feat: solution for 6a02e23da6fe2e4ac16acf65

solutions/6a02e23da6fe2e4ac16acf65/solution.py

@@ -1,58 +1,64 @@
-from pathlib import Path
-
-# LLM and embeddings via Ollama
-from langchain_ollama import ChatOllama, OllamaEmbeddings
-from langchain.tools import tool
-from langchain.agents import create_agent
-from langchain_core.documents import Document
+from langchain_ollama import OllamaEmbeddings
 from langchain_qdrant import QdrantVectorStore
 from qdrant_client import QdrantClient
 from qdrant_client.http.models import Distance, VectorParams
-from langchain_core.messages import HumanMessage
+from langchain_core.documents import Document
+from langchain_text_splitters import RecursiveCharacterTextSplitter
+from langchain.tools import tool
+from langchain.agents import create_agent
+from langchain_openai import ChatOpenAI
+from pydantic import SecretStr
 
-# ---------- Qdrant setup ----------
+# ---------- LLM and embeddings ----------
+llm = ChatOpenAI(
+    model="openai/gpt-oss-20b",
+    base_url='https://platform.brojs.ru/jrnl-bh/api/inference/v1',
+    api_key=SecretStr("jrnl_30283ab953615cbb6846ff9940a1eedce0b76d7b2f59a2394f29e74643e6a90d"),
+    temperature=0.7,
+)
+
+embeddings = OllamaEmbeddings(model_name="nomic-embed-text")
+
+# ---------- Qdrant client ----------
 client = QdrantClient(":memory:")
 client.create_collection(
-    collection_name="knowledge",
-    vectors_config=VectorParams(size=1024, distance=Distance.COSINE),
-)
-embeddings = OllamaEmbeddings(model="nomic-embed-text")
-vector_store = QdrantVectorStore(
-    client=client,
-    collection_name="knowledge",
-    embedding=embeddings,
+    collection_name="knowledge_base",
+    vectors_config=VectorParams(size=embeddings.embedding_size, distance=Distance.COSINE),
 )
+vector_store = QdrantVectorStore(client=client, collection_name="knowledge_base", embedding=embeddings)
+
+# ---------- Text splitter ----------
+splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
 
 # ---------- Tools ----------
 @tool
-def search_knowledge_base(query: str, max_results: int = 5) -> str:
-    """Search the knowledge base for relevant documents."""
-    docs_with_score = vector_store.similarity_search_with_score(query, k=max_results)
-    if not docs_with_score:
-        return "No results found."
-    return "\n".join(
-        f"{i+1}. {doc.page_content[:200]}..."
-        for i, (doc, _) in enumerate(docs_with_score)
-    )
+def add_to_knowledge_base(content: str, title: str) -> str:
+    """Add a document to the knowledge base."""
+    docs = [Document(page_content=c, metadata={"title": title}) for c in splitter.split_text(content)]
+    vector_store.add_documents(docs)
+    return f"Added {len(docs)} chunks titled '{title}'."
 
 @tool
-def add_to_knowledge_base(content: str, title: str = "") -> str:
-    """Add a new document to the knowledge base."""
-    doc = Document(page_content=content, metadata={"title": title})
-    vector_store.add_documents([doc])
-    return f"Document '{title}' added."
+def search_knowledge_base(query: str, max_results: int = 5) -> str:
+    """Search the knowledge base for relevant information."""
+    results = vector_store.similarity_search_with_score(query, k=max_results)
+    if not results:
+        return "No relevant documents found."
+    reply = ""
+    for i, (doc, score) in enumerate(results, start=1):
+        reply += f"{i}. ({score:.2f}) {doc.metadata.get('title', 'Untitled')}: {doc.page_content[:200]}...\n"
+    return reply.strip()
 
 # ---------- Agent ----------
-llm = ChatOllama(model="llama3")
-agent = create_agent(
-    model=llm,
-    tools=[search_knowledge_base, add_to_knowledge_base],
-    system_prompt="You are a helpful assistant that can search and store knowledge.",
+system_prompt = (
+    "You are an assistant that can search and add information to a knowledge base. "
+    "Use the tools `add_to_knowledge_base` and `search_knowledge_base` as needed."
 )
+agent = create_agent(model=llm, tools=[add_to_knowledge_base, search_knowledge_base], system_prompt=system_prompt)
 
 # ---------- CLI ----------
 def main():
-    print("RAG Agent CLI. Commands: /add <title> | <content>, /search <query>, /quit")
+    print("RAG Agent CLI. Commands: /add <title> <content>, /search <query>, /quit")
     while True:
         try:
             inp = input("> ").strip()
@@ -60,31 +66,30 @@ def main():
             break
         if not inp:
             continue
-        if inp.lower() in ("quit", "/quit"):
+        if inp.lower() in ("/quit", "exit"):
             print("Bye!")
             break
-
-        # Add document
-        if inp.startswith("/add "):
-            _, rest = inp.split(maxsplit=1)
-            try:
-                title, content = rest.split("|", 1)
-            except ValueError:
-                print("Usage: /add <title> | <content>")
+        if inp.startswith("/add"):
+            parts = inp.split(maxsplit=2)
+            if len(parts) < 3:
+                print("Usage: /add <title> <content>")
                 continue
-            res = agent.invoke({"messages": [HumanMessage(content=f"Add document {title}")]})
-            print(res.messages[-1].content)
-
-        # Search documents
-        elif inp.startswith("/search "):
-            query = inp[len("/search "):]
-            res = agent.invoke({"messages": [HumanMessage(content=f"Search for {query}")]})
-            print(res.messages[-1].content)
-
-        # General chat
+            title, content = parts[1], parts[2]
+            res_msg = agent.invoke({"messages": [{"role": "human", "content": f"/add {title} {content}"}]})
+            for msg in res_msg["messages"]:
+                if hasattr(msg, "tool_calls"):
+                    print(msg.tool_calls[0]["output"])
+        elif inp.startswith("/search"):
+            query = inp[len("/search"):].strip()
+            if not query:
+                print("Usage: /search <query>")
+                continue
+            res_msg = agent.invoke({"messages": [{"role": "human", "content": f"/search {query}"}]})
+            for msg in res_msg["messages"]:
+                if hasattr(msg, "tool_calls"):
+                    print(msg.tool_calls[0]["output"])
         else:
-            res = agent.invoke({"messages": [HumanMessage(content=inp)]})
-            print(res.messages[-1].content)
+            print("Unknown command. Use /add, /search, or /quit.")
 
 if __name__ == "__main__":
     main()