haiku.rag 0.3.2__py3-none-any.whl → 0.3.3__py3-none-any.whl

haiku/rag/qa/anthropic.py

@@ -37,75 +37,69 @@ try:
 
             messages: list[MessageParam] = [{"role": "user", "content": question}]
 
-            response = await anthropic_client.messages.create(
-                model=self._model,
-                max_tokens=4096,
-                system=self._system_prompt,
-                messages=messages,
-                tools=self.tools,
-                temperature=0.0,
-            )
-
-            if response.stop_reason == "tool_use":
-                messages.append({"role": "assistant", "content": response.content})
-
-                # Process tool calls
-                tool_results = []
-                for content_block in response.content:
-                    if isinstance(content_block, ToolUseBlock):
-                        if content_block.name == "search_documents":
-                            args = content_block.input
-                            query = (
-                                args.get("query", question)
-                                if isinstance(args, dict)
-                                else question
-                            )
-                            limit = (
-                                int(args.get("limit", 3))
-                                if isinstance(args, dict)
-                                else 3
-                            )
-
-                            search_results = await self._client.search(
-                                query, limit=limit
-                            )
-
-                            context_chunks = []
-                            for chunk, score in search_results:
-                                context_chunks.append(
-                                    f"Content: {chunk.content}\nScore: {score:.4f}"
+            max_rounds = 5  # Prevent infinite loops
+
+            for _ in range(max_rounds):
+                response = await anthropic_client.messages.create(
+                    model=self._model,
+                    max_tokens=4096,
+                    system=self._system_prompt,
+                    messages=messages,
+                    tools=self.tools,
+                    temperature=0.0,
+                )
+
+                if response.stop_reason == "tool_use":
+                    messages.append({"role": "assistant", "content": response.content})
+
+                    # Process tool calls
+                    tool_results = []
+                    for content_block in response.content:
+                        if isinstance(content_block, ToolUseBlock):
+                            if content_block.name == "search_documents":
+                                args = content_block.input
+                                query = (
+                                    args.get("query", question)
+                                    if isinstance(args, dict)
+                                    else question
+                                )
+                                limit = (
+                                    int(args.get("limit", 3))
+                                    if isinstance(args, dict)
+                                    else 3
+                                )
+
+                                search_results = await self._client.search(
+                                    query, limit=limit
+                                )
+
+                                context_chunks = []
+                                for chunk, score in search_results:
+                                    context_chunks.append(
+                                        f"Content: {chunk.content}\nScore: {score:.4f}"
+                                    )
+
+                                context = "\n\n".join(context_chunks)
+
+                                tool_results.append(
+                                    {
+                                        "type": "tool_result",
+                                        "tool_use_id": content_block.id,
+                                        "content": context,
+                                    }
                                 )
 
-                            context = "\n\n".join(context_chunks)
-
-                            tool_results.append(
-                                {
-                                    "type": "tool_result",
-                                    "tool_use_id": content_block.id,
-                                    "content": context,
-                                }
-                            )
-
-                if tool_results:
-                    messages.append({"role": "user", "content": tool_results})
-
-                    final_response = await anthropic_client.messages.create(
-                        model=self._model,
-                        max_tokens=4096,
-                        system=self._system_prompt,
-                        messages=messages,
-                        temperature=0.0,
-                    )
-                    if final_response.content:
-                        first_content = final_response.content[0]
+                    if tool_results:
+                        messages.append({"role": "user", "content": tool_results})
+                else:
+                    # No tool use, return the response
+                    if response.content:
+                        first_content = response.content[0]
                         if isinstance(first_content, TextBlock):
                             return first_content.text
                     return ""
 
-            if response.content:
-                first_content = response.content[0]
-                if isinstance(first_content, TextBlock):
-                    return first_content.text
+            # If we've exhausted max rounds, return empty string
             return ""
 
 except ImportError:

haiku/rag/qa/ollama.py

@@ -14,54 +14,51 @@ class QuestionAnswerOllamaAgent(QuestionAnswerAgentBase):
     async def answer(self, question: str) -> str:
         ollama_client = AsyncClient(host=Config.OLLAMA_BASE_URL)
 
-        # Define the search tool
-
         messages = [
             {"role": "system", "content": self._system_prompt},
             {"role": "user", "content": question},
         ]
 
-        # Initial response with tool calling
-        response = await ollama_client.chat(
-            model=self._model,
-            messages=messages,
-            tools=self.tools,
-            options=OLLAMA_OPTIONS,
-            think=False,
-        )
+        max_rounds = 5  # Prevent infinite loops
 
