XThomasBU
commited on
Commit
·
ab2ff67
1
Parent(s):
7bce799
updates
Browse files- code/main.py +33 -11
- code/modules/chat_processor/helpers.py +16 -28
- code/modules/config/constants.py +2 -2
code/main.py
CHANGED
@@ -21,6 +21,7 @@ from modules.chat_processor.helpers import (
|
|
21 |
get_time,
|
22 |
check_user_cooldown,
|
23 |
reset_tokens_for_user,
|
|
|
24 |
)
|
25 |
import copy
|
26 |
from typing import Optional
|
@@ -55,6 +56,24 @@ async def setup_data_layer():
|
|
55 |
return data_layer
|
56 |
|
57 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
58 |
class Chatbot:
|
59 |
def __init__(self, config):
|
60 |
"""
|
@@ -367,6 +386,10 @@ class Chatbot:
|
|
367 |
|
368 |
# update user info with last message time
|
369 |
user = cl.user_session.get("user")
|
|
|
|
|
|
|
|
|
370 |
|
371 |
print("\n\n User Tokens Left: ", user.metadata["tokens_left"])
|
372 |
|
@@ -406,6 +429,15 @@ class Chatbot:
|
|
406 |
user.metadata["in_cooldown"] = True
|
407 |
await update_user_info(user)
|
408 |
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
409 |
|
410 |
user.metadata["in_cooldown"] = False
|
411 |
|
@@ -486,18 +518,8 @@ class Chatbot:
|
|
486 |
print("Time taken to generate questions: ", time.time() - start_time)
|
487 |
|
488 |
# # update user info with token count
|
489 |
-
|
490 |
-
user.metadata["tokens_left"] = user.metadata["tokens_left"] - token_count
|
491 |
-
user.metadata["all_time_tokens_allocated"] = (
|
492 |
-
user.metadata["all_time_tokens_allocated"] - token_count
|
493 |
-
)
|
494 |
-
await reset_tokens_for_user(user) # regenerate tokens for the user
|
495 |
-
user.metadata["last_message_time"] = get_time()
|
496 |
-
await update_user_info(user)
|
497 |
|
498 |
-
tokens_left = user.metadata["tokens_left"]
|
499 |
-
if tokens_left < 0:
|
500 |
-
tokens_left = 0
|
501 |
answer_with_sources += (
|
502 |
'\n\n<footer><span style="font-size: 0.8em; text-align: right; display: block;">Tokens Left: '
|
503 |
+ str(tokens_left)
|
|
|
21 |
get_time,
|
22 |
check_user_cooldown,
|
23 |
reset_tokens_for_user,
|
24 |
+
get_user_details,
|
25 |
)
|
26 |
import copy
|
27 |
from typing import Optional
|
|
|
56 |
return data_layer
|
57 |
|
58 |
|
59 |
+
async def update_user_from_chainlit(user, token_count=0):
|
60 |
+
if "admin" not in user.metadata["role"]:
|
61 |
+
user.metadata["tokens_left"] = user.metadata["tokens_left"] - token_count
|
62 |
+
user.metadata["all_time_tokens_allocated"] = (
|
63 |
+
user.metadata["all_time_tokens_allocated"] - token_count
|
64 |
+
)
|
65 |
+
user.metadata["tokens_left_at_last_message"] = user.metadata[
|
66 |
+
"tokens_left"
|
67 |
+
] # tokens_left will keep regenerating outside of chainlit
|
68 |
+
user.metadata["last_message_time"] = get_time()
|
69 |
+
await update_user_info(user)
|
70 |
+
|
71 |
+
tokens_left = user.metadata["tokens_left"]
|
72 |
+
if tokens_left < 0:
|
73 |
+
tokens_left = 0
|
74 |
+
return tokens_left
|
75 |
+
|
76 |
+
|
77 |
class Chatbot:
|
78 |
def __init__(self, config):
|
79 |
"""
|
|
|
386 |
|
387 |
# update user info with last message time
|
388 |
user = cl.user_session.get("user")
|
389 |
+
await reset_tokens_for_user(user)
|
390 |
+
updated_user = await get_user_details(user.identifier)
|
391 |
+
user.metadata = updated_user.metadata
|
392 |
+
cl.user_session.set("user", user)
|
393 |
|
394 |
print("\n\n User Tokens Left: ", user.metadata["tokens_left"])
|
395 |
|
|
|
429 |
user.metadata["in_cooldown"] = True
|
430 |
await update_user_info(user)
|
431 |
return
|
432 |
+
else:
|
433 |
+
await cl.Message(
|
434 |
+
content=(
|
435 |
+
"Ah, seems like you don't have any tokens left...Please wait while we regenerate your tokens. Click "
|
436 |
+
'<a href="/cooldown" style="color: #0000CD; text-decoration: none;" target="_self">here</a> to view your token credits.'
