m[n]m
Distribution version 1.a
12/31/1999



                           ** UNIX Installation Instructions ** 


The distribution consists of the following files:


    mm.pl             manage INNOPAC loads
    mmftu.pl          extract MeSH headings from INNOPAC ftu report
    mmextract.pl      extract MARC records from NLM distribution file
    mm.ini            initialize system values
    mm.log            log of mm activities
    batch.html        sample INNOPAC procedures
    license           software license
    mmflow.gif        flowchart illustrating mm processes
    install.txt       this file
    


Installation
============

(1) Create a home directory for the programs. Add a subdirectory called "IO". 
Staff will need to ftp to and from the IO subdirectory.

(2) Copy distribution files to the home directory. All the *.pl files must be
executable; mm.log must be readable and writable, and mm.ini must be readable.

(3) If perl is anywhere other than /usr/bin/perl, the first line of each
perl script will have to be changed accordingly.

(4) ftp the current MARC MeSH distribution file from the NLM 
(http://www.nlm.nih.gov/marc) and put in the home directory. Remember it's a 
binary.

(5) Edit mm.ini to reflect local values. mm.ini controls the configuration and 
various options availabl. mm.ini is supposed to be self-documenting.



INNOPAC and M[n]M
================

mm.pl & mmftu.pl are INNOPAC specific. They are configured to work with INNOPAC style 
reports only. mmextract.pl, which can be run independently, is not INNOPAC specific; 
it will accept any correctly formatted list of MeSH headings. For more about INNOPAC, 
http://www.iii.com




Running the M[n]M system
=======================

See the file batch.html for staff oriented instructions. See the file mmflow.gif for 
flowchart of the procedures.

The following amounts to an outline of the system's functioning.

    (1) Staff generate INNOPAC ftu report (aka Report Heading Changes) and ftp 
    it to the IO directory.

    (2) mm.pl is invoked. It looks in the IO directory for any new ftu reports.
    If it finds one it:
      
        (i) invokes mmftu.pl on the report. mmftu.pl extracts new MeSH headings from
        the report.

        (ii) invokes mmextract.pl on the list of new MeSH headings. mmextract.pl
        looks each heading up in the NLM distribuiton file. If there is hit it
        writes the matching MARC record to a MARC load file in the IO directory.

        mmextract.pl writes a report on all its activities, including both a list
        of hits and non-hits to a "stats" file. Once mmextract.pl is done it sends 
		an email of the stats report to interested parties.

    (3) Staff ftp the MARC MeSH load file from the IO directory and load it to the 
    Innopac.


mm.pl can be invoked manually (presumably after one knows that a new ftu report has
been placed in the IO directory) or automatically via a cron job. 




Running mmextract.pl stand alone
--------------------------------

mmextract.pl can be run standalone, simply by invoking it from the command line, e.g.

     % ./mmextract.pl <search file> <load file> <stats file>
     

<search file> can be any ascii file of MeSH headings formatted:

      <heading>[<subfield delimiter<subheading>]

 e.g.,

      Neurodermatitis|xdiet therapy
      Roses
      Roses|xcultivation
      Roses|vinteractive multimedia


The resulting load file can be imported into any OPAC or other MARC system (provided that 
a reasonable means to import MARC records is provided).



*** eof ***