prepare_deploy_data.sh

#!/bin/bash
# Copies only the files needed by the dashboard from llm_project into data/
# Run once before pushing to HF Spaces.

set -e
PROJ="/data/ralmadamonter/llm_project"
DASH="/data/ralmadamonter/llm_dashboard/data"

echo "=== Copying evaluation results ==="
mkdir -p "$DASH/evaluation_llm_results"
cp -r "$PROJ/jsons/evaluation_llm_results/." "$DASH/evaluation_llm_results/"

echo "=== Copying SAMs data ==="
mkdir -p "$DASH/sams"
cp "$PROJ/jsons/sams/extracted_methods_sams_metrics.json" "$DASH/sams/"
cp "$PROJ/jsons/sams/sam_pdfs_enzymology.json"            "$DASH/sams/"
cp "$PROJ/jsons/sams/filtered.json"                        "$DASH/sams/"

echo "=== Copying azoreductases GT ==="
mkdir -p "$DASH/azoreductases_gt"
cp "$PROJ/jsons/azoreductases_gt/azo_metadata.json" "$DASH/azoreductases_gt/"

echo "=== Copying confusion matrix ==="
mkdir -p "$DASH/confusion_matrix"
cp "$PROJ/plots/classification_confusion_matrix/confusion_matrix_counts.csv" "$DASH/confusion_matrix/"

echo "=== Copying clustering (used files only — skipping .pkl, .png, .svg) ==="
find "$DASH/clustering" -name "*.pkl" -delete
find "$DASH/clustering" -name "*.png" -delete
find "$DASH/clustering" -name "*.svg" -delete
find "$DASH/clustering" -name "*_clusters_distribution_*.html" -delete
# Individual per-field CSVs (app uses ALL_FIELDS only)
find "$DASH/clustering" -name "*.csv" \
    ! -name "*_ALL_FIELDS.csv" \
    ! -name "*_FIELD_CLUSTER_METRICS.csv" \
    -delete

echo ""
echo "Done. Final data size:"
du -sh "$DASH"