Parsing the MeSH ASCII format in Python

Similar to our previously published UniPruot parser, MeSH provides an ASCII format that can easily be parsed using Python.

Just like the UniProt parser, this function yields MeSH entries represented by dictionaries. Example code is included at the bottom of the file.

#!/usr/bin/env python3
A simple parse for the MeSH ASCII format, as downloadable from here:

For a reference on MeSH headings (i.e. keys in the resulting map), see

Originally published on
from collections import defaultdict

__author__ = "Uli Koehler"
__copyright__ = "Copyright 2015, Uli Kö221hler"
__license__ = "CC0 1.0 Universal"
__version__ = "1.0"

def readMeSH(fin):
    Given a file-like object, generates MeSH objects, i.e.
    dictionaries with a list of values for each qualifier.

    Example: {"MH": ["Acetylcysteine"]}

    currentEntry = None
    for line in fin:
        line = line.strip()
        if not line:
        # Handle new record. MeSH explicitly marks this
        if line == "*NEWRECORD":
            # Yiel old entry, initialize new one
            if currentEntry:
                yield currentEntry
            currentEntry = defaultdict(list)
        # Line example: "MH = Acetylcysteine"
        key, _, value = line.partition(" = ")
        # Append to value list
    # If there is a non-empty entry left, yield it
    if currentEntry:
        yield currentEntry

if __name__ == "__main__":
    # Example of how to use readMeSH()
    import argparse
    parser = argparse.ArgumentParser()
    args = parser.parse_args()
    with open(args.file, "r") as infile:
        # readMeSH() yields MeSH objects, i.e. dictionaries
        for entry in readMeSH(infile):