#!/usr/bin/env nextflow

nextflow.enable.dsl=2

// Pipeline parameters
params.pdb = 's3://omic/eureka/Pocketminer/1HSG.pdb'
params.outdir = 's3://omic/eureka/Pocketminer/output'
params.model_path = '/workspace/gvp/models/pocketminer'
params.debug = false

process POCKETMINER {
    container 'harbor.cluster.omic.ai/omic/pocketminer:v2'

    input:
        val pdb_path

    output:
        path "results/*", emit: all_results
        path "run.log", emit: log

    script:
        def debug_flag = params.debug ? '--debug' : ''
        def outdir = params.outdir
        """
        set +e
        touch run.log
        mkdir -p results

        echo "=== Environment Debug ===" >> run.log
        echo "Input path: ${pdb_path}" >> run.log
        echo "Output path: ${outdir}" >> run.log

        PDB_FILE=""

        # Method 1: Direct path (if PVC mounted or local)
        if [ -f "${pdb_path}" ]; then
            echo "Found at direct path" >> run.log
            cp "${pdb_path}" input.pdb && PDB_FILE="input.pdb"
        fi

        # Method 2: Convert s3://omic/... to /omic/... (PVC mount path)
        if [ -z "\$PDB_FILE" ]; then
            PVC_PATH="\$(echo '${pdb_path}' | sed 's|^s3://|/|')"
            if [ -f "\$PVC_PATH" ]; then
                echo "Found at PVC path: \$PVC_PATH" >> run.log
                cp "\$PVC_PATH" input.pdb && PDB_FILE="input.pdb"
            else
                echo "Not found at PVC path: \$PVC_PATH" >> run.log
            fi
        fi

        # Method 3: Download from MinIO using boto3
        if [ -z "\$PDB_FILE" ]; then
            ENDPOINT="\${AWS_ENDPOINT_URL:-http://datalake-hl.datalake.svc.cluster.local:9000}"
            echo "Downloading via boto3 from: \$ENDPOINT" >> run.log
            python -c "
import os, sys, boto3
from botocore.client import Config
s3_path = '${pdb_path}'.replace('s3://', '')
bucket, key = s3_path.split('/', 1)
endpoint = os.environ.get('AWS_ENDPOINT_URL', 'http://datalake-hl.datalake.svc.cluster.local:9000')
print(f'Bucket: {bucket}, Key: {key}')
s3 = boto3.client('s3', endpoint_url=endpoint,
    aws_access_key_id=os.environ.get('AWS_ACCESS_KEY_ID', ''),
    aws_secret_access_key=os.environ.get('AWS_SECRET_ACCESS_KEY', ''),
    config=Config(signature_version='s3v4'), region_name='us-east-1')
s3.download_file(bucket, key, 'input.pdb')
print(f'Downloaded {os.path.getsize(\"input.pdb\")} bytes')
" >> run.log 2>&1
            if [ -f input.pdb ] && [ -s input.pdb ]; then
                PDB_FILE="input.pdb"
            fi
        fi

        set -e

        if [ -z "\$PDB_FILE" ] || [ ! -f input.pdb ]; then
            echo "ERROR: Could not retrieve PDB file" >> run.log
            cat run.log
            exit 1
        fi

        PDB_BASENAME="\$(basename '${pdb_path}' .pdb)"

        python /workspace/entrypoint.py \\
            --pdb input.pdb \\
            --output-folder results \\
            --output-name "\$PDB_BASENAME" \\
            --model-path ${params.model_path} \\
            ${debug_flag} 2>&1 | tee -a run.log

        # Upload results to S3
        echo "Uploading results to ${outdir}..." >> run.log
        python -c "
import os, boto3
from botocore.client import Config
outdir = '${outdir}'.replace('s3://', '')
bucket, prefix = outdir.split('/', 1)
endpoint = os.environ.get('AWS_ENDPOINT_URL', 'http://datalake-hl.datalake.svc.cluster.local:9000')
s3 = boto3.client('s3', endpoint_url=endpoint,
    aws_access_key_id=os.environ.get('AWS_ACCESS_KEY_ID', ''),
    aws_secret_access_key=os.environ.get('AWS_SECRET_ACCESS_KEY', ''),
    config=Config(signature_version='s3v4'), region_name='us-east-1')
for f in os.listdir('results'):
    filepath = os.path.join('results', f)
    if os.path.isfile(filepath):
        key = f'{prefix}/{f}'
        s3.upload_file(filepath, bucket, key)
        print(f'Uploaded {f} -> s3://{bucket}/{key}')
" >> run.log 2>&1

        echo "Pipeline completed successfully" >> run.log
        """
}

workflow {
    POCKETMINER(params.pdb)
}