charieswei
/
TradingAgents
mirror of https://github.com/TauricResearch/TradingAgents.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix(filter): replace tqdm with logger in batch news functions to fix I/O error 

tqdm writes to stderr immediately on __enter__, before any loop iteration.
In Streamlit's thread/subprocess context stderr can be a closed pipe, causing
'I/O operation on closed file' which _run_call catches and returns {} — so
the entire news enrichment step was silently skipped every run.

Replaced tqdm progress bars with logger.info() calls in:
- get_batch_stock_news_google() in openai.py
- get_batch_stock_news_openai() in openai.py
- Reddit DD parallel evaluation in reddit_api.py

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Youssef Aitousarrah 2026-02-21 13:55:13 -08:00
parent 21b33c6709
commit 61b731ac28
2 changed files with 82 additions and 89 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

111
tradingagents/dataflows/openai.py
View File

@ -113,18 +113,18 @@ def get_batch_stock_news_openai(
class PortfolioUpdate(BaseModel): class PortfolioUpdate(BaseModel):
items: List[TickerNews] items: List[TickerNews]
from tqdm import tqdm
client = _get_openai_client() client = _get_openai_client()
results = {} results = {}
total_batches = (len(tickers) + batch_size - 1) // batch_size
# Process in batches to avoid output token limits # Process in batches to avoid output token limits
with tqdm(total=len(tickers), desc="📰 OpenAI batch news", unit="ticker") as pbar: for i in range(0, len(tickers), batch_size):
for i in range(0, len(tickers), batch_size): batch = tickers[i : i + batch_size]
batch = tickers[i : i + batch_size] batch_num = i // batch_size + 1
logger.info(f"📰 OpenAI news batch {batch_num}/{total_batches}: {batch}")
# Request comprehensive news summaries for better ranker LLM context # Request comprehensive news summaries for better ranker LLM context
prompt = f"""Find the most significant news stories for {batch} from {start_date} to {end_date}. prompt = f"""Find the most significant news stories for {batch} from {start_date} to {end_date}.
Focus on business catalysts: earnings, product launches, partnerships, analyst changes, regulatory news. Focus on business catalysts: earnings, product launches, partnerships, analyst changes, regulatory news.
@ -135,32 +135,29 @@ For each ticker, provide a comprehensive summary (5-8 sentences) covering:
- Market reaction or implications - Market reaction or implications
- Any forward-looking statements or guidance""" - Any forward-looking statements or guidance"""
try: try:
completion = client.responses.parse( completion = client.responses.parse(
model="gpt-5-nano", model="gpt-5-nano",
tools=[{"type": "web_search"}], tools=[{"type": "web_search"}],
input=prompt, input=prompt,
text_format=PortfolioUpdate, text_format=PortfolioUpdate,
) )
# Extract structured output # Extract structured output
if completion.output_parsed: if completion.output_parsed:
for item in completion.output_parsed.items: for item in completion.output_parsed.items:
results[item.ticker.upper()] = item.news_summary results[item.ticker.upper()] = item.news_summary
else: else:
# Fallback if parsing failed # Fallback if parsing failed
logger.warning(f"Structured parsing returned None for batch: {batch}") logger.warning(f"Structured parsing returned None for batch: {batch}")
for ticker in batch:
results[ticker.upper()] = ""
except Exception as e:
logger.error(f"Error fetching batch news for {batch}: {e}")
# On error, set empty string for all tickers in batch
for ticker in batch: for ticker in batch:
results[ticker.upper()] = "" results[ticker.upper()] = ""
# Update progress bar except Exception as e:
pbar.update(len(batch)) logger.error(f"Error fetching batch news for {batch}: {e}")
# On error, set empty string for all tickers in batch
for ticker in batch:
results[ticker.upper()] = ""
return results return results
@ -218,15 +215,16 @@ def get_batch_stock_news_google(
).with_structured_output(PortfolioUpdate, method="json_schema") ).with_structured_output(PortfolioUpdate, method="json_schema")
results = {} results = {}
from tqdm import tqdm total_batches = (len(tickers) + batch_size - 1) // batch_size
# Process in batches # Process in batches
with tqdm(total=len(tickers), desc="📰 Google batch news", unit="ticker") as pbar: for i in range(0, len(tickers), batch_size):
for i in range(0, len(tickers), batch_size): batch = tickers[i : i + batch_size]
batch = tickers[i : i + batch_size] batch_num = i // batch_size + 1
logger.info(f"📰 Google news batch {batch_num}/{total_batches}: {batch}")
# Request comprehensive news summaries for better ranker LLM context # Request comprehensive news summaries for better ranker LLM context
prompt = f"""Find the most significant news stories for {batch} from {start_date} to {end_date}. prompt = f"""Find the most significant news stories for {batch} from {start_date} to {end_date}.
Focus on business catalysts: earnings, product launches, partnerships, analyst changes, regulatory news. Focus on business catalysts: earnings, product launches, partnerships, analyst changes, regulatory news.
@ -237,33 +235,30 @@ For each ticker, provide a comprehensive summary (5-8 sentences) covering:
- Market reaction or implications - Market reaction or implications
- Any forward-looking statements or guidance""" - Any forward-looking statements or guidance"""
try: try:
# Step 1: Perform Google search (grounded response) # Step 1: Perform Google search (grounded response)
raw_news = search_llm.invoke(prompt) raw_news = search_llm.invoke(prompt)
# Step 2: Structure the grounded results # Step 2: Structure the grounded results
structured_result = structured_llm.invoke( structured_result = structured_llm.invoke(
f"Using this verified news data: {raw_news.content}\n\n" f"Using this verified news data: {raw_news.content}\n\n"
f"Format the news for these tickers into the JSON structure: {batch}\n" f"Format the news for these tickers into the JSON structure: {batch}\n"
f"Include all tickers from the list, even if no news was found." f"Include all tickers from the list, even if no news was found."
) )
# Extract results # Extract results
if structured_result and hasattr(structured_result, "items"): if structured_result and hasattr(structured_result, "items"):
for item in structured_result.items: for item in structured_result.items:
results[item.ticker.upper()] = item.news_summary results[item.ticker.upper()] = item.news_summary
else: else:
logger.warning(f"Structured output invalid for batch: {batch}") logger.warning(f"Structured output invalid for batch: {batch}")
for ticker in batch:
results[ticker.upper()] = ""
except Exception as e:
logger.error(f"Error fetching Google batch news for {batch}: {e}")
# On error, set empty string for all tickers in batch
for ticker in batch: for ticker in batch:
results[ticker.upper()] = "" results[ticker.upper()] = ""
# Update progress bar except Exception as e:
pbar.update(len(batch)) logger.error(f"Error fetching Google batch news for {batch}: {e}")
# On error, set empty string for all tickers in batch
for ticker in batch:
results[ticker.upper()] = ""
return results return results

