(required=False, help_text="نام قبلی فیلد query"),
+                "user_id": drf_serializers.CharField(required=False, help_text="شناسه کاربر"),
+                "sensor_uuid": drf_serializers.CharField(required=False, help_text="نام قبلی فیلد user_id"),
            },
        ),
        responses={
@@ -50,19 +53,25 @@ class ChatView(APIView):
        examples=[
            OpenApiExample(
                "نمونه درخواست",
-                value={"message": "وضعیت خاک من چطوره؟", "sensor_uuid": "550e8400-e29b-41d4-a716-446655440000"},
+                value={
+                    "service_id": "support_bot",
+                    "user_id": "12345",
+                    "query": "How do I reset my password?",
+                },
                request_only=True,
            ),
        ],
    )
    def post(self, request: Request):
+        from .config import load_rag_config, get_service_config
+
        data = request.data if request.method == "POST" else request.query_params
-        message = data.get("message")
-        sensor_uuid = data.get("sensor_uuid")
-        logging.info("jhh")
+        service_id = data.get("service_id")
+        message = data.get("query", data.get("message"))
+        user_id = data.get("user_id", data.get("sensor_uuid"))
        if not message or not isinstance(message, str):
            return Response(
-                {"code": 400, "msg": "پارامتر message الزامی است."},
+                {"code": 400, "msg": "پارامتر query الزامی است."},
                status=status.HTTP_400_BAD_REQUEST,
            )
        message = str(message).strip()
@@ -71,22 +80,43 @@ class ChatView(APIView):
                {"code": 400, "msg": "پیام نباید خالی باشد."},
                status=status.HTTP_400_BAD_REQUEST,
            )
-        if not sensor_uuid or not isinstance(sensor_uuid, str):
+        if not service_id or not isinstance(service_id, str):
            return Response(
-                {"code": 400, "msg": "پارامتر sensor_uuid الزامی است."},
+                {"code": 400, "msg": "پارامتر service_id الزامی است."},
                status=status.HTTP_400_BAD_REQUEST,
            )
-        sensor_uuid = str(sensor_uuid).strip()
-        if not sensor_uuid:
+        service_id = str(service_id).strip()
+        if not service_id:
            return Response(
-                {"code": 400, "msg": "sensor_uuid نباید خالی باشد."},
+                {"code": 400, "msg": "service_id نباید خالی باشد."},
+                status=status.HTTP_400_BAD_REQUEST,
+            )
+        cfg = load_rag_config()
+        try:
+            service = get_service_config(service_id, cfg)
+        except KeyError:
+            return Response(
+                {"code": 400, "msg": f"service_id نامعتبر است: {service_id}"},
+                status=status.HTTP_400_BAD_REQUEST,
+            )
+        if user_id is not None:
+            user_id = str(user_id).strip()
+            if not user_id:
+                user_id = None
+        if service.use_user_embeddings and not user_id:
+            return Response(
+                {"code": 400, "msg": "برای این service_id، پارامتر user_id الزامی است."},
                status=status.HTTP_400_BAD_REQUEST,
            )
-        

        def generate():
            try:
-                for chunk in chat_rag_stream(message, sensor_uuid=sensor_uuid):
+                for chunk in chat_rag_stream(
+                    message,
+                    sensor_uuid=user_id,
+                    service_id=service_id,
+                    config=cfg,
+                ):
                    yield chunk
            except Exception as e:
                yield f"\n[خطا: {e}]"