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

examples/pipelines/providers/anthropic_manifold_pipeline.py

@@ -1,21 +1,20 @@
 """
 title: Anthropic Manifold Pipeline
-author: justinh-rahb
+author: justinh-rahb, sriparashiva
 date: 2024-06-20
-version: 1.3
+version: 1.4
 license: MIT
 description: A pipeline for generating text and processing images using the Anthropic API.
-requirements: requests, anthropic
+requirements: requests, sseclient-py
 environment_variables: ANTHROPIC_API_KEY
 """
 
 import os
-from anthropic import Anthropic, RateLimitError, APIStatusError, APIConnectionError
+import requests
-
+import json
-from schemas import OpenAIChatMessage
 from typing import List, Union, Generator, Iterator
 from pydantic import BaseModel
-import requests
+import sseclient
 
 from utils.pipelines.main import pop_system_message
 
@@ -32,7 +31,15 @@ class Pipeline:
         self.valves = self.Valves(
             **{"ANTHROPIC_API_KEY": os.getenv("ANTHROPIC_API_KEY", "your-api-key-here")}
         )
-        self.client = Anthropic(api_key=self.valves.ANTHROPIC_API_KEY)
+        self.url = 'https://api.anthropic.com/v1/messages'
+        self.update_headers()
+
+    def update_headers(self):
+        self.headers = {
+            'anthropic-version': '2023-06-01',
+            'content-type': 'application/json',
+            'x-api-key': self.valves.ANTHROPIC_API_KEY
+        }
 
     def get_anthropic_models(self):
         return [
@@ -51,8 +58,7 @@ class Pipeline:
         pass
 
     async def on_valves_updated(self):
-        self.client = Anthropic(api_key=self.valves.ANTHROPIC_API_KEY)
+        self.update_headers()
-        pass
 
     def pipelines(self) -> List[dict]:
         return self.get_anthropic_models()
@@ -131,21 +137,38 @@ class Pipeline:
             }
 
             if body.get("stream", False):
-                return self.stream_response(model_id, payload)
+                return self.stream_response(payload)
             else:
-                return self.get_completion(model_id, payload)
+                return self.get_completion(payload)
-        except (RateLimitError, APIStatusError, APIConnectionError) as e:
+        except Exception as e:
             return f"Error: {e}"
 
-    def stream_response(self, model_id: str, payload: dict) -> Generator:
+    def stream_response(self, payload: dict) -> Generator:
-        stream = self.client.messages.create(**payload)
+        response = requests.post(self.url, headers=self.headers, json=payload, stream=True)
 
-        for chunk in stream:
+        if response.status_code == 200:
-            if chunk.type == "content_block_start":
+            client = sseclient.SSEClient(response)
-                yield chunk.content_block.text
+            for event in client.events():
-            elif chunk.type == "content_block_delta":
+                try:
-                yield chunk.delta.text
+                    data = json.loads(event.data)
+                    if data["type"] == "content_block_start":
+                        yield data["content_block"]["text"]
+                    elif data["type"] == "content_block_delta":
+                        yield data["delta"]["text"]
+                    elif data["type"] == "message_stop":
+                        break
+                except json.JSONDecodeError:
+                    print(f"Failed to parse JSON: {event.data}")
+                except KeyError as e:
+                    print(f"Unexpected data structure: {e}")
+                    print(f"Full data: {data}")
+        else:
+            raise Exception(f"Error: {response.status_code} - {response.text}")
 
-    def get_completion(self, model_id: str, payload: dict) -> str:
+    def get_completion(self, payload: dict) -> str:
-        response = self.client.messages.create(**payload)
+        response = requests.post(self.url, headers=self.headers, json=payload)
-        return response.content[0].text
+        if response.status_code == 200:
+            res = response.json()
+            return res["content"][0]["text"] if "content" in res and res["content"] else ""
+        else:
+            raise Exception(f"Error: {response.status_code} - {response.text}")

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: 1.1
+version: 1.2
 license: MIT
 description: A pipeline for generating text using Google's GenAI models in Open-WebUI.
 requirements: google-generativeai
@@ -12,7 +12,7 @@ environment_variables: GOOGLE_API_KEY
 from typing import List, Union, Iterator
 import os
 
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
 
 import google.generativeai as genai
 from google.generativeai.types import GenerationConfig
@@ -25,13 +25,17 @@ class Pipeline:
         """Options to change from the WebUI"""
 
         GOOGLE_API_KEY: str = ""
+        USE_PERMISSIVE_SAFETY: bool = Field(default=False)
 
     def __init__(self):
         self.type = "manifold"
         self.id = "google_genai"
         self.name = "Google: "
 
-        self.valves = self.Valves(**{"GOOGLE_API_KEY": os.getenv("GOOGLE_API_KEY", "")})
+        self.valves = self.Valves(**{
+            "GOOGLE_API_KEY": os.getenv("GOOGLE_API_KEY", ""),
+            "USE_PERMISSIVE_SAFETY": False
+        })
         self.pipelines = []
 
         genai.configure(api_key=self.valves.GOOGLE_API_KEY)
@@ -41,6 +45,8 @@ class Pipeline:
         """This function is called when the server is started."""
 
         print(f"on_startup:{__name__}")
+        genai.configure(api_key=self.valves.GOOGLE_API_KEY)
+        self.update_pipelines()
 
     async def on_shutdown(self) -> None:
         """This function is called when the server is stopped."""
@@ -135,7 +141,15 @@ class Pipeline:
                 stop_sequences=body.get("stop", []),
             )
 
-            safety_settings = body.get("safety_settings")
+            if self.valves.USE_PERMISSIVE_SAFETY:
+                safety_settings = {
+                    genai.types.HarmCategory.HARM_CATEGORY_HARASSMENT: genai.types.HarmBlockThreshold.BLOCK_NONE,
+                    genai.types.HarmCategory.HARM_CATEGORY_HATE_SPEECH: genai.types.HarmBlockThreshold.BLOCK_NONE,
+                    genai.types.HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: genai.types.HarmBlockThreshold.BLOCK_NONE,
+                    genai.types.HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: genai.types.HarmBlockThreshold.BLOCK_NONE,
+                }
+            else:
+                safety_settings = body.get("safety_settings")
 
             response = model.generate_content(
                 contents,