code/scripts/collab_phase_e_eval_b_prime.sh

#!/bin/bash
# Run config (b'): planner-INDEP + validator SFT + fixer ORPO.
# This isolates the planner-collab vs planner-indep effect at matched fixer level
# when compared to config (d).

set -e
cd /home/datht/mats-sql-tist

EVAL_INPUT=data/sft_bird_with_evidence_dev_text2sql.json
[ ! -f "$EVAL_INPUT" ] && EVAL_INPUT=data/sft_bird_dev_with_evidence_text2sql_new.json

P_INDEP=alignment-handbook/output/qwen-coder0.5b-bird-planner-INDEP-orpo
V_SFT=alignment-handbook/output/qwen-coder0.5b-bird-validator-sft
F_ORPO=alignment-handbook/output/qwen-coder0.5b-bird-fixer-orpo

VLLM=/home/datht/anaconda3/envs/llm/bin/vllm

pkill -f "vllm serve" 2>/dev/null || true
sleep 6

CUDA_VISIBLE_DEVICES=0 $VLLM serve "$P_INDEP" \
    --served-model-name planner --port 8100 --dtype bfloat16 \
    --gpu-memory-utilization 0.45 --max-model-len 8192 \
    > /tmp/collab_eval_p.log 2>&1 &
CUDA_VISIBLE_DEVICES=1 $VLLM serve "$V_SFT" \
    --served-model-name validator --port 8101 --dtype bfloat16 \
    --gpu-memory-utilization 0.42 --max-model-len 8192 \
    > /tmp/collab_eval_v.log 2>&1 &
CUDA_VISIBLE_DEVICES=1 $VLLM serve "$F_ORPO" \
    --served-model-name fixer --port 8102 --dtype bfloat16 \
    --gpu-memory-utilization 0.42 --max-model-len 8192 \
    > /tmp/collab_eval_f.log 2>&1 &

for url in http://localhost:8100/v1/models http://localhost:8101/v1/models http://localhost:8102/v1/models; do
    for i in {1..120}; do
        curl --noproxy localhost -fs $url >/dev/null 2>&1 && break
        sleep 5
    done
done

OUT=eval_results/collab_b_prime_planner_indep_with_fixer_orpo_bird_dev.jsonl
PYTHONPATH=. /home/datht/anaconda3/envs/mats/bin/python scripts/run_pipeline_rollouts.py \
    --input_file "$EVAL_INPUT" \
    --output_file "$OUT" \
    --planner_host http://localhost:8100 \
    --validator_host http://localhost:8101 \
    --fixer_host http://localhost:8102 \
    --K 1 --K_val 1 --K_fix 1 --temperature 0.0 --top_p 1.0 \
    --max_questions -1 --n_threads 8

/home/datht/anaconda3/envs/mats/bin/python -c "
import json
correct = total = 0
with open('$OUT') as f:
    for line in f:
        d = json.loads(line)
        for t in d.get('trajectories', []):
            total += 1
            correct += int(t.get('is_fixed_correct', False))
print(f'b_prime EX = {correct}/{total} = {100*correct/max(total,1):.2f}%')
"

pkill -f "vllm serve" 2>/dev/null || true
echo "DONE_B_PRIME_EVAL"