Merge branch 'open-webui:main' into routellm-pipeline

examples/filters/langfuse_filter_pipeline.py

@@ -11,6 +11,7 @@ requirements: langfuse
 from typing import List, Optional
 from schemas import OpenAIChatMessage
 import os
+import uuid
 
 from utils.pipelines.main import get_last_user_message, get_last_assistant_message
 from pydantic import BaseModel
@@ -20,64 +21,36 @@ from langfuse.api.resources.commons.errors.unauthorized_error import Unauthorize
 
 class Pipeline:
     class Valves(BaseModel):
-        # List target pipeline ids (models) that this filter will be connected to.
-        # If you want to connect this filter to all pipelines, you can set pipelines to ["*"]
-        # e.g. ["llama3:latest", "gpt-3.5-turbo"]
         pipelines: List[str] = []
-
-        # Assign a priority level to the filter pipeline.
-        # The priority level determines the order in which the filter pipelines are executed.
-        # The lower the number, the higher the priority.
         priority: int = 0
-
-        # Valves
         secret_key: str
         public_key: str
         host: str
 
     def __init__(self):
-        # Pipeline filters are only compatible with Open WebUI
-        # You can think of filter pipeline as a middleware that can be used to edit the form data before it is sent to the OpenAI API.
         self.type = "filter"
-
-        # Optionally, you can set the id and name of the pipeline.
-        # Best practice is to not specify the id so that it can be automatically inferred from the filename, so that users can install multiple versions of the same pipeline.
-        # The identifier must be unique across all pipelines.
-        # The identifier must be an alphanumeric string that can include underscores or hyphens. It cannot contain spaces, special characters, slashes, or backslashes.
-        # self.id = "langfuse_filter_pipeline"
         self.name = "Langfuse Filter"
-
-        # Initialize
         self.valves = self.Valves(
             **{
-                "pipelines": ["*"],  # Connect to all pipelines
+                "pipelines": ["*"],
                 "secret_key": os.getenv("LANGFUSE_SECRET_KEY", "your-secret-key-here"),
                 "public_key": os.getenv("LANGFUSE_PUBLIC_KEY", "your-public-key-here"),
                 "host": os.getenv("LANGFUSE_HOST", "https://cloud.langfuse.com"),
             }
         )
-
         self.langfuse = None
         self.chat_generations = {}
-        pass
 
     async def on_startup(self):
-        # This function is called when the server is started.
         print(f"on_startup:{__name__}")
         self.set_langfuse()
-        pass
 
     async def on_shutdown(self):
-        # This function is called when the server is stopped.
         print(f"on_shutdown:{__name__}")
         self.langfuse.flush()
-        pass
 
     async def on_valves_updated(self):
-        # This function is called when the valves are updated.
-
         self.set_langfuse()
-        pass
 
     def set_langfuse(self):
         try:
@@ -97,6 +70,22 @@ class Pipeline:
 
     async def inlet(self, body: dict, user: Optional[dict] = None) -> dict:
         print(f"inlet:{__name__}")
+        print(f"Received body: {body}")
+        print(f"User: {user}")
+
+        # Check for presence of required keys and generate chat_id if missing
+        if "chat_id" not in body:
+            unique_id = f"SYSTEM MESSAGE {uuid.uuid4()}"
+            body["chat_id"] = unique_id
+            print(f"chat_id was missing, set to: {unique_id}")
+
+        required_keys = ["model", "messages"]
+        missing_keys = [key for key in required_keys if key not in body]
+        
+        if missing_keys:
+            error_message = f"Error: Missing keys in the request body: {', '.join(missing_keys)}"
+            print(error_message)
+            raise ValueError(error_message)
 
         trace = self.langfuse.trace(
             name=f"filter:{__name__}",
@@ -128,9 +117,6 @@ class Pipeline:
         user_message = get_last_user_message(body["messages"])
         generated_message = get_last_assistant_message(body["messages"])
 
