Upload 2 files
Browse files- requirements.txt +17 -0
- streamlit.py +423 -0
requirements.txt
ADDED
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
streamlit_chat
|
2 |
+
streamlit~=1.23.1
|
3 |
+
langchain~=0.0.225
|
4 |
+
sentence_transformers
|
5 |
+
utils~=1.0.1
|
6 |
+
cohere~=4.11.2
|
7 |
+
openai~=0.27.8
|
8 |
+
pypdf2~=3.0.1
|
9 |
+
tiktoken~=0.4.0
|
10 |
+
PyPDF2~=3.0.1
|
11 |
+
langchain~=0.0.231
|
12 |
+
chromadb~=0.3.27
|
13 |
+
yfinance~=0.2.25
|
14 |
+
yahooquery~=2.3.1
|
15 |
+
google-search-results~=2.4.2
|
16 |
+
fpdf~=1.7.2
|
17 |
+
pypdf~=3.12.1
|
streamlit.py
ADDED
@@ -0,0 +1,423 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
|
3 |
+
# Import necessary libraries
|
4 |
+
import os
|
5 |
+
import requests
|
6 |
+
import json
|
7 |
+
import yfinance as yf
|
8 |
+
from yahooquery import Ticker
|
9 |
+
from fpdf import FPDF
|
10 |
+
from typing import List, Union
|
11 |
+
import re
|
12 |
+
|
13 |
+
# Import components from langchain and other libraries
|
14 |
+
from langchain.agents import load_tools, initialize_agent, create_csv_agent, AgentType, AgentExecutor, LLMSingleActionAgent, AgentOutputParser
|
15 |
+
from langchain.agents.agent_toolkits import create_vectorstore_agent, VectorStoreToolkit, VectorStoreInfo
|
16 |
+
from langchain.llms import OpenAI, Cohere
|
17 |
+
from langchain.document_loaders import PyPDFLoader, TextLoader, DirectoryLoader
|
18 |
+
from langchain.vectorstores import Chroma
|
19 |
+
from langchain.embeddings import CohereEmbeddings, OpenAIEmbeddings
|
20 |
+
from langchain.indexes import VectorstoreIndexCreator
|
21 |
+
from langchain.chains import RetrievalQA
|
22 |
+
from langchain.evaluation.qa import QAEvalChain
|
23 |
+
from langchain.prompts import StringPromptTemplate
|
24 |
+
from langchain.tools.python.tool import PythonREPLTool
|
25 |
+
from langchain.python import PythonREPL
|
26 |
+
from langchain import LLMMathChain, SerpAPIWrapper, LLMChain
|
27 |
+
from langchain.schema import AgentAction, AgentFinish, OutputParserException
|
28 |
+
|
29 |
+
# Set up Streamlit header and sidebar
|
30 |
+
st.header('FinAI')
|
31 |
+
mod = None
|
32 |
+
with st.sidebar:
|
33 |
+
with st.form('Cohere/OpenAI'):
|
34 |
+
# User selects the model (OpenAI/Cohere) and enters API keys
|
35 |
+
model = st.radio('Choose OpenAI/Cohere', ('OpenAI', 'Cohere'))
|
36 |
+
api_key = st.text_input('Enter API key', type="password")
|
37 |
+
serpAI_key = st.text_input('Enter SERPAIAPI key', type="password")
|
38 |
+
submitted = st.form_submit_button("Submit")
|
39 |
+
|
40 |
+
# Check if API key is provided and set up the language model accordingly
|
41 |
+
if api_key:
|
42 |
+
if model == 'OpenAI':
|
43 |
+
os.environ["OPENAI_API_KEY"] = api_key
|
44 |
+
llm = OpenAI(temperature=0.3)
|
45 |
+
mod = 'OpenAI'
|
46 |
+
os.environ["SERPAPI_API_KEY"] = serpAI_key
|
47 |
+
elif model == 'Cohere':
|
48 |
+
os.environ["Cohere_API_KEY"] = api_key
|
49 |
+
llm = Cohere(cohere_api_key=api_key)
|
50 |
+
mod = 'Cohere'
|
51 |
+
os.environ["SERPAPI_API_KEY"] = serpAI_key
|
52 |
+
|
53 |
+
# Helper function to get company news from SERP API
|
54 |
+
def get_company_news(company_name):
|
55 |
+
# Set the parameters for the SERP API request
|
56 |
+
params = {
|
