In a project I’m currently working on in streamlit, I feed Groq various datapoints on a given stock and have it analyze them. However, some parts of the text Groq simply doesnt separate with spaces and I have no idea how to fix this. Any and all help is appreciated.
You can see here how it sticks everything together at certain points in the analysis
Hi, what model are you using? And does this usually happen when rendering LaTeX?
If you have any example requests that reproduce the issue, that would be helpful.
submit_ai_overview = st.button(f'{ticker} AI Overview')
if submit_ai_overview:
try:
if ticker == '':
st.warning('Please enter a ticker symbol')
else:
yf_ticker = yf.Ticker(ticker)
info = yf_ticker.info
metrics = {
'Current Price': info.get('regularMarketPrice'),
'Previous Close': info.get('previousClose'),
'Open': info.get('open'),
'Days Low': info.get('dayLow'),
'Days High': info.get('dayHigh'),
'Fifty Two Week Low': info.get('fiftyTwoWeekLow'),
'Fifty Two Week High': info.get('fiftyTwoWeekHigh'),
'Volume': info.get('volume'),
'Average Volume': info.get('averageVolume'),
'Market Cap': info.get('marketCap'),
'Beta': info.get('beta'),
'PE Ratio': info.get('trailingPE'),
'EPS': info.get('trailingEps'),
'Target Price': info.get('targetMeanPrice'),
}
def format_value(val):
if val is None:
return 'N/A'
if isinstance(val, (int, float)):
return f"{val:,.2f}"
if isinstance(val, datetime):
return val.strftime('%b %d, %Y')
return str(val)
metrics_formatted = {k: format_value(v) for k, v in metrics.items()}
prompt = f"""
You are a financial analyst. I will provide you with the stock ticker and a set of key financial metrics.
Your task is to write a clear, detailed, and professional summary of the company's current financial condition and performance, followed by an in-depth investment analysis.
Base your response only on the provided data. If any important data is missing, clearly state the limitation.
Ticker Symbol: {ticker}
Key Financial Metrics:
{metrics_formatted}
Instructions for your analysis:
1. **Company Overview** — Briefly describe what the company does (if identifiable from the ticker symbol).
2. **Financial Health** — Discuss profitability, liquidity, leverage, and efficiency based on the metrics.
3. **Growth & Trends** — Identify trends from the given values and note whether the company appears to be growing, stable, or declining.
4. **Valuation** — If valuation metrics are included (like P/E ratio, Price/Sales), analyze if the stock might be overvalued or undervalued.
5. **Risks & Concerns** — Highlight any red flags or concerning financial ratios.
6. **Investment Outlook** — Provide a reasoned outlook for the stock based on the metrics.
Write in full sentences, and keep your tone objective and data-driven.
"""
try:
# noinspection PyTypeChecker
response = groq_client.chat.completions.create(
model="llama3-8b-8192",
messages=[
{"role": "system",
"content": """You are a financial analyst. I will provide you with the stock ticker and a set of key financial metrics.
Your task is to write a clear, detailed, and professional summary of the company's current financial condition and performance, followed by an in-depth investment analysis.
Base your response only on the provided data. If any important data is missing, clearly state the limitation."""},
{"role": "user", "content": prompt}
],
temperature=0
)
answer = response.choices[0].message.content.strip()
st.markdown(f"**AI Analysis:** {answer}")
except Exception as e:
st.error(f"AI request failed: {e}")Here’s a snippet of code from yesterday that consistently brought up the issue. Currently by switching from using metrics_formatted to multiple variables containing smaller amounts of data I’ve gotten it to work about 99% of the time (and when it does have concatenation issues its only a word or two). I’m currently using model llama3-8b-8192.