WIP

compose.yml

@@ -9,25 +9,25 @@ services:
       - "${OLLAMA_PORT:-11434}:11434"
     restart: "no"
 
-  ai-qdrant:
+  # ai-qdrant:
-    container_name: ai-qdrant
+  #   container_name: ai-qdrant
-    image: qdrant/qdrant
+  #   image: qdrant/qdrant
-    env_file: .env
+  #   env_file: .env
-    ports:
+  #   ports:
-      - "${QDRANT_PORT:-6333}:6333"
+  #     - "${QDRANT_PORT:-6333}:6333"
-    volumes:
+  #   volumes:
-      - ./.data/qdrant/storage:/qdrant/storage
+  #     - ./.data/qdrant/storage:/qdrant/storage
-    restart: "no"
+  #   restart: "no"
-    profiles: ["rag"]
+  #   profiles: ["rag"]
 
-  ai-webui:
+  # ai-webui:
-    container_name: ai-webui
+  #   container_name: ai-webui
-    image: ghcr.io/open-webui/open-webui:main
+  #   image: ghcr.io/open-webui/open-webui:main
-    env_file: .env
+  #   env_file: .env
-    volumes:
+  #   volumes:
-      - ./.data/webui:/app/backend/data
+  #     - ./.data/webui:/app/backend/data
-    ports:
+  #   ports:
-      - "${OWEBUI_PORT:-9999}:8080"
+  #     - "${OWEBUI_PORT:-9999}:8080"
-    extra_hosts:
+  #   extra_hosts:
-      - "host.docker.internal:host-gateway"
+  #     - "host.docker.internal:host-gateway"
-    restart: "no"
+  #   restart: "no"

rag/README.md

@@ -8,7 +8,7 @@
 cd ..; ./up; cd -
 python3 -m venv .venv
 source .venv/bin/activate
-pip install beautifulsoup4 markdownify sentence-transformers qdrant-client langchain transformers
+pip install beautifulsoup4 markdownify sentence-transformers qdrant-client langchain transformers ollama
 pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu
 ./download.sh 123456789 # <<== pageId страницы в Confluence
 python3 convert.py
@@ -66,7 +66,7 @@ rag/
 ```bash
 python3 -m venv .venv
 source ./venv/bin/activate
-pip install beautifulsoup4 markdownify sentence-transformers qdrant-client langchain transformers
+pip install beautifulsoup4 markdownify sentence-transformers qdrant-client langchain transformers ollama
 pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu
 ```
 

rag/rag.py

@@ -1,10 +1,9 @@
 import os
-import requests
-import json
 import time
 import sys
 from qdrant_client import QdrantClient
 from sentence_transformers import SentenceTransformer, CrossEncoder
+import ollama
 
 DEFAULT_CHAT_MODEL = "openchat:7b"
 DEFAULT_EMBED_MODEL = "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2"
@@ -38,33 +37,26 @@ class RagSystem:
         self.qdrant_port = qdrant_port
         self.chat_model = chat_model
         self.emb_model = SentenceTransformer(embed_model)
-        self.qdrant = QdrantClient(host=args.qdrant_host, port=args.qdrant_port)
+        self.qdrant = QdrantClient(host=qdrant_host, port=qdrant_port)
         self.use_rank = use_rank
         if self.use_rank:
             self.rank_model = CrossEncoder(rank_model)
         self.conversation_history = []
+        self.ollama = ollama.Client(base_url=ollama_url)
 
     def check_chat_model(self):
-        response = requests.get(f"{self.ollama_url}/api/tags")
+        models = self.ollama.list()
-        if response.status_code != 200:
+        return any(model.name == self.chat_model for model in models)
-            return False
-        for model in response.json().get("models", []):
-            if model["name"] == self.chat_model:
-                return True
-        return False
 
     def install_chat_model(self, model: str = DEFAULT_CHAT_MODEL):
         try:
-            response = requests.post(f"{self.ollama_url}/api/pull", json={"model": model})
+            result = self.ollama.pull(model)
-            if response.status_code == 200:
+            print(f"Модель {model} установлена успешно")
-                print(f"Модель {self.chat_model} установлена успешно")
-            else:
-                print(f"Ошибка установки модели: {response.text}")
         except Exception as e:
-            print(f"Ошибка проверки модели: {str(e)}")
+            print(f"Ошибка установки модели: {str(e)}")
 
     def load_chat_model(self):
-        requests.post(f"{self.ollama_url}/api/generate", json={"model": self.chat_model}, timeout=600)
+        self.ollama.generate(model=self.chat_model, keep_alive=True)
 
     def search_qdrant(self, query: str, doc_count: int = DEFAULT_TOP_K, collection_name = DEFAULT_QDRANT_COLLECTION):
         query_vec = self.emb_model.encode(query, show_progress_bar=False).tolist()
@@ -100,85 +92,71 @@ class RagSystem:
         return ranked_docs[:top_n]
 