57 |
+
"engine": "google",
|
58 |
+
"tbm": "nws",
|
59 |
+
"q": company_name,
|
60 |
+
"api_key": os.environ["SERPAPI_API_KEY"],
|
61 |
+
}
|
62 |
+
|
63 |
+
# Send the request and get the response data in JSON format
|
64 |
+
response = requests.get('https://serpapi.com/search', params=params)
|
65 |
+
data = response.json()
|
66 |
+
|
67 |
+
return data.get('news_results')
|
68 |
+
|
69 |
+
# Helper function to write news data to a file
|
70 |
+
def write_news_to_file(news, filename):
|
71 |
+
with open(filename, 'w') as file:
|
72 |
+
for news_item in news:
|
73 |
+
if news_item is not None:
|
74 |
+
title = news_item.get('title', 'No title')
|
75 |
+
link = news_item.get('link', 'No link')
|
76 |
+
date = news_item.get('date', 'No date')
|
77 |
+
file.write(f"Title: {title}\n")
|
78 |
+
file.write(f"Link: {link}\n")
|
79 |
+
file.write(f"Date: {date}\n\n")
|
80 |
+
|
81 |
+
# Helper function to get stock evolution data from Yahoo Finance API
|
82 |
+
def get_stock_evolution(company_name, period="1y"):
|
83 |
+
# Get the stock information using yfinance
|
84 |
+
stock = yf.Ticker(company_name)
|
85 |
+
|
86 |
+
# Get historical market data for the specified period
|
87 |
+
hist = stock.history(period=period)
|
88 |
+
|
89 |
+
# Convert the DataFrame to a string with a specific format
|
90 |
+
data_string = hist.to_string()
|
91 |
+
|
92 |
+
# Save the historical market data to a CSV file
|
93 |
+
hist.to_csv('stocks_data.csv')
|
94 |
+
|
95 |
+
# Get financials data
|
96 |
+
fin = stock.get_financials()
|
97 |
+
fin.to_csv('fin_data.csv')
|
98 |
+
|
99 |
+
# Append the string to the "investment.txt" file
|
100 |
+
with open("investment.txt", "a") as file:
|
101 |
+
file.write(f"\nStock Evolution for {company_name}:\n")
|
102 |
+
file.write(data_string)
|
103 |
+
file.write("\n")
|
104 |
+
|
105 |
+
# Helper function to get financial statements from Yahoo Finance API
|
106 |
+
def get_financial_statements(ticker):
|
107 |
+
# Create a Ticker object
|
108 |
+
company = Ticker(ticker)
|
109 |
+
|
110 |
+
# Get financial data for balance sheet, cash flow, income statement, and valuation measures
|
111 |
+
balance_sheet = company.balance_sheet().to_string()
|
112 |
+
cash_flow = company.cash_flow(trailing=False).to_string()
|
113 |
+
income_statement = company.income_statement().to_string()
|
114 |
+
valuation_measures = str(company.valuation_measures) # This one might already be a dictionary or string
|
115 |
+
|
116 |
+
# Write data to "investment.txt" file
|
117 |
+
with open("investment.txt", "a") as file:
|
118 |
+
file.write("\nBalance Sheet\n")
|
119 |
+
file.write(balance_sheet)
|
120 |
+
file.write("\nCash Flow\n")
|
121 |
+
file.write(cash_flow)
|
122 |
+
file.write("\nIncome Statement\n")
|
123 |
+
file.write(income_statement)
|
124 |
+
file.write("\nValuation Measures\n")
|
125 |
+
file.write(valuation_measures)
|
126 |
+
|
127 |
+
# Helper function to fetch data from different sources and store it in "investment.txt" file
|
128 |
+
def get_data(company_name, company_ticker, period="1y", filename="investment.txt"):
|
129 |
+
news = get_company_news(company_name)
|
130 |
+
if news:
|
131 |
+
write_news_to_file(news, filename)
|
132 |
+
else:
|
133 |
+
print("No news found.")
