Piteball
/
chatgpt
mirror of https://github.com/knrd1/chatgpt.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

multiple endpoints support

This commit is contained in:
konrad 2023-04-01 17:28:07 +01:00
parent 8e4f86d73d
commit 1a285bd24b
2 changed files with 72 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
README.md
View File

@ -59,11 +59,14 @@ ChatGPT will interact only if you mention its nickname:
```
### Model endpoint compatibility
ChatGPT uses API endpoint v1/chat/completions. Following models are compatible.
ChatGPT IRC bot can use two API endpoints: /v1/chat/completions and /v1/completions
Following models support endpoint /v1/chat/completions:
```
gpt-4, gpt-4-0314, gpt-4-32k, gpt-4-32k-0314, gpt-3.5-turbo, gpt-3.5-turbo-0301
```
Models that support /v1/completions:
```
text-davinci-003, text-davinci-002, text-curie-001, text-babbage-001, text-ada-001, davinci, curie, babbage, ada
```
More details about models: https://platform.openai.com/docs/models
To use API endpoint v1/completions, use models that support "completions", e.g. "text-davinci-003"
See branch "completions" for Python code that supports this API endpoint.

95
chatgpt.py
View File

@ -48,7 +48,7 @@ while True:
irc.send(bytes("USER " + ident + " 0 * :" + realname + "\n", "UTF-8"))
irc.send(bytes("NICK " + nickname + "\n", "UTF-8"))
irc.send(bytes("JOIN " + ",".join(channels) + "\n", "UTF-8"))
print ("connected to:" + server)
print ("Connected to:" + server)
break
except:
print("Connection failed. Retrying in 5 seconds...")
@ -79,33 +79,68 @@ while True:
elif command == "PRIVMSG" and chunk[2].startswith("#") and chunk[3] == ":" + nickname + ":":
channel = chunk[2].strip()
question = data.split(nickname + ":")[1].strip()
try:
response = openai.ChatCompletion.create(
model=model,
messages=[{"role": role, "content": question}],
temperature=temperature,
max_tokens=max_tokens,
top_p=top_p,
frequency_penalty=frequency_penalty,
presence_penalty=presence_penalty,
request_timeout=request_timeout
)
answers = [x.strip() for x in response.choices[0].message.content.strip().split('\n')]
for answer in answers:
while len(answer) > 0:
if len(answer) <= 392:
irc.send(bytes("PRIVMSG " + channel + " :" + answer + "\n", "UTF-8"))
answer = ""
else:
last_space_index = answer[:392].rfind(" ")
if last_space_index == -1:
last_space_index = 392
irc.send(bytes("PRIVMSG " + channel + " :" + answer[:last_space_index] + "\n", "UTF-8"))
answer = answer[last_space_index:].lstrip()
except openai.error.Timeout as e:
print("Error: " + str(e))
irc.send(bytes("PRIVMSG " + channel + " :API call timed out. Try again later.\n", "UTF-8"))
except Exception as e:
print("Error: " + str(e))
irc.send(bytes("PRIVMSG " + channel + " :API call failed. Try again later.\n", "UTF-8"))
if model in ["gpt-4", "gpt-4-0314", "gpt-4-32k", "gpt-4-32k-0314", "gpt-3.5-turbo", "gpt-3.5-turbo-0301"]:
try:
response = openai.ChatCompletion.create(
model=model,
messages=[{"role": role, "content": question}],
temperature=temperature,
max_tokens=max_tokens,
top_p=top_p,
frequency_penalty=frequency_penalty,
presence_penalty=presence_penalty,
request_timeout=request_timeout
)
answers = [x.strip() for x in response.choices[0].message.content.strip().split('\n')]
for answer in answers:
while len(answer) > 0:
if len(answer) <= 392:
irc.send(bytes("PRIVMSG " + channel + " :" + answer + "\n", "UTF-8"))
answer = ""
else:
last_space_index = answer[:392].rfind(" ")
if last_space_index == -1:
last_space_index = 392
irc.send(bytes("PRIVMSG " + channel + " :" + answer[:last_space_index] + "\n", "UTF-8"))
answer = answer[last_space_index:].lstrip()
except openai.error.Timeout as e:
print("Error: " + str(e))
irc.send(bytes("PRIVMSG " + channel + " :API call timed out. Try again later.\n", "UTF-8"))
except Exception as e:
print("Error: " + str(e))
irc.send(bytes("PRIVMSG " + channel + " :API call failed. Try again later.\n", "UTF-8"))
elif model in ["text-davinci-003", "text-davinci-002", "text-curie-001", "text-babbage-001", "text-ada-001", "davinci", "curie", "babbage", "ada"]:
try:
response = openai.Completion.create(
model=model,
prompt="Q: " + question + "\nA:",
temperature=temperature,
max_tokens=max_tokens,
top_p=top_p,
frequency_penalty=frequency_penalty,
presence_penalty=presence_penalty,
request_timeout=request_timeout
)
answers = [x.strip() for x in response.choices[0].text.strip().split('\n')]
for answer in answers:
while len(answer) > 0:
if len(answer) <= 392:
irc.send(bytes("PRIVMSG " + channel + " :" + answer + "\n", "UTF-8"))
answer = ""
else:
last_space_index = answer[:392].rfind(" ")
if last_space_index == -1:
last_space_index = 392
irc.send(bytes("PRIVMSG " + channel + " :" + answer[:last_space_index] + "\n", "UTF-8"))
answer = answer[last_space_index:].lstrip()
except openai.error.Timeout as e:
print("Error: " + str(e))
irc.send(bytes("PRIVMSG " + channel + " :API call timed out. Try again later.\n", "UTF-8"))
except Exception as e:
print("Error: " + str(e))
irc.send(bytes("PRIVMSG " + channel + " :API call failed. Try again later.\n", "UTF-8"))
else:
print("Invalid model.")
irc.send(bytes("PRIVMSG " + channel + " :Invalid model.\n", "UTF-8"))
continue
time.sleep(1)