Source code for matchms.filtering.metadata_processing.clean_compound_name

import logging
import re
from matchms.typing import SpectrumType


logger = logging.getLogger("matchms")


def clean_compound_name(spectrum_in: SpectrumType) -> SpectrumType:
    """Clean compound name.

    A list of frequently seen name additions that do not belong to the compound
    name will be removed.
    """
    if spectrum_in is None:
        return None

    spectrum = spectrum_in.clone()

    if spectrum.get("compound_name", None) is not None:
        name = spectrum.get("compound_name")
    else:
        assert spectrum.get("name", None) in [None, ""], ("Found 'name' but not 'compound_name' in metadata",
                                                          "Apply 'add_compound_name' filter first.")
        return spectrum

    # Clean compound name
    name_cleaned = _remove_parts_by_regular_expression(name)
    name_cleaned = _remove_known_non_compound_parts(name_cleaned)
    name_cleaned = _remove_misplaced_mass(name_cleaned)
    if name_cleaned != name:
        spectrum.set("compound_name", name_cleaned)
        logger.info("Added cleaned compound name: %s", name_cleaned)

    return spectrum


def _remove_parts_by_regular_expression(name: str):
    """Clean name string by removing known parts that don't belong there."""
    name = name.strip()
    # remove type NCGC00180417-03_C31H40O16_
    name = re.split(r"[A-Z]{3,6}[0-9]{8,12}-[0-9]{2,5}_[A-Z,0-9]{4,15}_", name)[-1]
    # remove type NCGC00160232-01! or MLS001142816-01!
    name = re.split(r"[A-Z]{3,6}[0-9]{8,12}-[0-9]{2,3}\!", name)[-1]
    # remove type Massbank:EA008813 option1|option2|option3
    name = re.split(r"((Massbank:)|(MassbankEU:))[A-Z]{2}[0-9]{5,6}.*\|", name)[-1]
    # remove type Massbank:EA008813 or MassbankEU:EA008813
    name = re.split(r"((Massbank:)|(MassbankEU:))[A-Z]{2}[0-9]{5,6}", name)[-1]
    # remove type HMDB:HMDB00943-1336
    name = re.split(r"HMDB:HMDB[0-9]{4,7}-[0-9]{1,7}", name)[-1]
    # remove type MoNA:662599
    name = re.split(r"MoNA:[0-9]{5,10}", name)[-1]
    # ReSpect:PS013405 option1|option2|option3...
    name = re.split(r"ReSpect:[A-Z]{2,3}[0-9]{6}.*\|", name)[-1]
    # ReSpect:PS013405 option1
    name = re.split(r"[A-Z]{2,3}[0-9]{6}( )", name)[-1]
    # remove type 0072_2-Mercaptobenzothiaz
    name = re.split(r"^[0-9]{4}_", name)[-1]
    # remove type nameofcompound_CID20_170920 or Spiraeoside_HCD30_170919
    name = re.split(r"_((HCD)|(CID))[0-9]{2}_[0-9]{5,6}$", name)[0]
    # Removes the collision energy from the compound name. Also allows for occurances of - 40.0 eV Unknown
    name = re.split(r"(?: - )?[0-9]+(?:\.[0-9]+)? ?[eE][Vv](?: Unknown)?$", name)[0]
    return name


def _remove_known_non_compound_parts(name: str):
    """Remove known non compound-name strings from name."""
    parts_remove = ["Spectral Match to",
                    "from NIST14",
                    "Massbank:"]
    for part in parts_remove:
        name = name.replace(part, "")
    return name.strip("; ")


def _remove_misplaced_mass(name: str):
    """Remove occasionally occurring parent mass addition to name."""
    regex_mass = r"^[0-9]{2,4}\.[0-9]$"
    end_part = name.split(" ")[-1]
    if re.search(regex_mass, end_part) is not None:
        return name.replace(end_part, "").strip()
    return name