implement streaming for hf-textgen

README.md

@@ -52,3 +52,4 @@ should probably clear the `generation_time` time column in the `prompts` table.
 - Convince Oobabooga to implement concurrent generation
 - Make sure stats work when starting from an empty database
 - Make sure we're correctly canceling requests when the client cancels
+- Implement auth and tokens on the websocket endpoint. Maybe add something to the instruct prompt and the remove it before proxying??

llm_server/config.py

@@ -19,8 +19,8 @@ config_default_vars = {
 config_required_vars = ['token_limit', 'concurrent_gens', 'mode', 'llm_middleware_name']
 
 mode_ui_names = {
-    'oobabooga': ('Text Gen WebUI (ooba)', 'Blocking API url'),
+    'oobabooga': ('Text Gen WebUI (ooba)', 'Blocking API url', 'Streaming API url'),
-    'hf-textgen': ('Text Gen WebUI (ooba)', 'Blocking API url'),
+    'hf-textgen': ('Text Gen WebUI (ooba)', 'Blocking API url', 'Streaming API url'),
 }
 
 

llm_server/database.py

@@ -50,7 +50,7 @@ def log_prompt(ip, token, prompt, response, gen_time, parameters, headers, backe
 
     if not is_error:
         if not response_tokens:
-            response_tokens = len(tokenizer.encode(response))
+            response_tokens = len(tokenizer.encode(response, disallowed_special=()))
     else:
         response_tokens = None
 

llm_server/opts.py

@@ -11,7 +11,7 @@ database_path = './proxy-server.db'
 auth_required = False
 log_prompts = False
 frontend_api_client = ''
-full_client_api = None
+base_client_api = None
 http_host = None
 verify_ssl = True
 show_num_prompts = True

llm_server/routes/v1/__init__.py

@@ -6,18 +6,16 @@ from ... import opts
 bp = Blueprint('v1', __name__)
 
 
-# openai_bp = Blueprint('/v1', __name__)
-
 @bp.before_request
 def before_request():
     if not opts.http_host:
         opts.http_host = request.headers.get("Host")
-    if not opts.full_client_api:
+    if not opts.base_client_api:
-        opts.full_client_api = f'https://{request.headers.get("Host")}/{opts.frontend_api_client.strip("/")}'
+        opts.base_client_api = f'{request.headers.get("Host")}/{opts.frontend_api_client.strip("/")}'
     if request.endpoint != 'v1.get_stats':
         response = require_api_key()
         if response is not None:
             return response
 
 
-from . import generate, info, proxy
+from . import generate, info, proxy, generate_stream

llm_server/routes/v1/generate_stats.py

@@ -81,7 +81,8 @@ def generate_stats():
         },
         'online': online,
         'endpoints': {
-            'blocking': opts.full_client_api,
+            'blocking': f'https://{opts.base_client_api}',
+            'streaming': f'wss://{opts.base_client_api}/v1/stream',
         },
         'queue': {
             'processing': active_gen_workers,

llm_server/routes/v1/generate_stream.py

@@ -0,0 +1,84 @@
+import json
+import time
+
+import requests
+from flask import request
+
+from ..helpers.client import format_sillytavern_err
+from ... import opts
+from ...database import log_prompt
+from ...helpers import indefinite_article
+from ...llm.hf_textgen.generate import prepare_json
+from ...stream import sock
+
+
+@sock.route('/api/v1/stream')  # TODO: use blueprint route???
+def stream(ws):
+    start_time = time.time()
+    if request.headers.get('cf-connecting-ip'):
+        client_ip = request.headers.get('cf-connecting-ip')
+    elif request.headers.get('x-forwarded-for'):
+        client_ip = request.headers.get('x-forwarded-for').split(',')[0]
+    else:
+        client_ip = request.remote_addr
+    token = request.headers.get('X-Api-Key')
+
+    message_num = 0
+    while ws.connected:
+        message = ws.receive()
+        data = json.loads(message)
+
+        if opts.mode == 'hf-textgen':
+            response = requests.post(f'{opts.backend_url}/generate_stream', json=prepare_json(data), stream=True, verify=False)
+
+            # Be extra careful when getting attributes from the response object
+            try:
+                response_status_code = response.status_code
+            except:
+                response_status_code = 0
+
+            details = {}
+            generated_text = ''
+
+            # Iterate over each line in the response
+            for line in response.iter_lines():
+                # Decode the line to a string
+                line = line.decode('utf-8')
+                # If the line starts with 'data:', remove the prefix and parse the remaining string as JSON
+                if line.startswith('data:'):
+                    line = line[5:]
+                    json_data = json.loads(line)
+                    details = json_data.get('details', {})
+                    generated_text = json_data.get('generated_text', '')
+
+                    if json_data.get('error'):
+                        error_type = json_data.get('error_type')
+                        error_type_string = 'returned an error' if opts.mode == 'oobabooga' else f'returned {indefinite_article(error_type)} {error_type} error'
+                        generated_text = format_sillytavern_err(
+                            f'Backend ({opts.mode}) {error_type_string}: {json_data.get("error")}',
+                            f'HTTP CODE {response_status_code}')
+                        ws.send(json.dumps({
+                            'event': 'text_stream',
+                            'message_num': message_num,
+                            'text': generated_text
+                        }))
+                        break
+                    else:
+                        ws.send(json.dumps({
+                            'event': 'text_stream',
+                            'message_num': message_num,
+                            'text': json_data['token']['text']
+                        }))
+                        message_num += 1
+
+        ws.send(json.dumps({
+            'event': 'stream_end',
+            'message_num': message_num
+        }))
+
+        end_time = time.time()
+        elapsed_time = end_time - start_time
+        parameters = data.copy()
+        del parameters['prompt']
+
+        log_prompt(client_ip, token, data['prompt'], generated_text, elapsed_time, parameters, dict(request.headers), response_status_code, response_tokens=details['generated_tokens'])

