Add extension example, replace input_hijack with chat_input_modifier (#3307)

extensions/api/blocking_api.py

@@ -7,10 +7,15 @@ from modules import shared
 from modules.chat import generate_chat_reply
 from modules.LoRA import add_lora_to_model
 from modules.models import load_model, unload_model
-from modules.models_settings import (get_model_settings_from_yamls,
-                                     update_model_parameters)
-from modules.text_generation import (encode, generate_reply,
-                                     stop_everything_event)
+from modules.models_settings import (
+    get_model_settings_from_yamls,
+    update_model_parameters
+)
+from modules.text_generation import (
+    encode,
+    generate_reply,
+    stop_everything_event
+)
 from modules.utils import get_available_models
 
 

extensions/api/streaming_api.py

@@ -2,12 +2,15 @@ import asyncio
 import json
 from threading import Thread
 
-from websockets.server import serve
-
-from extensions.api.util import build_parameters, try_start_cloudflared, with_api_lock
+from extensions.api.util import (
+    build_parameters,
+    try_start_cloudflared,
+    with_api_lock
+)
 from modules import shared
 from modules.chat import generate_chat_reply
 from modules.text_generation import generate_reply
+from websockets.server import serve
 
 PATH = '/api/v1/stream'
 

extensions/api/util.py

@@ -10,7 +10,6 @@ from modules import shared
 from modules.chat import load_character_memoized
 from modules.presets import load_preset_memoized
 
-
 # We use a thread local to store the asyncio lock, so that each thread
 # has its own lock.  This isn't strictly necessary, but it makes it
 # such that if we can support multiple worker threads in the future,