paper_review.sh

# Set environment variables for the OpenAI API
export OPENAI_API_KEY="Bearer ak-90mn12op34qr56st78uv90wx12yz34ab"
export OPENAI_API_BASE="https://models-proxy.stepfun-inc.com/v1/chat/completions"

# Define variables for the research topic and paper location
topic="llm_in_context_learning"
topic_description="llm_in_context_learning_interpretability_cognitive_science"
# Specify the model to use for the review
review_model="deepseek-r1-0528-volce"


# Get the current file's directory (basefile)
basefile=$(cd `dirname $0`; pwd)

# Path to the PDF of the original paper
ori_paper_path="${basefile}/research_agent/paper_result/${topic}/${topic_description}/latex/iclrfake_conference.pdf"

# Directory to save the review results
output_paper_path="${basefile}/research_agent/paper_result/${topic}/${topic_description}/${review_model}/"

# Directory containing the paper's source files (one level up from the PDF)
data_dir="${basefile}/research_agent/paper_result/${topic}/${topic_description}/latex/"


echo "Processing $ori_paper_path"

cd "${basefile}/review_agent"

python3 run_paper_review_cli_single.py \
    --paper_path $ori_paper_path \
    --output_dir $output_paper_path \
    --openai_client_type openai \
    --data_dir $data_dir \
    --model_name $review_model