import os import json from tqdm import tqdm from suno_utils.utils.text import read_jsonl, write_jsonl if __name__ == "__main__": # read ids from the file with open( "/home/christian/code/christian/metadata/auk_id_to_artist_ids.json", "r" ) as f: auk_id_to_artist_ids = json.load(f) print(len(auk_id_to_artist_ids)) base_dir = "/app2/suno/data/diffusion/v0" for subset in ["val", "tr"]: metas_filename = f"metas_{subset}_aligned_v3.jsonl" print(os.path.join(base_dir, metas_filename)) metas = read_jsonl(os.path.join(base_dir, metas_filename)) print(len(metas)) new_metas = [] artist_count = 0 for meta in tqdm(metas): # check if this meta id is in the auk_id_to_artist_ids map if meta["id"] in auk_id_to_artist_ids: meta["artist_ids"] = auk_id_to_artist_ids[meta["id"]] artist_count += 1 new_metas.append(meta) print(len(new_metas), artist_count) write_jsonl( new_metas, os.path.join(base_dir, f"metas_{subset}_aligned_v3_artist_ids.jsonl"), ) print( f"Wrote {len(new_metas)} metas to {os.path.join(base_dir, f'metas_{subset}_aligned_v3_artist_ids.jsonl')}" )