-        if response.get("message", {}).get("tool_calls"):
-            for tool_call in response["message"]["tool_calls"]:
-                if tool_call["function"]["name"] == "search_documents":
-                    args = tool_call["function"]["arguments"]
-                    query = args.get("query", question)
-                    limit = int(args.get("limit", 3))
+        for _ in range(max_rounds):
+            response = await ollama_client.chat(
+                model=self._model,
+                messages=messages,
+                tools=self.tools,
+                options=OLLAMA_OPTIONS,
+                think=False,
+            )
 
-                    search_results = await self._client.search(query, limit=limit)
+            if response.get("message", {}).get("tool_calls"):
+                messages.append(response["message"])
 
-                    context_chunks = []
-                    for chunk, score in search_results:
-                        context_chunks.append(
-                            f"Content: {chunk.content}\nScore: {score:.4f}"
-                        )
+                for tool_call in response["message"]["tool_calls"]:
+                    if tool_call["function"]["name"] == "search_documents":
+                        args = tool_call["function"]["arguments"]
+                        query = args.get("query", question)
+                        limit = int(args.get("limit", 3))
 
-                    context = "\n\n".join(context_chunks)
+                        search_results = await self._client.search(query, limit=limit)
 
-                    messages.append(response["message"])
-                    messages.append(
-                        {
-                            "role": "tool",
-                            "content": context,
-                            "tool_call_id": tool_call.get("id", "search_tool"),
-                        }
-                    )
+                        context_chunks = []
+                        for chunk, score in search_results:
+                            context_chunks.append(
+                                f"Content: {chunk.content}\nScore: {score:.4f}"
+                            )
 
-            final_response = await ollama_client.chat(
-                model=self._model,
-                messages=messages,
-                think=False,
-                options=OLLAMA_OPTIONS,
-            )
-            return final_response["message"]["content"]
-        else:
-            return response["message"]["content"]
+                        context = "\n\n".join(context_chunks)
+
+                        messages.append(
+                            {
+                                "role": "tool",
+                                "content": context,
+                                "tool_call_id": tool_call.get("id", "search_tool"),
+                            }
+                        )
+            else:
+                # No tool calls, return the response
+                return response["message"]["content"]
+
+        # If we've exhausted max rounds, return empty string
+        return ""

haiku/rag/qa/openai.py

@@ -24,8 +24,6 @@ try:
         async def answer(self, question: str) -> str:
             openai_client = AsyncOpenAI()
 
-            # Define the search tool
-
             messages: list[ChatCompletionMessageParam] = [
                 ChatCompletionSystemMessageParam(
                     role="system", content=self._system_prompt
@@ -33,69 +31,70 @@ try:
                 ChatCompletionUserMessageParam(role="user", content=question),
             ]
 
-            # Initial response with tool calling
-            response = await openai_client.chat.completions.create(
-                model=self._model,
-                messages=messages,
-                tools=self.tools,
-                temperature=0.0,
-            )
-
-            response_message = response.choices[0].message
-
-            if response_message.tool_calls:
-                messages.append(
-                    ChatCompletionAssistantMessageParam(
-                        role="assistant",
-                        content=response_message.content,
-                        tool_calls=[
-                            {
-                                "id": tc.id,
-                                "type": "function",
-                                "function": {
-                                    "name": tc.function.name,
-                                    "arguments": tc.function.arguments,
-                                },
-                            }
-                            for tc in response_message.tool_calls
-                        ],
-                    )
+            max_rounds = 5  # Prevent infinite loops
+
+            for _ in range(max_rounds):
+                response = await openai_client.chat.completions.create(
+                    model=self._model,
+                    messages=messages,
+                    tools=self.tools,
+                    temperature=0.0,
                 )
 
-                for tool_call in response_message.tool_calls:
-                    if tool_call.function.name == "search_documents":
-                        import json
+                response_message = response.choices[0].message
+
+                if response_message.tool_calls:
+                    messages.append(
+                        ChatCompletionAssistantMessageParam(
+                            role="assistant",
+                            content=response_message.content,
+                            tool_calls=[
+                                {
+                                    "id": tc.id,
+                                    "type": "function",
+                                    "function": {
+                                        "name": tc.function.name,
+                                        "arguments": tc.function.arguments,
+                                    },
+                                }
+                                for tc in response_message.tool_calls
+                            ],
+                        )
+                    )
 
