NAME

mok - an awk for molecules

SYNOPSIS

    mok [OPTION]...  'CODE' FILE...

DESCRIPTION

The purpose of mok is to read all the molecules found in the files that are given in the command line, and for each molecule execute the CODE that is given. The CODE is given in Perl and it has at its disposal all of the methods of the PerlMol toolkit.

This mini-language is intended to provide a powerful environment for writing "molecular one-liners" for extracting and munging chemical information. It was inspired by the AWK programming language by Aho, Kernighan, and Weinberger, the SMARTS molecular pattern description language by Daylight, Inc., and the Perl programming language by Larry Wall.

Mok takes its name from Ookla the Mok, an unforgettable character from the animated TV series "Thundarr the Barbarian", and from shortening "molecular awk". For more details about the Mok mini-language, see LANGUAGE SPECIFICATION below.

Mok is part of the PerlMol project, <http://www.perlmol.org>.

OPTIONS

-3: Generate 3D coordinates using Chemistry::3DBuilder.
-a: "Aromatize" each molecule as it is read. This is needed for example for matching SMARTS patterns that use aromaticity or ring primitives.
-b: Find bonds. Use it when reading files with no bond information but 3D coordinates to detect the bonds if needed (for example, if you want to do match a pattern that includes bonds). If the file has explicit bonds, mok will not try to find the bonds, but it will reassign the bond orders from scratch.
-c CLASS: Use CLASS instead of Chemistry::Mol to read molecules
-d: Delete dummy atoms after reading each molecule. A dummy atom is defined as an atom with an unknown symbol (i.e., it doesn't appear on the periodic table), or an atomic number of zero.
-D: Print debugging information, such as the way the input program was tokenized and parsed into blocks and subs. This may be useful for diagnosing syntax errors when the default error mesage is not informative enough.
-f FILE: Run the code from FILE instead of the command line
-h: Print usage information and exit
-p TYPE: Parse patterns using the specified TYPE. Default: 'smarts'. Other options are 'smiles' and 'midas'.
-t TYPE: Assume that every file has the specified TYPE. Available types depend on which Chemistry::File modules are installed, but currently available types include mdl, sdf, smiles, formula, mopac, pdb.

LANGUAGE SPECIFICATION

A Mok script consists of a sequence of pattern-action statements and optional subroutine definitions, in a manner very similar to the AWK language.

    pattern_type:/pattern/options { action statements }
    { action statements }
    sub name { statements }
    BEGIN { statements }
    END { statements }
    # comment

When the whole program consists of one unconditional action block, the braces may be omitted.

Program execution is as follows:

1) The BEGIN block is executed as soon as it's compiled, before any other actions are taken.

2) For each molecule in the files given in the command line, each pattern is applied in turn; if the pattern matches, the corresponding statement block is executed. The pattern is optional; statement blocks without a pattern are executed unconditionally. Subroutines are only executed when called explicitly.

3) Finally, the END block is executed.

The statements are evaluated as Perl statements in the Chemistry::Mok::UserCode::Default package. The following chemistry modules are conveniently loaded by default:

    Chemistry::Mol;
    Chemistry::Atom ':all';
    Chemistry::Bond;
    Chemistry::Pattern;
    Chemistry::Pattern::Atom;
    Chemistry::Pattern::Bond;
    Chemistry::File;
    Chemistry::File::*;
    Math::VectorReal ':all';

Besides these, there is one more function available for convenience: "println", which is defined by "sub println { print "\@_", "\n" }".

Pattern Specification

The pattern must be a SMARTS string readable by the Chemistry::File::SMARTS module, unless a different type is specified by means of the -p option or a pattern_type is given explicitly before the pattern itself. The pattern is given within slashes, in a way reminiscent of AWK and Perl regular expressions. As in Perl, certain one-letter options may be included after the closing slash. An option is turned on by giving the corresponing lowercase letter and turned off by giving the corresponding uppercase letter.

g/G: Match globally (default: off). When not present, the Mok interpreter only matches a molecule once; when present, it tries matching again in other parts of the molecule. For example, /C/ matches butane only once (at an unspecified atom), while /C/g matches four times (once at each atom).
o/O: Overlap (default: on). When set and matching globally, matches may overlap. For example, /CC/go pattern could match twice on propane, but /CC/gO would match only once.
p/P: Permute (default: off). Sometimes there is more than one way of matching the same set of pattern atoms on the same set of molecule atoms. If true, return these "redundant" matches. For example, /CC/gp could match ethane with two different permutations (forwards and backwards).

Special Variables

When blocks with action statements are executed, some variables are defined automatically. The variables are local, so you can do whatever you want with them with no side effects. However, the objects themselves may be altered by using their methods.

NOTE: Mok 0.10 defined $file, $mol, $match, and $patt in lowercase. While they still work, the lowercase variables are deprecated and may be removed in the future.

$FILE: The current filename.
$MOL: A reference to the current molecule as a Chemistry::Mol object.
$MATCH: A reference to the current match as a Chemistry::Pattern object.
$PATT: The current pattern as a string.
$FH: The current input filehandle. This provides low-level access in case you want to rewind or seek into the file, tell the current position, etc. Playing with $FH may break things if you are not careful. Use at your own risk!
@A: The atoms that were matched. It is defined as @A = $MATCH->atom_map if a pattern was used, or @A = $MOL->atoms within an unconditional block. Remember that this is a Perl array, so it is zero-based, unlike the one-based numbering used by most file types and some PerlMol methods.
@B: The bonds that were matched. It is defined as @A = $MATCH->bond_map if a pattern was used, or @A = $MOL->bonds within an unconditional block. Remember Remember that this is a Perl array, so it is zero-based, unlike the one-based numbering used by most file types and some PerlMol methods.

Special Blocks

Within action blocks, the following block names can be used with Perl funcions such as "next" and "last":

MATCH
BLOCK
MOL
FILE

EXAMPLES

Print the names of all the molecules found in all the .sdf files in the current directory:

    mok 'println $MOL->name' *.sdf

Find esters among *.mol; print the filename, molecule name, and formula:

    mok '/C(=O)OC/{ printf "$FILE: %s (%s)\n", 
        $MOL->name, $MOL->formula }' *.mol

Find out the total number of atoms:

    mok '{ $n += $MOL->atoms } END { print "Total: $n atoms\n" }' *.mol

Find out the average C-S bond length:

    mok '/CS/g{ $n++; $len += $B[0]->length }
        END { printf "Average C-S bond length: %.3f\n", $len/$n; }' *.mol

Convert PDB files to MDL molfiles:

    mok '{ $FILE =~ s/pdb/mol/; $MOL->write($FILE, format => "mdlmol") }' *.pdb

Find molecules with a given formula by overriding the formula pattern type globally (this example requires Chemistry::FormulatPattern):

    mok -p formula_pattern '/C6H12O6/{ println $MOL->name }' *.sdf

Find molecules with a given formula by overriding the formula pattern type just for one specific pattern. This can be used when more than one pattern type is needed in one script.

    mok 'formula_pattern:/C6H12O6/{ println $MOL->name }' *.sdf

VERSION

0.25

AUTHOR

Ivan Tubert-Brohman <itub@cpan.org>