from pathlib import Path import os from dotenv import load_dotenv load_dotenv() PATH_DATA = Path.home() / "git_repos" / "CHI_corpus_analysis" / "data" PATH_RAW_CHI_METADATA = PATH_DATA / "CHI_raw.xls" PATH_CLEAN_CHI_METADATA = PATH_DATA / "CHI_filtered.parquet" PATH_CLEAN_CHI_METADATA_POSITIONS = PATH_DATA / "CHI_metadata.parquet" PATH_CLEAN_CHI_METADATA_CLUSTERS = PATH_DATA / "CHI_metadata_clusters.parquet" PATH_CLEAN_CHI_CLUSTERS_TOP_WORDS = PATH_DATA / "CHI_cluster_top_words.parquet" PATH_EMBEDDINGS = PATH_DATA / "embeddings.parquet" PATH_EMBEDDINGS_10d = PATH_DATA / "embeddings_10d.parquet" PATH_SPARSE_EMBEDDINGS = PATH_DATA / "sparse_embeddings.parquet" DENSE_VECTOR_NAME = "dense" SPARSE_VECTOR_NAME = "sparse" # why is v2 not supported? # ValueError: Model prithivida/Splade_PP_en_v2 is not supported in SparseTextEmbedding.Please check the supported models using `SparseTextEmbedding.list_supported_models()` SBERT_MODEL_NAME = "BAAI/bge-base-en-v1.5" SPARSE_MODEL_NAME = "prithivida/Splade_PP_en_v1" DEFAULT_QUERY = "doctors in participatory design" APP_NAME = "CHI papers search engine" CHROMA_DB_PATH = PATH_DATA / "chroma_db_vectors" QDRANT_DB_PATH = PATH_DATA / "qdrant_db_vectors" COLLECTION_NAME = "chi_collection" COLLECTION_HYBRID_NAME = "hybrid_collection" QDRANT_URL = os.getenv("QDRANT_URL", "http://qdrant:6333") QDRANT_KEY = os.getenv("QDRANT_KEY", None) MIN_COSINE_SCORE = float(os.getenv("MIN_COSINE_SCORE", 0.65)) CHROMA_COLLECTION_NAME = "chi_collection" SQL_DB_PATH_LOCAL = PATH_DATA / "metadata.db" METADATA_TABLE_NAME = "METADATA_TABLE" CLUSTER_TOP_WORDS_TABLE_NAME = "CLUSTER_TOP_WORDS" DISTANCE_METRIC = "cosine" SQL_HOST = os.getenv("SQL_HOST") SQL_PORT = os.getenv("SQL_PORT") SQL_DATABASE = os.getenv("SQL_DATABASE") SQL_USER_NAME = os.getenv("SQL_USER_NAME") SQL_PASSWORD = os.getenv("SQL_PASSWORD") SQL_EXTERNAL_URL = os.getenv("SQL_EXTERNAL_URL") SQL_INTERNAL_URL = os.getenv("SQL_INTERNAL_URL") SQL_BASE = f"{SQL_EXTERNAL_URL}" if SQL_EXTERNAL_URL else f'sqlite:///{SQL_DB_PATH_LOCAL}' "postgresql://chi_metadata_user:DJy5LJpogeOGD7ziw1i40zr2YmJ3Nq4z@dpg-d2bkj1re5dus738b67dg-a.frankfurt-postgres.render.com:5432/chi_metadata"