-                        args = json.loads(tool_call.function.arguments)
-                        query = args.get("query", question)
-                        limit = int(args.get("limit", 3))
+                    for tool_call in response_message.tool_calls:
+                        if tool_call.function.name == "search_documents":
+                            import json
 
-                        search_results = await self._client.search(query, limit=limit)
+                            args = json.loads(tool_call.function.arguments)
+                            query = args.get("query", question)
+                            limit = int(args.get("limit", 3))
 
-                        context_chunks = []
-                        for chunk, score in search_results:
-                            context_chunks.append(
-                                f"Content: {chunk.content}\nScore: {score:.4f}"
+                            search_results = await self._client.search(
+                                query, limit=limit
                             )
 
-                        context = "\n\n".join(context_chunks)
+                            context_chunks = []
+                            for chunk, score in search_results:
+                                context_chunks.append(
+                                    f"Content: {chunk.content}\nScore: {score:.4f}"
+                                )
+
+                            context = "\n\n".join(context_chunks)
 
-                        messages.append(
-                            ChatCompletionToolMessageParam(
-                                role="tool",
-                                content=context,
-                                tool_call_id=tool_call.id,
+                            messages.append(
+                                ChatCompletionToolMessageParam(
+                                    role="tool",
+                                    content=context,
+                                    tool_call_id=tool_call.id,
+                                )
                             )
-                        )
+                else:
+                    # No tool calls, return the response
+                    return response_message.content or ""
 
-                final_response = await openai_client.chat.completions.create(
-                    model=self._model,
-                    messages=messages,
-                    temperature=0.0,
-                )
-                return final_response.choices[0].message.content or ""
-            else:
-                return response_message.content or ""
+            # If we've exhausted max rounds, return empty string
+            return ""
 
 except ImportError:
     pass

haiku/rag/qa/prompts.py

@@ -1,7 +1,20 @@
 SYSTEM_PROMPT = """
-You are a helpful assistant that uses a RAG library to answer the user's prompt.
-Your task is to provide a concise and accurate answer based on the provided context.
-You should ask the provided tools to find relevant documents and then use the content of those documents to answer the question.
-Never make up information, always use the context to answer the question.
-If the context does not contain enough information to answer the question, respond with "I cannot answer that based on the provided context."
+You are a knowledgeable assistant that helps users find information from a document knowledge base.
+
+Your process:
+1. When a user asks a question, use the search_documents tool to find relevant information
+2. Search with specific keywords and phrases from the user's question
+3. Review the search results and their relevance scores
+4. If you need additional context, perform follow-up searches with different keywords
+5. Provide a comprehensive answer based only on the retrieved documents
+
+Guidelines:
+- Base your answers strictly on the provided document content
+- Quote or reference specific information when possible
+- If multiple documents contain relevant information, synthesize them coherently
+- Indicate when information is incomplete or when you need to search for additional context
+- If the retrieved documents don't contain sufficient information, clearly state: "I cannot find enough information in the knowledge base to answer this question."
+- For complex questions, consider breaking them down and performing multiple searches
+
+Be concise, and always maintain accuracy over completeness. Prefer short, direct answers that are well-supported by the documents.
 """

{haiku_rag-0.3.2.dist-info → haiku_rag-0.3.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: haiku.rag
-Version: 0.3.2
+Version: 0.3.3
 Summary: Retrieval Augmented Generation (RAG) with SQLite
 Author-email: Yiorgis Gozadinos <ggozadinos@gmail.com>
 License: MIT
@@ -116,3 +116,4 @@ Full documentation at: https://ggozad.github.io/haiku.rag/
 - [Configuration](https://ggozad.github.io/haiku.rag/configuration/) - Environment variables
 - [CLI](https://ggozad.github.io/haiku.rag/cli/) - Command reference
 - [Python API](https://ggozad.github.io/haiku.rag/python/) - Complete API docs
+- [Benchmarks](https://ggozad.github.io/haiku.rag/benchmarks/) - Performance Benchmarks

