joey0629
/
chat_function
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

init

Browse Source
master
josephwang 11 months ago
commit 3f515e9a88
8 changed files with 265 additions and 0 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 8
      .idea/.gitignore
  2. 10
      .idea/chatbot.iml
  3. 6
      .idea/inspectionProfiles/profiles_settings.xml
  4. 7
      .idea/misc.xml
  5. 8
      .idea/modules.xml
  6. 6
      .idea/vcs.xml
  7. 99
      function_calling.py
  8. 121
      langchain_ver.py

8
.idea/.gitignore vendored
Unescape View File

@ -0,0 +1,8 @@
# Default ignored files
/shelf/
/workspace.xml
# Editor-based HTTP Client requests
/httpRequests/
# Datasource local storage ignored files
/dataSources/
/dataSources.local.xml

10
.idea/chatbot.iml
Unescape View File

@ -0,0 +1,10 @@
<?xml version="1.0" encoding="UTF-8"?>
<module type="PYTHON_MODULE" version="4">
<component name="NewModuleRootManager">
<content url="file://$MODULE_DIR$">
<excludeFolder url="file://$MODULE_DIR$/.venv" />
</content>
<orderEntry type="inheritedJdk" />
<orderEntry type="sourceFolder" forTests="false" />
</component>
</module>

6
.idea/inspectionProfiles/profiles_settings.xml
Unescape View File

@ -0,0 +1,6 @@
<component name="InspectionProjectProfileManager">
<settings>
<option name="USE_PROJECT_PROFILE" value="false" />
<version value="1.0" />
</settings>
</component>

7
.idea/misc.xml
Unescape View File

@ -0,0 +1,7 @@
<?xml version="1.0" encoding="UTF-8"?>
<project version="4">
<component name="Black">
<option name="sdkName" value="Python 3.10 (chatbot)" />
</component>
<component name="ProjectRootManager" version="2" project-jdk-name="Python 3.10 (chatbot)" project-jdk-type="Python SDK" />
</project>

8
.idea/modules.xml
Unescape View File

@ -0,0 +1,8 @@
<?xml version="1.0" encoding="UTF-8"?>
<project version="4">
<component name="ProjectModuleManager">
<modules>
<module fileurl="file://$PROJECT_DIR$/.idea/chatbot.iml" filepath="$PROJECT_DIR$/.idea/chatbot.iml" />
</modules>
</component>
</project>

6
.idea/vcs.xml
Unescape View File

@ -0,0 +1,6 @@
<?xml version="1.0" encoding="UTF-8"?>
<project version="4">
<component name="VcsDirectoryMappings">
<mapping directory="$PROJECT_DIR$" vcs="Git" />
</component>
</project>

99
function_calling.py
Unescape View File

@ -0,0 +1,99 @@
import json
from openai import OpenAI
from tenacity import retry, wait_random_exponential, stop_after_attempt
import sys
GPT_MODEL = "gpt-4-turbo"
client = OpenAI(api_key="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")
@retry(wait=wait_random_exponential(multiplier=1, max=40), stop=stop_after_attempt(3))
def chat_completion_request(messages, tools=None, tool_choice=None, model=GPT_MODEL):
try:
response = client.chat.completions.create(
model=model,
messages=messages,
tools=tools,
tool_choice=tool_choice,
)
return response
except Exception as e:
print("Unable to generate ChatCompletion response")
print(f"Exception: {e}")
sys.exit(1) # Add this line
def get_backtest(symbols, sliding_window="six months", frequency="semi-annually", function="sharpe ratio"):
symbol = " , ".join(symbols)
return f"Backtest result for {symbol} Annualized return: 10%, Annualized Sharpe ratio: 1.5, Annualized volatility: 20%, Maximum drawdown: 5%, Alpha: 0.1, Beta: 0.8, VaR10: 5%, R2: 0.9"
def backest_main(query):
tools = [
{
"type": "function",
"function": {
"name": "get_backtest",
"description": "Get the portfolio backtesting result by combined with a list of symbol, sliding window, optimize frequency and optimize function",
"parameters": {
"type": "object",
"properties": {
"symbol": {
"type": "array",
"description": "An array of multiple portfolio symbol to be backtested if the symbol is Taiwan Stock exchage the code ex: TSMC to 2330.TW , if there is multiple symbol return a python list format",
"items": {
"type": "string",
"description": "The symbol of the stock",
},
},
"sliding window": {
"type": "string",
"enum": ["one month", "three months", "six months", "one year"],
"description": "The sliding window size to be backtested in one month, three months , six months, one year",
},
"frequency": {
"type": "string",
"enum": ["monthly", "quarterly", "semi-annually", "annually"],
"description": "The optimize frequency to be backtested in monthly, quarterly, semi-annually, annually",
},
"function": {
"type": "string",
"enum": ["sharpe ratio", "sortino ratio", "volatility", "utility function"],
"description": "The optimize function to be backtested in sharpe ratio, sortino ratio,volatility, utility function",
},
},
"required": ["symbol", "sliding window", "frequency", "function"],
},
}
}
]
messages = []
sec_message = []
messages.append({"role": "system", "content": "You are a software developer who is writing a function to get the portfolio backtesting result by using different multiple symbol, sliding window, optimize frequency and optimize function , "
"only the symbol is required, the sliding window, frequency and function are optional. The sliding window size can be default by six months. The optimize frequency can be default by semi-annually. The optimize function can be default by sharpe ratio."})
messages.append({"role": "user", "content": query})
chat_response = chat_completion_request(messages,tools=tools)
assistant_messages = chat_response.choices[0].message.tool_calls
messages.append(assistant_messages)
available_functions = {
"get_backtest": get_backtest,
} # only one function in this example, but you can have multiple
for tool_call in assistant_messages:
function_name = tool_call.function.name
function_to_call = available_functions[function_name]
function_args = json.loads(tool_call.function.arguments)
function_response = function_to_call(
symbols=function_args.get("symbol"),
sliding_window=function_args.get("sliding window"),
frequency=function_args.get("frequency"),
function=function_args.get("function"),
)
sec_message.append({"role": "system","content": "You are a professional financial analyst. The user will provide you with some results from their backtesting of an investment portfolio using the Efficient Frontier calculation. These results include annualized return, annualized Sharpe ratio, annualized volatility, maximum drawdown, Alpha, Beta, VaR10, and R2. Please provide professional advice in Traditional Chinese based on these reports. "})
sec_message.append({"role": "user","content": function_response})
result_messages = chat_completion_request(sec_message)
return result_messages.choices[0].message.content
query = "我想要使用台積電和蘋果股票來進行最大夏普比率的回測"
article = backest_main(query)
print(article)