llm_server/stream.py

@@ -0,0 +1,8 @@
+from flask_sock import Sock
+
+sock = Sock()
+
+
+def init_socketio(app):
+    global sock
+    sock.init_app(app)

requirements.txt

@@ -7,4 +7,5 @@ tiktoken
 gunicorn
 redis
 gevent
-async-timeout
+async-timeout
+flask-socketio

server.py

@@ -17,6 +17,7 @@ from llm_server.routes.queue import start_workers
 from llm_server.routes.stats import SemaphoreCheckerThread, process_avg_gen_time
 from llm_server.routes.v1 import bp
 from llm_server.routes.v1.generate_stats import generate_stats
+from llm_server.stream import init_socketio
 from llm_server.threads import MainBackgroundThread
 
 script_path = os.path.dirname(os.path.realpath(__file__))
@@ -88,6 +89,7 @@ SemaphoreCheckerThread().start()
 app = Flask(__name__)
 cache.init_app(app)
 cache.clear()  # clear redis cache
+init_socketio(app)
 # with app.app_context():
 #     current_app.tokenizer = tiktoken.get_encoding("cl100k_base")
 app.register_blueprint(bp, url_prefix='/api/v1/')
@@ -100,8 +102,8 @@ app.register_blueprint(bp, url_prefix='/api/v1/')
 @app.route('/api')
 @cache.cached(timeout=10, query_string=True)
 def home():
-    if not opts.full_client_api:
+    if not opts.base_client_api:
-        opts.full_client_api = f'https://{request.headers.get("Host")}/{opts.frontend_api_client.strip("/")}'
+        opts.base_client_api = f'{request.headers.get("Host")}/{opts.frontend_api_client.strip("/")}'
     stats = generate_stats()
 
     if not bool(redis.get('backend_online')) or not stats['online']:
@@ -131,10 +133,12 @@ def home():
                            analytics_tracking_code=analytics_tracking_code,
                            info_html=info_html,
                            current_model=running_model,
-                           client_api=opts.full_client_api,
+                           client_api=f'https://{opts.base_client_api}',
+                           ws_client_api=f'wss://{opts.base_client_api}/v1/stream',
                            estimated_wait=estimated_wait_sec,
                            mode_name=mode_ui_names[opts.mode][0],
                            api_input_textbox=mode_ui_names[opts.mode][1],
+                           streaming_input_textbox=mode_ui_names[opts.mode][2],
                            context_size=opts.context_size,
                            stats_json=json.dumps(stats, indent=4, ensure_ascii=False),
                            extra_info=hf_textget_info if opts.mode == 'hf-textgen' else '',

templates/home.html

@@ -71,6 +71,7 @@
     <div class="info-box">
         <p><strong>Current Model:</strong> <span id="model">{{ current_model }}</span></p>
         <p><strong>Client API URL:</strong> {{ client_api }}</p>
+        <p><strong>Streaming API URL:</strong> {{ ws_client_api }}</p>
         <p><strong>Estimated Wait Time:</strong> <span id="estimatedWait">{{ estimated_wait }}</span></p>
         {{ info_html|safe }}
     </div>
@@ -83,6 +84,10 @@
             <ol>
                 <li>Set your API type to <kbd>{{ mode_name }}</kbd></li>
                 <li>Enter <kbd>{{ client_api }}</kbd> in the <kbd>{{ api_input_textbox }}</kbd> textbox.</li>
+                <li>Enter <kbd>{{ ws_client_api }}</kbd> in the <kbd>{{ streaming_input_textbox }}</kbd> textbox.</li>
+                <li>If using a token, check the <kbd>Mancer AI</kbd> checkbox and enter your token in the <kbd>Mancer
+                    API key</kbd> textbox.
+                </li>
                 <li>Click <kbd>Connect</kbd> to test the connection.</li>
                 <li>Open your preset config and set <kbd>Context Size</kbd> to {{ context_size }}.</li>
                 <li>Follow this guide to get set up: <a href="https://rentry.org/freellamas" target="_blank">rentry.org/freellamas</a>