add support for default target (#111)

* add support for default target

* add more fixes

arch/arch_config_schema.yaml

@@ -67,6 +67,8 @@ properties:
           type: boolean
         description:
           type: string
+        auto_llm_dispatch_on_response:
+          type: boolean
         parameters:
           type: array
           items:

arch/src/stream_context.rs

@@ -39,6 +39,7 @@ enum ResponseHandlerType {
     FunctionCall,
     ZeroShotIntent,
     ArchGuard,
+    DefaultTarget,
 }
 
 pub struct CallContext {
@@ -179,12 +180,16 @@ impl StreamContext {
 
         let prompt_target_names = prompt_targets
             .iter()
+            // exclude default target
+            .filter(|(_, prompt_target)| !prompt_target.default.unwrap_or(false))
             .map(|(name, _)| name.clone())
             .collect();
 
         let similarity_scores: Vec<(String, f64)> = prompt_targets
             .iter()
-            .map(|(prompt_name, _prompt_target)| {
+            // exclude default prompt target
+            .filter(|(_, prompt_target)| !prompt_target.default.unwrap_or(false))
+            .map(|(prompt_name, _)| {
                 let default_embeddings = HashMap::new();
                 let pte = prompt_target_embeddings
                     .get(prompt_name)
@@ -331,34 +336,84 @@ impl StreamContext {
 
         // check to ensure that the prompt target similarity score is above the threshold
         if prompt_target_similarity_score < prompt_target_intent_matching_threshold
-            && !arch_assistant
+            || arch_assistant
         {
+            debug!("intent score is low or arch assistant is handling the conversation");
             // if arch fc responded to the user message, then we don't need to check the similarity score
             // it may be that arch fc is handling the conversation for parameter collection
             if arch_assistant {
                 info!("arch assistant is handling the conversation");
             } else {
-                info!(
-                    "prompt target below limit: {:.3}, threshold: {:.3}, continue conversation with user",
-                    prompt_target_similarity_score,
-                    prompt_target_intent_matching_threshold
-                );
+                debug!("checking for default prompt target");
+                if let Some(default_prompt_target) = self
+                    .prompt_targets
+                    .read()
+                    .unwrap()
+                    .values()
+                    .find(|pt| pt.default.unwrap_or(false))
+                {
+                    debug!("default prompt target found");
+                    let endpoint = default_prompt_target.endpoint.clone().unwrap();
+                    let upstream_path: String = endpoint.path.unwrap_or(String::from("/"));
+
+                    let upstream_endpoint = endpoint.name;
+                    let mut params = HashMap::new();
+                    params.insert(
+                        ARCH_MESSAGES_KEY.to_string(),
+                        callout_context.request_body.messages.clone(),
+                    );
+                    let arch_messages_json = serde_json::to_string(&params).unwrap();
+                    debug!("no prompt target found with similarity score above threshold, using default prompt target");
+                    let token_id = match self.dispatch_http_call(
+                        &upstream_endpoint,
+                        vec![
+                            (":method", "POST"),
+                            (":path", &upstream_path),
+                            (":authority", &upstream_endpoint),
+                            ("content-type", "application/json"),
+                            ("x-envoy-max-retries", "3"),
+                            (
+                                "x-envoy-upstream-rq-timeout-ms",
+                                ARCH_FC_REQUEST_TIMEOUT_MS.to_string().as_str(),
+                            ),
+                        ],
+                        Some(arch_messages_json.as_bytes()),
+                        vec![],
+                        Duration::from_secs(5),
+                    ) {
+                        Ok(token_id) => token_id,
+                        Err(e) => {
+                            let error_msg =
+                                format!("Error dispatching HTTP call for default-target: {:?}", e);
+                            return self
+                                .send_server_error(error_msg, Some(StatusCode::BAD_REQUEST));
+                        }
+                    };
+
+                    self.metrics.active_http_calls.increment(1);
+                    callout_context.response_handler_type = ResponseHandlerType::DefaultTarget;
+                    callout_context.prompt_target_name = Some(default_prompt_target.name.clone());
+                    if self.callouts.insert(token_id, callout_context).is_some() {
+                        panic!("duplicate token_id")
+                    }
+                    return;
+                }
                 self.resume_http_request();
                 return;
             }
         }
 
-        let prompt_target = self
-            .prompt_targets
-            .read()
-            .unwrap()
-            .get(&prompt_target_name)
-            .unwrap()
-            .clone();
+        let prompt_target = match self.prompt_targets.read().unwrap().get(&prompt_target_name) {
+            Some(prompt_target) => prompt_target.clone(),
+            None => {
+                return self.send_server_error(
+                    format!("Prompt target not found: {}", prompt_target_name),
+                    None,
+                );
+            }
+        };
 