|
134 |
+
|
135 |
+
get_stock_evolution(company_ticker, period)
|
136 |
+
|
137 |
+
get_financial_statements(company_ticker)
|
138 |
+
|
139 |
+
# Helper function to call the language model for financial analysis based on user request
|
140 |
+
def financial_analyst(request):
|
141 |
+
# Print the request received from the user
|
142 |
+
print(f"Received request: {request}")
|
143 |
+
|
144 |
+
# Use OpenAI GPT-3.5 Turbo model to analyze the request and generate a response
|
145 |
+
response = openai.ChatCompletion.create(
|
146 |
+
model="gpt-3.5-turbo-16k",
|
147 |
+
messages=[{
|
148 |
+
"role": "user",
|
149 |
+
"content": f"Given the user request, what is the company name and the company stock ticker ?: {request}?"
|
150 |
+
}],
|
151 |
+
functions=[{
|
152 |
+
"name": "get_data",
|
153 |
+
"description": "Get financial data on a specific company for investment purposes",
|
154 |
+
"parameters": {
|
155 |
+
"type": "object",
|
156 |
+
"properties": {
|
157 |
+
"company_name": {
|
158 |
+
"type": "string",
|
159 |
+
"description": "The name of the company",
|
160 |
+
},
|
161 |
+
"company_ticker": {
|
162 |
+
"type": "string",
|
163 |
+
"description": "The ticker of the stock of the company"
|
164 |
+
},
|
165 |
+
"period": {
|
166 |
+
"type": "string",
|
167 |
+
"description": "The period of analysis"
|
168 |
+
},
|
169 |
+
"filename": {
|
170 |
+
"type": "string",
|
171 |
+
"description": "The filename to store data"
|
172 |
+
}
|
173 |
+
},
|
174 |
+
"required": ["company_name", "company_ticker"],
|
175 |
+
},
|
176 |
+
}],
|
177 |
+
function_call={"name": "get_data"},
|
178 |
+
)
|
179 |
+
|
180 |
+
# Extract the arguments and company information from the response
|
181 |
+
message = response["choices"][0]["message"]
|
182 |
+
if message.get("function_call"):
|
183 |
+
arguments = json.loads(message["function_call"]["arguments"])
|
184 |
+
company_name = arguments["company_name"]
|
185 |
+
company_ticker = arguments["company_ticker"]
|
186 |
+
|
187 |
+
# Call the function to fetch and store financial data
|
188 |
+
get_data(company_name, company_ticker)
|
189 |
+
|
190 |
+
# Read the contents of the "investment.txt" file for the response
|
191 |
+
with open("investment.txt", "r") as file:
|
192 |
+
content = file.read()[:14000]
|
193 |
+
|
194 |
+
# Use OpenAI GPT-3.5 Turbo model again to provide a detailed investment thesis
|
195 |
+
second_response = openai.ChatCompletion.create(
|
196 |
+
model="gpt-3.5-turbo-16k",
|
197 |
+
messages=[
|
198 |
+
{
|
199 |
+
"role": "user",
|
200 |
+
"content": request
|
201 |
+
},
|
202 |
+
message,
|
203 |
+
{
|
204 |
+
"role": "system",
|
205 |
+
"content": """write a detailed investment thesis to answer
|
206 |
+
the user request. Provide numbers to justify
|
207 |
+
your assertions, a lot ideally. Never mention
|
208 |
+
something like this:
|
209 |
+
However, it is essential to consider your own risk
|
210 |
+
tolerance, financial goals, and time horizon before
|
211 |
+
making any investment decisions. It is recommended
|
212 |
+
to consult with a financial advisor or do further
|
213 |
+
research to gain more insights into the company's
|
214 |
+
fundamentals and market trends. The user
|
215 |
+
already knows that"""
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"role": "assistant",
|
219 |
+
"content": content,
|
220 |
+
},
|
221 |
+
],
|
222 |
+
)
|
223 |
+
|
224 |
+
return second_response["choices"][0]["message"]["content"]
|
225 |
+
|
226 |
+
# Helper function to generate a PDF with text and images
|
227 |
+