-        # Update usage cost based on the length of the input and output messages
-        # Below does not reflect the actual cost of the API
-        # You can adjust the cost based on your requirements
         generation.end(
             output=generated_message,
             usage={

examples/pipelines/providers/google_manifold_pipeline.py

@@ -2,7 +2,7 @@
 title: Google GenAI Manifold Pipeline
 author: Marc Lopez (refactor by justinh-rahb)
 date: 2024-06-06
-version: 1.2
+version: 1.3
 license: MIT
 description: A pipeline for generating text using Google's GenAI models in Open-WebUI.
 requirements: google-generativeai
@@ -128,10 +128,13 @@ class Pipeline:
                             "parts": [{"text": message["content"]}]
                         })
             
-            if system_message:
+            if "gemini-1.5" in model_id:
-                contents.insert(0, {"role": "user", "parts": [{"text": f"System: {system_message}"}]})
+                model = genai.GenerativeModel(model_name=model_id, system_instruction=system_message)
+            else:
+                if system_message:
+                    contents.insert(0, {"role": "user", "parts": [{"text": f"System: {system_message}"}]})
                 
-            model = genai.GenerativeModel(model_name=model_id)
+                model = genai.GenerativeModel(model_name=model_id)
 
             generation_config = GenerationConfig(
                 temperature=body.get("temperature", 0.7),

examples/pipelines/providers/litellm_manifold_pipeline.py

@@ -2,7 +2,7 @@
 title: LiteLLM Manifold Pipeline
 author: open-webui
 date: 2024-05-30
-version: 1.0
+version: 1.0.1
 license: MIT
 description: A manifold pipeline that uses LiteLLM.
 """
@@ -46,12 +46,15 @@ class Pipeline:
                 "LITELLM_PIPELINE_DEBUG": os.getenv("LITELLM_PIPELINE_DEBUG", False),
             }
         )
-        self.pipelines = []
+        # Get models on initialization
+        self.pipelines = self.get_litellm_models()
         pass
 
     async def on_startup(self):
         # This function is called when the server is started.
         print(f"on_startup:{__name__}")
+        # Get models on startup
+        self.pipelines = self.get_litellm_models()
         pass
 
     async def on_shutdown(self):
@@ -85,7 +88,7 @@ class Pipeline:
                     for model in models["data"]
                 ]
             except Exception as e:
-                print(f"Error: {e}")
+                print(f"Error fetching models from LiteLLM: {e}")
                 return [
                     {
                         "id": "error",
@@ -93,6 +96,7 @@ class Pipeline:
                     },
                 ]
         else:
+            print("LITELLM_BASE_URL not set. Please configure it in the valves.")
             return []
 
     def pipe(

examples/pipelines/providers/perplexity_manifold_pipeline.py

@@ -3,6 +3,9 @@ from pydantic import BaseModel
 import os
 import requests
 
+from utils.pipelines.main import pop_system_message
+
+
 class Pipeline:
     class Valves(BaseModel):
         PERPLEXITY_API_BASE_URL: str = "https://api.perplexity.ai"
@@ -26,14 +29,28 @@ class Pipeline:
 
         # List of models
         self.pipelines = [
-            {"id": "llama-3-sonar-large-32k-online", "name": "Llama 3 Sonar Large 32K Online"},
+            {
-            {"id": "llama-3-sonar-small-32k-online", "name": "Llama 3 Sonar Small 32K Online"},
+                "id": "llama-3.1-sonar-large-128k-online",
-            {"id": "llama-3-sonar-large-32k-chat", "name": "Llama 3 Sonar Large 32K Chat"},
+                "name": "Llama 3.1 Sonar Large 128k Online"
-            {"id": "llama-3-sonar-small-32k-chat", "name": "Llama 3 Sonar Small 32K Chat"},
+            },
-            {"id": "llama-3-8b-instruct", "name": "Llama 3 8B Instruct"},
+            {
-            {"id": "llama-3-70b-instruct", "name": "Llama 3 70B Instruct"},
+                "id": "llama-3.1-sonar-small-128k-online",
-            {"id": "mixtral-8x7b-instruct", "name": "Mixtral 8x7B Instruct"},
+                "name": "Llama 3.1 Sonar Small 128k Online"
-            {"id": "related", "name": "Related"}
+            },
+            {
+                "id": "llama-3.1-sonar-large-128k-chat",
+                "name": "Llama 3.1 Sonar Large 128k Chat"
+            },
+            {
+                "id": "llama-3.1-sonar-small-128k-chat",
+                "name": "Llama 3.1 Sonar Small 128k Chat"
+            },
+            {
+                "id": "llama-3.1-8b-instruct", "name": "Llama 3.1 8B Instruct"
+            },
+            {
+                "id": "llama-3.1-70b-instruct", "name": "Llama 3.1 70B Instruct"
+            }
         ]
         pass
 