|
437 |
+
),
|
438 |
+
author=SYSTEM,
|
439 |
+
).send()
|
440 |
+
return
|
441 |
|
442 |
user.metadata["in_cooldown"] = False
|
443 |
|
|
|
518 |
print("Time taken to generate questions: ", time.time() - start_time)
|
519 |
|
520 |
# # update user info with token count
|
521 |
+
tokens_left = await update_user_from_chainlit(user, token_count)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
522 |
|
|
|
|
|
|
|
523 |
answer_with_sources += (
|
524 |
'\n\n<footer><span style="font-size: 0.8em; text-align: right; display: block;">Tokens Left: '
|
525 |
+ str(tokens_left)
|
code/modules/chat_processor/helpers.py
CHANGED
@@ -206,46 +206,34 @@ async def reset_tokens_for_user(user_info):
|
|
206 |
elapsed_time_in_seconds = (current_time - last_message_time).total_seconds()
|
207 |
|
208 |
# Current token count (can be negative)
|
209 |
-
current_tokens = user_info["metadata"].get("
|
|
|
210 |
|
211 |
# Maximum tokens that can be regenerated
|
212 |
max_tokens = user_info["metadata"].get("max_tokens", TOKENS_LEFT)
|
213 |
|
214 |
# Calculate how many tokens should have been regenerated proportionally
|
215 |
if current_tokens < max_tokens:
|
216 |
-
|
217 |
-
|
218 |
-
|
219 |
-
|
220 |
-
|
221 |
-
else:
|
222 |
-
time_to_full_regen = REGEN_TIME * (
|
223 |
-
1 - current_tokens / max_tokens
|
224 |
-
) # less time to regenerate if tokens left is positive
|
225 |
-
|
226 |
-
# Calculate the proportion of this time that has elapsed
|
227 |
-
proportion_of_time_elapsed = elapsed_time_in_seconds / time_to_full_regen
|
228 |
-
|
229 |
-
# Ensure the proportion doesn't exceed 1.0
|
230 |
-
proportion_of_time_elapsed = min(proportion_of_time_elapsed, 1.0)
|
231 |
-
|
232 |
-
# Calculate the tokens to regenerate based on the elapsed proportion
|
233 |
tokens_to_regenerate = int(
|
234 |
-
|
235 |
)
|
236 |
|
237 |
-
# Ensure
|
238 |
-
|
239 |
-
new_token_count = min(current_tokens + tokens_to_regenerate, max_tokens)
|
240 |
|
241 |
-
|
242 |
-
|
243 |
-
|
244 |
|
245 |
-
|
246 |
-
|
247 |
|
248 |
-
|
249 |
|
250 |
|
251 |
async def get_thread_step_info(thread_id):
|
|
|
206 |
elapsed_time_in_seconds = (current_time - last_message_time).total_seconds()
|
207 |
|
208 |
# Current token count (can be negative)
|
209 |
+
current_tokens = user_info["metadata"].get("tokens_left_at_last_message", 0)
|
210 |
+
current_tokens = min(current_tokens, TOKENS_LEFT)
|
211 |
|
212 |
# Maximum tokens that can be regenerated
|
213 |
max_tokens = user_info["metadata"].get("max_tokens", TOKENS_LEFT)
|
214 |
|
215 |
# Calculate how many tokens should have been regenerated proportionally
|
216 |
if current_tokens < max_tokens:
|
217 |
+
|
218 |
+
# Calculate the regeneration rate per second based on REGEN_TIME for full regeneration
|
219 |
+
regeneration_rate_per_second = max_tokens / REGEN_TIME
|
220 |
+
|
221 |
+
# Calculate how many tokens should have been regenerated based on the elapsed time
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
222 |
tokens_to_regenerate = int(
|
223 |
+
elapsed_time_in_seconds * regeneration_rate_per_second
|
224 |
)
|
225 |
|
226 |
+
# Ensure the new token count does not exceed max_tokens
|
227 |
+
new_token_count = min(current_tokens + tokens_to_regenerate, max_tokens)
|
|
|
228 |
|
229 |
+
print(
|
230 |
+
f"\n\n Adding {tokens_to_regenerate} tokens to the user, Time elapsed: {elapsed_time_in_seconds} seconds, Tokens after regeneration: {new_token_count}, Tokens before: {current_tokens} \n\n"
|
231 |
+
)
|
232 |
|
233 |
+
# Update the user's token count
|
234 |
+
user_info["metadata"]["tokens_left"] = new_token_count
|
235 |
|
236 |
+
await update_user_info(user_info)
|
237 |
|
238 |
|
239 |
async def get_thread_step_info(thread_id):
|
code/modules/config/constants.py
CHANGED
@@ -4,8 +4,8 @@ import os
|
|
4 |
load_dotenv()
|
5 |
|
6 |
TIMEOUT = 60
|
7 |
-
COOLDOWN_TIME =
|
8 |
-
REGEN_TIME =
|
9 |
TOKENS_LEFT = 2000
|
10 |
ALL_TIME_TOKENS_ALLOCATED = 1000000
|
11 |
|
|
|
4 |
load_dotenv()
|
5 |
|
6 |
TIMEOUT = 60
|
7 |
+
COOLDOWN_TIME = 60
|
8 |
+
REGEN_TIME = 180
|
9 |
TOKENS_LEFT = 2000
|
10 |
ALL_TIME_TOKENS_ALLOCATED = 1000000
|
11 |
|