def generate_pdf(text, image_paths):
|
228 |
+
# Create a FPDF object
|
229 |
+
pdf = FPDF()
|
230 |
+
|
231 |
+
# Add a page
|
232 |
+
pdf.add_page()
|
233 |
+
|
234 |
+
# Set style and size of font for the PDF
|
235 |
+
pdf.set_font("Arial", size=12)
|
236 |
+
|
237 |
+
# Set left and right margins
|
238 |
+
pdf.set_left_margin(20)
|
239 |
+
pdf.set_right_margin(20)
|
240 |
+
|
241 |
+
# Add multi-cell with line break for the text
|
242 |
+
pdf.multi_cell(0, 10, text)
|
243 |
+
|
244 |
+
# Move to the next line after the text
|
245 |
+
pdf.ln()
|
246 |
+
|
247 |
+
# Add a new page for the images
|
248 |
+
pdf.add_page()
|
249 |
+
|
250 |
+
# Add the first image to the PDF
|
251 |
+
pdf.image(image_paths[0], x=20, y=pdf.get_y(), w=175)
|
252 |
+
|
253 |
+
# Calculate the y-coordinate for the second image
|
254 |
+
second_image_y = pdf.get_y() + 150
|
255 |
+
|
256 |
+
# Add the second image to the PDF
|
257 |
+
pdf.image(image_paths[1], x=20, y=second_image_y, w=175)
|
258 |
+
|
259 |
+
# Save the PDF with the given file name
|
260 |
+
pdf.output("output.pdf")
|
261 |
+
|
262 |
+
# Helper function to generate graphs using langchain and save as images
|
263 |
+
def graphs(path, prompt):
|
264 |
+
agent = create_csv_agent(
|
265 |
+
OpenAI(temperature=0),
|
266 |
+
path,
|
267 |
+
verbose=True,
|
268 |
+
agent_type=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
|
269 |
+
)
|
270 |
+
agent.run(prompt)
|
271 |
+
|
272 |
+
|
273 |
+
# Category 1: Revenue and Sales
|
274 |
+
revenue_sales_questions = [
|
275 |
+
"What was the total revenue generated by the company during the fiscal year?",
|
276 |
+
"How does the revenue of the current year compare to the previous year?",
|
277 |
+
"Which product/service contributed the most to the company's revenue?",
|
278 |
+
"Did the company experience any significant changes in sales volume or pricing?",
|
279 |
+
"Are there any notable trends or patterns in the revenue growth of the company over the past few years?",
|
280 |
+
"What were the geographical regions or markets where the company generated the highest revenue?",
|
281 |
+
"Has the company introduced any new revenue streams or business lines?",
|
282 |
+
"Were there any extraordinary events or factors that affected the company's revenue performance?",
|
283 |
+
"How does the revenue composition of the company compare to its competitors in the industry?",
|
284 |
+
"Are there any forecasts or projections for future revenue growth provided in the report?"
|
285 |
+
]
|
286 |
+
|
287 |
+
# Category 2: Expenses and Costs
|
288 |
+
expenses_costs_questions = [
|
289 |
+
"What were the major expense categories for the company during the fiscal year?",
|
290 |
+
"How do the expenses of the current year compare to the previous year?",
|
291 |
+
"Did the company implement any cost-saving measures or efficiency improvements?",
|
292 |
+
"Were there any significant changes in the cost of raw materials or production inputs?",
|
293 |
+
"How does the company's expense ratio compare to industry benchmarks?",
|
294 |
+
"Did the company incur any one-time or non-recurring expenses during the year?",
|
295 |
+
"Are there any trends or patterns in the company's cost structure over the past few years?",
|
296 |
+
"Has the company invested in research and development (R&D) or capital expenditures?",
|
297 |
+
"What were the employee-related costs and benefits provided by the company?",
|
298 |
+
"Are there any forecasts or projections for future cost management initiatives provided in the report?"