@@ -59,6 +76,12 @@ class Pipeline:
         # This is where you can add your custom pipelines like RAG.
         print(f"pipe:{__name__}")
 
+        system_message, messages = pop_system_message(messages)
+        system_prompt = "You are a helpful assistant."
+        if system_message is not None:
+            system_prompt = system_message["content"]
+
+        print(system_prompt)
         print(messages)
         print(user_message)
 
@@ -71,8 +94,8 @@ class Pipeline:
         payload = {
             "model": model_id,
             "messages": [
-                {"role": "system", "content": "Be precise and concise."},
+                {"role": "system", "content": system_prompt},
-                {"role": "user", "content": user_message}
+                *messages
             ],
             "stream": body.get("stream", True),
             "return_citations": True,
@@ -124,17 +147,21 @@ class Pipeline:
         except Exception as e:
             return f"Error: {e}"
 
+
 if __name__ == "__main__":
     import argparse
 
     parser = argparse.ArgumentParser(description="Perplexity API Client")
-    parser.add_argument("--api-key", type=str, required=True, help="API key for Perplexity")
+    parser.add_argument("--api-key", type=str, required=True,
-    parser.add_argument("--prompt", type=str, required=True, help="Prompt to send to the Perplexity API")
+                        help="API key for Perplexity")
+    parser.add_argument("--prompt", type=str, required=True,
+                        help="Prompt to send to the Perplexity API")
 
     args = parser.parse_args()
 
     pipeline = Pipeline()
     pipeline.valves.PERPLEXITY_API_KEY = args.api_key
-    response = pipeline.pipe(user_message=args.prompt, model_id="llama-3-sonar-large-32k-online", messages=[], body={"stream": False})
+    response = pipeline.pipe(
+        user_message=args.prompt, model_id="llama-3-sonar-large-32k-online", messages=[], body={"stream": False})
 
     print("Response:", response)

main.py

@@ -26,6 +26,7 @@ import time
 import json
 import uuid
 import sys
+import subprocess
 
 
 from config import API_KEY, PIPELINES_DIR
@@ -105,12 +106,45 @@ def get_all_pipelines():
 
     return pipelines
 
+def parse_frontmatter(content):
+    frontmatter = {}
+    for line in content.split('\n'):
+        if ':' in line:
+            key, value = line.split(':', 1)
+            frontmatter[key.strip().lower()] = value.strip()
+    return frontmatter
+
+def install_frontmatter_requirements(requirements):
+    if requirements:
+        req_list = [req.strip() for req in requirements.split(',')]
+        for req in req_list:
+            print(f"Installing requirement: {req}")
+            subprocess.check_call([sys.executable, "-m", "pip", "install", req])
+    else:
+        print("No requirements found in frontmatter.")
 
 async def load_module_from_path(module_name, module_path):
-    spec = importlib.util.spec_from_file_location(module_name, module_path)
-    module = importlib.util.module_from_spec(spec)
 
     try:
+        # Read the module content
+        with open(module_path, 'r') as file:
+            content = file.read()
+
+        # Parse frontmatter
+        frontmatter = {}
+        if content.startswith('"""'):
+            end = content.find('"""', 3)
+            if end != -1:
+                frontmatter_content = content[3:end]
+                frontmatter = parse_frontmatter(frontmatter_content)
+
+        # Install requirements if specified
+        if 'requirements' in frontmatter:
+            install_frontmatter_requirements(frontmatter['requirements'])
+
+        # Load the module
+        spec = importlib.util.spec_from_file_location(module_name, module_path)
+        module = importlib.util.module_from_spec(spec)
         spec.loader.exec_module(module)
         print(f"Loaded module: {module.__name__}")
         if hasattr(module, "Pipeline"):