         info!("prompt_target name: {:?}", prompt_target_name);
-
-        //TODO: handle default function resolver type
         let mut chat_completion_tools: Vec<ChatCompletionTool> = Vec::new();
         for pt in self.prompt_targets.read().unwrap().values() {
             // only extract entity names
@@ -761,6 +816,83 @@ impl StreamContext {
             )
         }
     }
+
+    fn default_target_handler(&self, body: Vec<u8>, callout_context: CallContext) {
+        let prompt_target = self
+            .prompt_targets
+            .read()
+            .unwrap()
+            .get(callout_context.prompt_target_name.as_ref().unwrap())
+            .unwrap()
+            .clone();
+        debug!(
+            "response received for default target: {}",
+            prompt_target.name
+        );
+        // check if the default target should be dispatched to the LLM provider
+        if !prompt_target.auto_llm_dispatch_on_response.unwrap_or(false) {
+            let default_target_response_str = String::from_utf8(body).unwrap();
+            debug!(
+                "sending response back to developer: {}",
+                default_target_response_str
+            );
+            self.send_http_response(
+                StatusCode::OK.as_u16().into(),
+                vec![("Powered-By", "Katanemo")],
+                Some(default_target_response_str.as_bytes()),
+            );
+            // self.resume_http_request();
+            return;
+        }
+        debug!("default_target: sending api response to default llm");
+        let chat_completions_resp: ChatCompletionsResponse = match serde_json::from_slice(&body) {
+            Ok(chat_completions_resp) => chat_completions_resp,
+            Err(e) => {
+                return self.send_server_error(
+                    format!("Error deserializing default target response: {:?}", e),
+                    None,
+                );
+            }
+        };
+        let api_resp = chat_completions_resp.choices[0]
+            .message
+            .content
+            .as_ref()
+            .unwrap();
+        let mut messages = callout_context.request_body.messages;
+
+        // add system prompt
+        match prompt_target.system_prompt.as_ref() {
+            None => {}
+            Some(system_prompt) => {
+                let system_prompt_message = Message {
+                    role: SYSTEM_ROLE.to_string(),
+                    content: Some(system_prompt.clone()),
+                    model: None,
+                    tool_calls: None,
+                };
+                messages.push(system_prompt_message);
+            }
+        }
+
+        messages.push(Message {
+            role: USER_ROLE.to_string(),
+            content: Some(api_resp.clone()),
+            model: None,
+            tool_calls: None,
+        });
+        let chat_completion_request = ChatCompletionsRequest {
+            model: GPT_35_TURBO.to_string(),
+            messages,
+            tools: None,
+            stream: callout_context.request_body.stream,
+            stream_options: callout_context.request_body.stream_options,
+        };
+        let json_resp = serde_json::to_string(&chat_completion_request).unwrap();
+        debug!("sending response back to default llm: {}", json_resp);
+        self.set_http_request_body(0, json_resp.len(), json_resp.as_bytes());
+        self.resume_http_request();
+    }
 }
 
 // HttpContext is the trait that allows the Rust code to interact with HTTP objects.
@@ -1067,6 +1199,9 @@ impl Context for StreamContext {
                     self.function_call_response_handler(body, callout_context)
                 }
                 ResponseHandlerType::ArchGuard => self.arch_guard_handler(body, callout_context),
+                ResponseHandlerType::DefaultTarget => {
+                    self.default_target_handler(body, callout_context)
+                }
             }
         } else {
             self.send_server_error(

demos/function_calling/api_server/Dockerfile

@@ -16,4 +16,4 @@ COPY --from=builder /runtime /usr/local
 COPY /app /app
 WORKDIR /app
 
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "80"]
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "80", "--log-level", "info"]

demos/function_calling/api_server/app/main.py

@@ -1,12 +1,13 @@
+import json
 import random
 from fastapi import FastAPI, Response
 from datetime import datetime, date, timedelta, timezone
 import logging
 from pydantic import BaseModel
-import pytz
 
-logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
-logger = logging.getLogger(__name__)
+
+logger = logging.getLogger('uvicorn.error')
+logger.setLevel(logging.INFO)
 
 app = FastAPI()
 
@@ -58,18 +59,28 @@ async def insurance_claim_details(req: InsuranceClaimDetailsRequest, res: Respon
 
     return claim_details
 
-@app.get("/current_time")
-async def current_time(timezone: str):
-    tz = None
-    try:
-      timezone.strip('"')
-      tz = pytz.timezone(timezone)
-    except pytz.exceptions.UnknownTimeZoneError:
-        return {
-            "error": "Invalid timezone: {}".format(timezone)
-        }
-    current_time = datetime.now(tz)
-    return {
-        "timezone": timezone,
-        "current_time": current_time.strftime("%Y-%m-%d %H:%M:%S %Z")
-    }
+
+class DefaultTargetRequest(BaseModel):
+   arch_messages: list
+
+@app.post("/default_target")
+async def default_target(req: DefaultTargetRequest, res: Response):
+    logger.info(f"Received arch_messages: {req.arch_messages}")
+    resp = {
+            "choices": [
+              {
+                "message": {
+                  "role": "assistant",
+                  "content": "hello world from api server"
+                },
+                "finish_reason": "completed",
+                "index": 0
+              }
+            ],
+            "model": "api_server",
+            "usage": {
+              "completion_tokens": 0
+            }
+          }
+    logger.info(f"sending response: {json.dumps(resp)}")
+    return resp

demos/function_calling/api_server/requirements.txt

@@ -1,3 +1,3 @@
 fastapi
 uvicorn
-pytz
+pyyaml

demos/function_calling/arch_config.yaml

@@ -47,21 +47,6 @@ prompt_targets:
       - Use farenheight for temperature
       - Use miles per hour for wind speed
 