|
299 |
+
]
|
300 |
+
|
301 |
+
# Category 3: Profitability and Financial Ratios
|
302 |
+
profitability_ratios_questions = [
|
303 |
+
"What was the net profit or net income generated by the company during the fiscal year?",
|
304 |
+
"How does the profitability of the current year compare to the previous year?",
|
305 |
+
"What is the company's gross profit margin and how has it changed over time?",
|
306 |
+
"Did the company experience any changes in operating profit or operating margin?",
|
307 |
+
"What is the return on assets (ROA) and return on equity (ROE) for the company?",
|
308 |
+
"Has the company improved its profitability compared to its competitors in the industry?",
|
309 |
+
"Are there any trends or patterns in the company's profitability ratios over the past few years?",
|
310 |
+
"Did the company face any challenges or risks that impacted its profitability?",
|
311 |
+
"How does the company's profitability ratios compare to industry benchmarks?",
|
312 |
+
"Are there any forecasts or projections for future profitability provided in the report?"
|
313 |
+
]
|
314 |
+
|
315 |
+
# Category 4: Cash Flow and Liquidity
|
316 |
+
cash_flow_liquidity_questions = [
|
317 |
+
"What was the operating cash flow generated by the company during the fiscal year?",
|
318 |
+
"How does the cash flow from operations of the current year compare to the previous year?",
|
319 |
+
"Did the company experience any significant changes in its working capital management?",
|
320 |
+
"What were the major sources and uses of cash for the company during the year?",
|
321 |
+
"Has the company made any significant investments or divestments during the year?",
|
322 |
+
"How does the company's cash conversion cycle compare to industry benchmarks?",
|
323 |
+
"Are there any trends or patterns in the company's cash flow statement over the past few years?",
|
324 |
+
"What is the company's current ratio and quick ratio for assessing liquidity?",
|
325 |
+
"Did the company undertake any debt financing or equity financing activities?",
|
326 |
+
"Are there any forecasts or projections for future cash flow or liquidity provided in the report?"
|
327 |
+
]
|
328 |
+
# Define the options for the dropdown menu
|
329 |
+
categories = [
|
330 |
+
"Revenue and Sales",
|
331 |
+
"Expenses and Costs",
|
332 |
+
"Profitability and Financial Ratios",
|
333 |
+
"Cash Flow and Liquidity"
|
334 |
+
]
|
335 |
+
# Create a textbox to enter company's name
|
336 |
+
company_name = st.text_input("Enter the company's name:")
|
337 |
+
|
338 |
+
uploaded_file = st.file_uploader(f"Upload an Annual Report of {company_name} if available (PDF).", type=['pdf'])
|
339 |
+
toolkit = None
|
340 |
+
if uploaded_file is not None:
|
341 |
+
st.write("File uploaded successfully!")