{haiku_rag-0.3.2.dist-info → haiku_rag-0.3.3.dist-info}/RECORD

@@ -15,11 +15,11 @@ haiku/rag/embeddings/ollama.py,sha256=hWdrTiuJwNSRYCqP0WP-z6XXA3RBGkAiknZMsPLH0q
 haiku/rag/embeddings/openai.py,sha256=reh8AykG2f9f5hhRDmqSsjiuCPi9SsXfe2YEZFlxXk8,550
 haiku/rag/embeddings/voyageai.py,sha256=jc0JywdLJD3Ee1MUv1m8MhWCEo0enNnVcrIBtUvD-Ss,534
 haiku/rag/qa/__init__.py,sha256=oso98Ypti7mBLTJ6Zk71YaSJ9Rgc89QXp9RSB6zSpYs,1501
-haiku/rag/qa/anthropic.py,sha256=lzHRQxpEv6Qd6iBIqexUgWnq-ITqytppwkfOuRGWdDs,4556
+haiku/rag/qa/anthropic.py,sha256=6I6cf6ySNkYbmDFdy22sA8r3GO5moiiH75tJnHcgJQA,4448
 haiku/rag/qa/base.py,sha256=4ZTM_l5FAZ9cA0f8NeqRJiUAmjatwCTmSoclFw0gTFQ,1349
-haiku/rag/qa/ollama.py,sha256=poShrse-RgLTwa5gbVzoERNTrn5QRpovJCZKYkIpOZI,2393
-haiku/rag/qa/openai.py,sha256=yBbSjGlG4Lo5p2B2NOTa5C6JceX0OJ1jXar_ABFZYYI,3849
-haiku/rag/qa/prompts.py,sha256=dAz2HjD4eJ8tcW534Tx7EuFOs6pSv2kPr7yrHnHtS0E,535
+haiku/rag/qa/ollama.py,sha256=-UtNFErYlA_66g3WLU6lK38a1Y5zhAL6s_uZ5AP0TFs,2381
+haiku/rag/qa/openai.py,sha256=dF32sGgVt8mZi5oVxByaeECs9NqLjvDiZnnpJBsrHm8,3968
+haiku/rag/qa/prompts.py,sha256=578LJGZJ0LQ_q7ccyj5hLabtHo8Zcfw5-DiLGN9lC-w,1200
 haiku/rag/store/__init__.py,sha256=hq0W0DAC7ysqhWSP2M2uHX8cbG6kbr-sWHxhq6qQcY0,103
 haiku/rag/store/engine.py,sha256=BeYZRZ08zaYeeu375ysnAL3tGz4roA3GzP7WRNwznCo,2603
 haiku/rag/store/models/__init__.py,sha256=s0E72zneGlowvZrFWaNxHYjOAUjgWdLxzdYsnvNRVlY,88
@@ -29,8 +29,8 @@ haiku/rag/store/repositories/__init__.py,sha256=uIBhxjQh-4o3O-ck8b7BQ58qXQTuJdPv
 haiku/rag/store/repositories/base.py,sha256=cm3VyQXhtxvRfk1uJHpA0fDSxMpYN-mjQmRiDiLsQ68,1008
 haiku/rag/store/repositories/chunk.py,sha256=gik7ZPOK3gCoG6tU1pGueAZBPmJxIb7obYFUhwINrYg,16497
 haiku/rag/store/repositories/document.py,sha256=xpWOpjHFbhVwNJ1gpusEKNY6l_Qyibg9y_bdHCwcfpk,7133
-haiku_rag-0.3.2.dist-info/METADATA,sha256=0A8BVZDp38i_xLznvkrZBq3f3OYtWPtqBx_U2eHRIow,3931
-haiku_rag-0.3.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-haiku_rag-0.3.2.dist-info/entry_points.txt,sha256=G1U3nAkNd5YDYd4v0tuYFbriz0i-JheCsFuT9kIoGCI,48
-haiku_rag-0.3.2.dist-info/licenses/LICENSE,sha256=eXZrWjSk9PwYFNK9yUczl3oPl95Z4V9UXH7bPN46iPo,1065
-haiku_rag-0.3.2.dist-info/RECORD,,
+haiku_rag-0.3.3.dist-info/METADATA,sha256=nDI-sy2F8h7qr9hK1S7VQLOMRcWYP1clxJYxNVB1AaA,4019
+haiku_rag-0.3.3.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+haiku_rag-0.3.3.dist-info/entry_points.txt,sha256=G1U3nAkNd5YDYd4v0tuYFbriz0i-JheCsFuT9kIoGCI,48
+haiku_rag-0.3.3.dist-info/licenses/LICENSE,sha256=eXZrWjSk9PwYFNK9yUczl3oPl95Z4V9UXH7bPN46iPo,1065
+haiku_rag-0.3.3.dist-info/RECORD,,