Update app.py
Browse files
app.py
CHANGED
@@ -6,6 +6,7 @@ import json
|
|
6 |
import random
|
7 |
import uuid
|
8 |
import concurrent.futures
|
|
|
9 |
from datetime import datetime, timedelta
|
10 |
from apscheduler.schedulers.background import BackgroundScheduler
|
11 |
from flask import Flask, request, jsonify, Response, stream_with_context
|
@@ -33,6 +34,10 @@ valid_keys_global = []
|
|
33 |
executor = concurrent.futures.ThreadPoolExecutor(max_workers=20)
|
34 |
model_key_indices = {}
|
35 |
|
|
|
|
|
|
|
|
|
36 |
def get_credit_summary(api_key):
|
37 |
"""
|
38 |
使用 API 密钥获取额度信息。
|
@@ -394,7 +399,19 @@ scheduler.add_job(refresh_models, 'interval', hours=1)
|
|
394 |
|
395 |
@app.route('/')
|
396 |
def index():
|
397 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
398 |
|
399 |
@app.route('/check_tokens', methods=['POST'])
|
400 |
def check_tokens():
|
@@ -616,6 +633,10 @@ def handsome_chat_completions():
|
|
616 |
f"输出的内容: {response_content_replaced}"
|
617 |
)
|
618 |
|
|
|
|
|
|
|
|
|
619 |
return Response(
|
620 |
stream_with_context(generate()),
|
621 |
content_type=response.headers['Content-Type']
|
@@ -678,6 +699,13 @@ def handsome_chat_completions():
|
|
678 |
f"用户的内容: {user_content_replaced}, "
|
679 |
f"输出的内容: {response_content_replaced}"
|
680 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
681 |
return jsonify(response_json)
|
682 |
|
683 |
except requests.exceptions.RequestException as e:
|
|
|
6 |
import random
|
7 |
import uuid
|
8 |
import concurrent.futures
|
9 |
+
import threading
|
10 |
from datetime import datetime, timedelta
|
11 |
from apscheduler.schedulers.background import BackgroundScheduler
|
12 |
from flask import Flask, request, jsonify, Response, stream_with_context
|
|
|
34 |
executor = concurrent.futures.ThreadPoolExecutor(max_workers=20)
|
35 |
model_key_indices = {}
|
36 |
|
37 |
+
request_timestamps = []
|
38 |
+
token_counts = []
|
39 |
+
data_lock = threading.Lock()
|
40 |
+
|
41 |
def get_credit_summary(api_key):
|
42 |
"""
|
43 |
使用 API 密钥获取额度信息。
|
|
|
399 |
|
400 |
@app.route('/')
|
401 |
def index():
|
402 |
+
current_time = time.time()
|
403 |
+
one_minute_ago = current_time - 60
|
404 |
+
|
405 |
+
with data_lock:
|
406 |
+
# Clean up old data
|
407 |
+
while request_timestamps and request_timestamps[0] < one_minute_ago:
|
408 |
+
request_timestamps.pop(0)
|
409 |
+
token_counts.pop(0)
|
410 |
+
|
411 |
+
rpm = len(request_timestamps)
|
412 |
+
tpm = sum(token_counts)
|
413 |
+
|
414 |
+
return jsonify({"rpm": rpm, "tpm": tpm})
|
415 |
|
416 |
@app.route('/check_tokens', methods=['POST'])
|
417 |
def check_tokens():
|
|
|
633 |
f"输出的内容: {response_content_replaced}"
|
634 |
)
|
635 |
|
636 |
+
with data_lock:
|
637 |
+
request_timestamps.append(time.time())
|
638 |
+
token_counts.append(prompt_tokens+completion_tokens)
|
639 |
+
|
640 |
return Response(
|
641 |
stream_with_context(generate()),
|
642 |
content_type=response.headers['Content-Type']
|
|
|
699 |
f"用户的内容: {user_content_replaced}, "
|
700 |
f"输出的内容: {response_content_replaced}"
|
701 |
)
|
702 |
+
with data_lock:
|
703 |
+
request_timestamps.append(time.time())
|
704 |
+
if "prompt_tokens" in response_json["usage"] and "completion_tokens" in response_json["usage"]:
|
705 |
+
token_counts.append(response_json["usage"]["prompt_tokens"] + response_json["usage"]["completion_tokens"])
|
706 |
+
else:
|
707 |
+
token_counts.append(0)
|
708 |
+
|
709 |
return jsonify(response_json)
|
710 |
|
711 |
except requests.exceptions.RequestException as e:
|