     def generate_answer(self, sys_prompt: str, user_prompt: str):
-        url = f"{self.ollama_url}/api/generate"
+        try:
-        body = {
+            with self.ollama.generate(
-            "model": self.chat_model,
+                model=self.chat_model,
-            "system": sys_prompt,
+                prompt=sys_prompt + "\n" + user_prompt,
-            "prompt": user_prompt,
+               options={
-            "stream": False,
+                   "temperature": 0.5,
-            "options": {
+               },
-                "temperature": 0.5,
+               stream=False,
-                # "top_p": 0.2,
+           ) as generator:
-            },
+                response = next(generator)
-        }
+                if response.error:
-
+                    raise RuntimeError(f"Ошибка генерации: {response.error}")
-        response = requests.post(url, json=body, timeout=900)
+                self.last_response = response
-        if response.status_code != 200:
+                return response.output
-            return f"Ошибка генерации ответа: {response.status_code} {response.text}"
+        except Exception as e:
-        self.response = response.json()
+            print(f"Ошибка генерации ответа: {str(e)}")
-        return self.response["response"]
+            return str(e)
 
     def generate_answer_stream(self, sys_prompt: str, user_prompt: str):
-        url = f"{self.ollama_url}/api/generate"
+        try:
-        body = {
+            generator = self.ollama.generate(
-            "model": self.chat_model,
+                model=self.chat_model,
-            "system": sys_prompt,
+                prompt=sys_prompt + "\n" + user_prompt,
-            "prompt": user_prompt,
+               options={
-            "stream": True,
+                   "temperature": 0.5,
-            "options": {
+               },
-                "temperature": 0.5,
+               stream=True,
-                # "top_p": 0.2,
+           )
-            },
+            answer = ""
-        }
+            for response in generator:
-        resp = requests.post(url, json=body, stream=True, timeout=900)
+                if response.data:
-        if resp.status_code != 200:
+                    yield response.data
-            raise RuntimeError(f"Ошибка генерации ответа: {resp.status_code} {resp.text}")
+                    answer += response.data
-
+                if response.done:
-        answer = ""
+                    self.last_response = response
-        self.response = None
+                    break
-        for chunk in resp.iter_lines():
+            return answer
-            if chunk:
+        except Exception as e:
-                try:
+            print(f"Ошибка стриминга: {str(e)}")
-                    decoded_chunk = chunk.decode('utf-8')
+            return str(e)
-                    data = json.loads(decoded_chunk)
-                    if "response" in data:
-                        yield data["response"]
-                        answer += data["response"]
-                    if "done" in data and data["done"] is True:
-                        self.response = data
-                        break
-                    elif "error" in data:
-                        answer += f" | Ошибка стриминга ответа: {data['error']}"
-                        break
-                except json.JSONDecodeError as e:
-                    answer += f" | Ошибка конвертации чанка: {chunk.decode('utf-8')} - {e}"
-                except Exception as e:
-                    answer += f" | Ошибка обработки чанка: {e}"
 
     def get_prompt_eval_count(self):
-        if not self.response:
+        if not hasattr(self, "last_response"):
             return 0
-        return self.response["prompt_eval_count"]
+        return self.last_response.prompt_eval_count or 0
 
     def get_prompt_eval_duration(self):
-        if not self.response:
+        if not hasattr(self, "last_response"):
             return 0
-        return self.response["prompt_eval_duration"] / (10 ** 9)
+        return self.last_response.prompt_eval_duration / (10 ** 9)
 
     def get_eval_count(self):
-        if not self.response:
+        if not hasattr(self, "last_response"):
             return 0
-        return self.response["eval_count"]
+        return self.last_response.eval_count or 0
 
     def get_eval_duration(self):
-        if not self.response:
+        if not hasattr(self, "last_response"):
             return 0
-        return self.response["eval_duration"] / (10 ** 9)
+        return self.last_response.eval_duration / (10 ** 9)
 
     def get_total_duration(self):
-        if not self.response:
+        if not hasattr(self, "last_response"):
             return 0
-        return self.response["total_duration"] / (10 ** 9)
+        return self.last_response.total_duration / (10 ** 9)
 
     def get_tps(self):
         eval_count = self.get_eval_count()
@@ -360,19 +338,23 @@ Context:
 
     def process_query(self, sys_prompt: str, user_prompt: str, streaming: bool = DEFAULT_STREAM):
         answer = ""
-        # try:
         if streaming:
             self.print_v(text="\nГенерация потокового ответа (^C для остановки)...\n")
             print(f"<<< ", end='', flush=True)
-            for token in self.rag.generate_answer_stream(sys_prompt, user_prompt):
+            try:
-                answer += token
+                for token in self.rag.generate_answer_stream(sys_prompt, user_prompt):
-                print(token, end='', flush=True)
+                    answer += token
+                    print(token, end='', flush=True)
+            except KeyboardInterrupt:
+                print("\n*** Генерация ответа прервана")
+                return answer
         else:
             self.print_v(text="\nГенерация ответа (^C для остановки)...\n")
-            answer = self.rag.generate_answer(sys_prompt, user_prompt)
+            try:
-            print(f"<<< {answer}\n")
+                answer = self.rag.generate_answer(sys_prompt, user_prompt)
-        # except RuntimeError as e:
+            except KeyboardInterrupt:
-            # answer = str(e)
+                print("\n*** Генерация ответа прервана")
+                return ""
 
         print(f"\n===================================================")
         return answer