|
342 |
+
file_contents = uploaded_file.read()
|
343 |
+
save_path = uploaded_file.name
|
344 |
+
with open(save_path, "wb") as f:
|
345 |
+
f.write(file_contents)
|
346 |
+
print(save_path)
|
347 |
+
loader = PyPDFLoader(save_path) #Step 1.1
|
348 |
+
documents = loader.load()
|
349 |
+
#1.2
|
350 |
+
text_splitter = CharacterTextSplitter(chunk_size=2000, chunk_overlap=0) #Splitting the text and creating chunks
|
351 |
+
docs = text_splitter.split_documents(documents)
|
352 |
+
if(mod=="OpenAI"):
|
353 |
+
embeddings = OpenAIEmbeddings()
|
354 |
+
if(mod=="Cohere"):
|
355 |
+
embeddings = CohereEmbeddings(cohere_api_key=api_key)
|
356 |
+
store = Chroma.from_documents(docs,embeddings)
|
357 |
+
vectorstore_info = VectorStoreInfo(
|
358 |
+
name="starbucks",
|
359 |
+
description="Starbucks financials",
|
360 |
+
vectorstore=store,
|
361 |
+
)
|
362 |
+
# llm = OpenAI(temperature=0.3)
|
363 |
+
toolkit = VectorStoreToolkit(llm=llm,vectorstore_info=vectorstore_info)
|
364 |
+
|
365 |
+
|
366 |
+
# Create a dropdown using the `selectbox` function
|
367 |
+
selected_category = st.selectbox("Select a category:", categories)
|
368 |
+
|
369 |
+
if(selected_category=="Revenue and Sales"):
|
370 |
+
selected_ques= st.selectbox("Select a category:", revenue_sales_questions)
|
371 |
+
if(selected_category=="Expenses and Costs"):
|
372 |
+
selected_ques= st.selectbox("Select a category:", expenses_costs_questions)
|
373 |
+
if(selected_category=="Profitability and Financial Ratios"):
|
374 |
+
selected_ques= st.selectbox("Select a category:", profitability_ratios_questions)
|
375 |
+
if(selected_category=="Cash Flow and Liquidity"):
|
376 |
+
selected_ques= st.selectbox("Select a category:", cash_flow_liquidity_questions)
|
377 |
+
|
378 |
+
# st.write(selected_ques)
|
379 |
+
|
380 |
+
ans=[]
|
381 |
+
if (st.button("Submit")):
|
382 |
+
output_res=[]
|
383 |
+
|
384 |
+
st.write("Company Name: " + company_name)
|
385 |
+
|
386 |
+
# llm = OpenAI(temperature=0.3)
|
387 |
+
tools = load_tools(["serpapi", "llm-math"], llm=llm)
|
388 |
+
agent = initialize_agent(llm = llm,
|
389 |
+
toolkit = toolkit,
|
390 |
+
agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
|
391 |
+
tools = tools,
|
392 |
+
verbose=True)
|
393 |
+
if(mod=='OpenAI'):
|
394 |
+
ans = financial_analyst(company_name)
|
395 |
+
st.write("Question asked by the user is " + selected_ques)
|
396 |
+
response = agent.run(selected_ques+f" Consider {ans}")
|
397 |
+
st.write(response)
|
398 |
+
st.write("Report")
|
399 |
+
st.write(ans)
|
400 |
+
prompt2="Make a line graph with Date as the x label and Closing Value as the y label and save the graph as an image file name of file as 'img2.png'"
|
401 |
+
prompt1="Make a line graph, x axis lables should be rotation = 90, save the graph as an image file and name of file as 'img1.png'"
|
402 |
+
graphs('stocks_data.csv',prompt1)
|
403 |
+
graphs('fin_data.csv',prompt2)
|
404 |
+
image_path = ['img1.png','img2.png']
|
405 |
+
generate_pdf(ans,image_path)
|
406 |
+
st.write("PDF generated successfully! Click below to download.")
|
407 |
+
# Download link
|
408 |
+
with open("output.pdf", "rb") as f:
|
409 |
+
st.download_button("Download PDF", f.read(), file_name="output.pdf", mime="application/pdf")
|
410 |
+
else:
|
411 |
+
try:
|
412 |
+
st.write("Question asked by the user is " + selected_ques)
|
413 |
+
response = agent.run(f"As a financial data analyst, your task is to thoroughly analyze \
|
414 |
+
the annual financial report of a company and provide accurate answers based solely on the data presented \
|
415 |
+
in the document. It is important to strictly adhere to the information provided in the report and \
|
416 |
+
refrain from making any assumptions or speculations. \
|
417 |
+
If necessary, you may utilize appropriate tools and formulas to derive the required answers.\
|
418 |
+
prompt = {selected_ques}")
|
419 |
+
print(response)
|
420 |
+
# response = agent.run(selected_ques)
|
421 |
+
st.write(response)
|
422 |
+
except:
|
423 |
+
st.write("Cohere Key Cannot give out the desired outputs. Pls provide OpenAI key for better results or try again!")
|