60
tradingagents/dataflows/reddit_api.py
View File

@ -342,7 +342,8 @@ def get_reddit_undiscovered_dd(
top_n: Annotated[int, "Number of top DD posts to return"] = 10, top_n: Annotated[int, "Number of top DD posts to return"] = 10,
num_comments: Annotated[int, "Number of top comments to include"] = 10, num_comments: Annotated[int, "Number of top comments to include"] = 10,
llm_evaluator=None, # Will be passed from discovery graph llm_evaluator=None, # Will be passed from discovery graph
) -> str: as_list: bool = False,
) -> str | list:
""" """
Find high-quality undiscovered DD using LLM evaluation. Find high-quality undiscovered DD using LLM evaluation.
@ -383,18 +384,19 @@ def get_reddit_undiscovered_dd(
if not submission.selftext or len(submission.selftext) < 200: if not submission.selftext or len(submission.selftext) < 200:
continue continue
# Get top comments for community validation
submission.comment_sort = "top"
submission.comments.replace_more(limit=0)
top_comments = [] top_comments = []
for comment in submission.comments[:num_comments]: if llm_evaluator:
if hasattr(comment, "body") and hasattr(comment, "score"): # Get top comments for community validation
top_comments.append( submission.comment_sort = "top"
{ submission.comments.replace_more(limit=0)
"body": comment.body[:1000], # Include more of each comment for comment in submission.comments[:num_comments]:
"score": comment.score, if hasattr(comment, "body") and hasattr(comment, "score"):
} top_comments.append(
) {
"body": comment.body[:1000], # Include more of each comment
"score": comment.score,
}
)
candidate_posts.append( candidate_posts.append(
{ {
@ -517,27 +519,11 @@ Extract all stock ticker symbols mentioned in the post or comments."""
return post return post
# Parallel evaluation with progress tracking # Parallel evaluation
try: logger.info(f"Scanning {len(candidate_posts)} Reddit posts with LLM...")
from tqdm import tqdm
use_tqdm = True
except ImportError:
use_tqdm = False
with ThreadPoolExecutor(max_workers=10) as executor: with ThreadPoolExecutor(max_workers=10) as executor:
futures = [executor.submit(evaluate_post, post) for post in candidate_posts] futures = [executor.submit(evaluate_post, post) for post in candidate_posts]
evaluated = [f.result() for f in as_completed(futures)]
if use_tqdm:
# With progress bar
evaluated = []
for future in tqdm(
as_completed(futures), total=len(futures), desc=" Evaluating posts"
):
evaluated.append(future.result())
else:
# Without progress bar (fallback)
evaluated = [f.result() for f in as_completed(futures)]
# Filter quality threshold (55+ = decent DD) # Filter quality threshold (55+ = decent DD)
quality_dd = [p for p in evaluated if p["quality_score"] >= 55] quality_dd = [p for p in evaluated if p["quality_score"] >= 55]
@ -559,6 +545,18 @@ Extract all stock ticker symbols mentioned in the post or comments."""
candidate_posts.sort(key=lambda x: x["full_length"] + (x["score"] * 10), reverse=True) candidate_posts.sort(key=lambda x: x["full_length"] + (x["score"] * 10), reverse=True)
top_dd = candidate_posts[:top_n] top_dd = candidate_posts[:top_n]
if as_list:
if not llm_evaluator:
import re
ticker_pattern = r"\$([A-Z]{2,5})\b|^([A-Z]{2,5})\s"
for post in top_dd:
matches = re.findall(ticker_pattern, post["title"] + " " + post["text"])
tickers = list(set([t[0] or t[1] for t in matches if t[0] or t[1]]))
post["ticker"] = tickers[0] if tickers else ""
post["quality_score"] = 75 # default to Medium priority
return top_dd
if not top_dd: if not top_dd:
return f"# Undiscovered DD\n\nNo high-quality DD found (scanned {len(candidate_posts)} posts)." return f"# Undiscovered DD\n\nNo high-quality DD found (scanned {len(candidate_posts)} posts)."