constants.py

"""
Constants for the Antibody Developability Benchmark
"""

import os
from huggingface_hub import HfApi
import pandas as pd

ASSAY_LIST = ["AC-SINS_pH7.4", "PR_CHO", "HIC", "Tm2", "Titer"]
ASSAY_RENAME = {
    "AC-SINS_pH7.4": "Self-association",
    "PR_CHO": "Polyreactivity",
    "HIC": "Hydrophobicity",
    "Tm2": "Thermostability",
    "Titer": "Titer",
}
ASSAY_DESCRIPTION = {
    "AC-SINS_pH7.4": "Self association by AC-SINS at pH 7.4",
    "PR_CHO": "Polyreactivity by bead-based method against CHO SMP",
    "HIC": "Hydrophobicity by HIC",
    "Tm2": "Thermostability by nanoDSF",
    "Titer": "Titer by Valita",
}
ASSAY_EMOJIS = {
    "AC-SINS_pH7.4": "🧲",
    "PR_CHO": "🎯",
    "HIC": "💧",
    "Tm2": "🌡️",
    "Titer": "🧪",
}
ASSAY_HIGHER_IS_BETTER = {
    "HIC": False,
    "Tm2": True,
    "Titer": True,
    "PR_CHO": False,
    "AC-SINS_pH7.4": False,
}
# Tabs with emojis
ABOUT_TAB_NAME = "📖 About / Rules"
FAQ_TAB_NAME = "❓ FAQs"
SUBMIT_TAB_NAME = "✉️ Submit"

REGISTRATION_CODE = os.environ.get("REGISTRATION_CODE")
TERMS_URL = "https://euphsfcyogalqiqsawbo.supabase.co/storage/v1/object/public/gdpweb/pdfs/2025%20Ginkgo%20Antibody%20Developability%20Prediction%20Competition%202025-08-28-v2.pdf"
SLACK_URL = "https://join.slack.com/t/bitsinbio/shared_invite/zt-3dqigle2b-e0dEkfPPzzWL055j_8N_eQ"
TUTORIAL_URL = "https://huggingface.co/blog/ginkgo-datapoints/making-antibody-embeddings-and-predictions"
GITHUB_URL = "https://github.com/ginkgobioworks/abdev-benchmark"

# Input CSV file requirements
REQUIRED_COLUMNS: list[str] = [
    "antibody_name",
]
# Cross validation
CV_COLUMN = "hierarchical_cluster_IgG_isotype_stratified_fold"
# Sequence files
SEQUENCES_FILE_DICT = {
    "GDPa1": "data/GDPa1_v1.2_sequences.csv",
    "GDPa1_cross_validation": "data/GDPa1_v1.2_sequences.csv",
    "Heldout Test Set": "data/heldout-set-sequences.csv",
}
# GDPa1 dataset
GDPa1_path = "hf://datasets/ginkgo-datapoints/GDPa1/GDPa1_v1.2_20250814.csv"

# Huggingface API
TOKEN = os.environ.get("HF_TOKEN")
CACHE_PATH = os.getenv("HF_HOME", ".")
API = HfApi(token=TOKEN)

# Huggingface repos
ORGANIZATION = "ginkgo-datapoints"
SUBMISSIONS_REPO = f"{ORGANIZATION}/abdev-bench-submissions"
RESULTS_REPO = f"{ORGANIZATION}/abdev-bench-results"

# Leaderboard dataframes
LEADERBOARD_RESULTS_COLUMNS = [
    "user",
    "assay",
    "spearman",
    "dataset",
    "model",
    "submission_time",
]  # The columns expected from the results dataset
LEADERBOARD_DISPLAY_COLUMNS = [
    "user",
    "property",
    "spearman",
    "dataset",
    "model",
    "submission_time",
]  # After changing assay to property (pretty formatting)
LEADERBOARD_COLUMNS_RENAME = {
    "spearman": "Spearman Correlation",
    "dataset": "Dataset",
    "user": "User",
    "submission_time": "Submission Time",
    "model": "Model Name",
    "property": "Property",
}
BASELINE_USERNAMES = ["loodvanniekerkginkgo", "jmollerginkgobioworks", "sritter-ginkgobioworks"]


def LEADERBOARD_COLUMNS_RENAME_LIST(columns: list[str]) -> list[str]:
    return list(map(lambda x: LEADERBOARD_COLUMNS_RENAME.get(x, x), columns))


# First deadline: 2025-10-14 23:59:59 EST
FIRST_DEADLINE = pd.to_datetime("2025-10-14 23:59:59").tz_localize("US/Eastern")
SECOND_DEADLINE = pd.to_datetime("2025-11-18 23:59:59").tz_localize("US/Eastern")