121
langchain_ver.py
Unescape View File

@ -0,0 +1,121 @@
import json
from tenacity import retry, wait_random_exponential, stop_after_attempt
import sys
from langchain.chains import LLMChain
from langchain.prompts import PromptTemplate
from langchain.llms import OpenAI as LangChainOpenAI
GPT_MODEL = "gpt-4o"
client = LangChainOpenAI(api_key="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", model_name=GPT_MODEL)
@retry(wait=wait_random_exponential(multiplier=1, max=40), stop=stop_after_attempt(3))
def chat_completion_request(messages, tools=None):
try:
response = client(messages=messages, tools=tools)
return response
except Exception as e:
print("Unable to generate ChatCompletion response")
print(f"Exception: {e}")
sys.exit(1)
def get_backtest(symbols, sliding_window="six months", frequency="semi-annually", function="sharpe ratio"):
symbol = ", ".join(symbols)
return f"Backtest result for {symbol}: Annualized return: 10%, Annualized Sharpe ratio: 1.5, Annualized volatility: 20%, Maximum drawdown: 5%, Alpha: 0.1, Beta: 0.8, VaR10: 5%, R2: 0.9"
def backtest_main(query):
tools = {
"get_backtest": {
"description": "Get the portfolio backtesting result by combining a list of symbols, sliding window, optimization frequency, and optimization function",
"parameters": {
"symbol": {
"type": "array",
"description": "An array of multiple portfolio symbols to be backtested. If the symbol is Taiwan Stock Exchange, the code should be formatted as TSMC to 2330.TW. If there are multiple symbols, return in a Python list format",
},
"sliding window": {
"type": "string",
"enum": ["one month", "three months", "six months", "one year"],
"description": "The sliding window size to be backtested: one month, three months, six months, or one year",
},
"frequency": {
"type": "string",
"enum": ["monthly", "quarterly", "semi-annually", "annually"],
"description": "The optimization frequency to be backtested: monthly, quarterly, semi-annually, or annually",
},
"function": {
"type": "string",
"enum": ["sharpe ratio", "sortino ratio", "volatility", "utility function"],
"description": "The optimization function to be backtested: Sharpe ratio, Sortino ratio, volatility, or utility function",
},
},
"required": ["symbol", "sliding window", "frequency", "function"],
}
}
initial_prompt = PromptTemplate.from_template(
"You are a software developer writing a function to get the portfolio backtesting result using different symbols, sliding window, optimization frequency, and optimization function. "
"Only the symbol is required; the sliding window, frequency, and function are optional. The sliding window size can default to six months, the optimization frequency can default to semi-annually, and the optimization function can default to Sharpe ratio."
)
secondary_prompt = PromptTemplate.from_template(
"You are a professional financial analyst. The user will provide you with some results from their backtesting of an investment portfolio using the Efficient Frontier calculation. These results include annualized return, annualized Sharpe ratio, annualized volatility, maximum drawdown, Alpha, Beta, VaR10, and R2. Please provide professional advice in Traditional Chinese based on these reports."
)
main_chain = LLMChain(
prompt=initial_prompt,
llm=client,
)
tool_chain = LLMChain(
prompt=secondary_prompt,
llm=client,
)
messages = [{"role": "system", "content": initial_prompt.template}, {"role": "user", "content": query}]
chat_response = main_chain.run(messages=messages)
if not chat_response or "choices" not in chat_response or not chat_response["choices"]:
print("No valid response received from the initial prompt.")
return
assistant_messages = chat_response["choices"][0].get("message", {}).get("tool_calls", [])
if not assistant_messages:
print("No tool calls received in the assistant messages.")
return
function_responses = []
available_functions = {
"get_backtest": get_backtest,
}
for tool_call in assistant_messages:
function_name = tool_call["function"]["name"]
function_to_call = available_functions.get(function_name)
if function_to_call:
function_args = json.loads(tool_call["function"]["arguments"])
function_response = function_to_call(
symbols=function_args.get("symbol"),
sliding_window=function_args.get("sliding window"),
frequency=function_args.get("frequency"),
function=function_args.get("function"),
)
function_responses.append(function_response)
if not function_responses:
print("No valid function responses generated.")
return
sec_messages = [{"role": "system", "content": secondary_prompt.template}, {"role": "user", "content": function_responses[0]}]
result_messages = tool_chain.run(messages=sec_messages)
if not result_messages or "choices" not in result_messages or not result_messages["choices"]:
print("No valid response received from the secondary prompt.")
return
return result_messages["choices"][0]["message"]["content"]
query = "我想要使用台積電和蘋果股票來進行最大夏普比率的回測"
article = backtest_main(query)
if article:
print(article)
Write Preview
Loading…
Cancel
Save