-  - name: system_time
-    description: This function provides the current system time.
-    parameters:
-      - name: timezone
-        description: The city for which the weather forecast is requested.
-        default: US/Pacific
-        type: string
-    endpoint:
-      name: api_server
-      path: /current_time
-    system_prompt: |
-      You are a helpful system time provider. Use system time data that is provided to you. Please following following guidelines when responding to user queries:
-      - Use 12 hour time format
-      - Use AM/PM for time
-
   - name: insurance_claim_details
     description: This function resolver provides insurance claim details for a given policy number.
     parameters:
@@ -80,6 +65,18 @@ prompt_targets:
       You are a helpful insurance claim details provider. Use insurance claim data that is provided to you. Please following following guidelines when responding to user queries:
       - Use policy number to retrieve insurance claim details
 
+  - name: default_target
+    default: true
+    description: This is the default target for all unmatched prompts.
+    endpoint:
+      name: api_server
+      path: /default_target
+    system_prompt: |
+      You are a helpful assistant. Use the information that is provided to you.
+    # if it is set to false arch will send response that it received from this prompt target to the user
+    # if true arch will forward the response to the default LLM
+    auto_llm_dispatch_on_response: true
+
 ratelimits:
   - provider: gpt-3.5-turbo
     selector:

model_server/Dockerfile

@@ -1,41 +1,25 @@
-FROM python:3.10 AS base
+FROM python:3.10 AS builder
 
-#
-# builder
-#
-FROM base AS builder
-
-WORKDIR /src
-RUN pip install --upgrade pip
-
-# Install git (needed for cloning the repository)
-RUN apt-get update && apt-get install -y git && apt-get clean
-
-COPY requirements.txt /src/
-
-RUN pip install --prefix=/runtime --force-reinstall -r requirements.txt
-
-COPY . /src
-
-#
-# output
-#
+COPY requirements.txt .
+RUN pip install --prefix=/runtime -r requirements.txt
 
 FROM python:3.10-slim AS output
 
+# curl is needed for health check in docker-compose
+RUN apt-get update && apt-get install -y curl && apt-get clean && rm -rf /var/lib/apt/lists/*
+
+COPY --from=builder /runtime /usr/local
+
+WORKDIR /src
+
 # specify list of models that will go into the image as a comma separated list
 # following models have been tested to work with this image
 # "sentence-transformers/all-MiniLM-L6-v2,sentence-transformers/all-mpnet-base-v2,thenlper/gte-base,thenlper/gte-large,thenlper/gte-small"
 ENV MODELS="BAAI/bge-large-en-v1.5"
 
-COPY --from=builder /runtime /usr/local
-
-COPY ./ /app
-WORKDIR /app
-
-RUN apt-get update && apt-get install -y \
-  curl \
-  && rm -rf /var/lib/apt/lists/*
+COPY ./app ./app
+COPY ./guard_model_config.yaml .
+COPY ./openai_params.yaml .
 
 # comment it out for now as we don't want to download the model every time we build the image
 # we will mount host cache to docker image to avoid downloading the model every time

model_server/app/arch_fc/arch_fc.py

@@ -9,6 +9,10 @@ import yaml
 from openai import OpenAI
 import os
 
+logging.basicConfig(
+    level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s"
+)
+logger = logging.getLogger(__name__)
 
 with open("openai_params.yaml") as f:
     params = yaml.safe_load(f)
@@ -20,7 +24,6 @@ mode = os.getenv("MODE", "cloud")
 if mode not in ["cloud", "local-gpu", "local-cpu"]:
     raise ValueError(f"Invalid mode: {mode}")
 arch_api_key = os.getenv("ARCH_API_KEY", "vllm")
-logger = logging.getLogger("uvicorn.error")
 
 handler = None
 if ollama_model.startswith("Arch"):
@@ -28,17 +31,12 @@ if ollama_model.startswith("Arch"):
 else:
     handler = BoltHandler()
 
-
-# app = FastAPI()
-
 if mode == "cloud":
     client = OpenAI(
         base_url=fc_url,
         api_key="EMPTY",
     )
-    models = client.models.list()
-    model = models.data[0].id
-    chosen_model = model
+    chosen_model = "fc-cloud"
     endpoint = fc_url
 else:
     client = OpenAI(
@@ -47,12 +45,12 @@ else:
     )
     chosen_model = ollama_model
     endpoint = ollama_endpoint
+
 logger.info(f"serving mode: {mode}")
 logger.info(f"using model: {chosen_model}")
 logger.info(f"using endpoint: {endpoint}")
 
 
-
 async def chat_completion(req: ChatMessage, res: Response):
     logger.info("starting request")
     tools_encoded = handler._format_system(req.tools)