#!/usr/bin/env python3

"""
Example: Query multiple TCGA projects at once.
"""

from pathlib import Path

from tcga_downloader.manifest import (
    ManifestRecord,
    format_manifest_stats,
    get_manifest_stats,
    write_manifest,
)
from tcga_downloader.query import query_multiple_projects


def main():
    projects = ["TCGA-BRCA", "TCGA-LUAD", "TCGA-COAD"]
    data_type = "Gene Expression"
    manifest_path = Path("multi_project_manifest.tsv")

    print(f"Querying {len(projects)} projects: {', '.join(projects)}...")
    print(f"Data type: {data_type}")

    hits = query_multiple_projects(projects, data_type)

    records = [
        ManifestRecord(
            file_id=h["file_id"],
            file_name=h["file_name"],
            data_type=h["data_type"],
            data_format=h["data_format"],
            size=int(h["file_size"]),
            md5=h["md5sum"],
        )
        for h in hits
    ]

    write_manifest(records, manifest_path, fmt="tsv")

    stats = get_manifest_stats(records)
    print("\nManifest Statistics:")
    print(format_manifest_stats(stats))


if __name__ == "__main__":
    main()