patt – locate and output user-defined patterns in a Humdrum input
`` patt [-m] -f templatefile [-s regexp] [-t output_tag] [inputfile ...]``
The patt command is used to locate occurrences of a user-defined pattern in some Humdrum input. The patterns sought may span innumerable data records. Occurrences of the pattern may be identified by line number, echoed intact in the output stream, or tagged by a user-defined marker in a **patt output spine.
The pattern sought must be defined as a separate “template” file. The template file is identified using the -f command option.
Pattern templates consist of one or more records. Each record specifies a regular expression pattern. The input is scanned from beginning to end, in order to find passages that match the defined template. In order for a match to take place, successive records in the input stream must match the regular expressions given in each of the corresponding records of the template. For example, if the template is 9 lines in length, then the input stream must contain 9 successive lines of matching data in order for a pattern match to be successful.
The patt command implements a full UNIX regular expression syntax. Each line in the template file represents an independent regular expression. For example, the template:
1 2 3
will match inputs such as the following:
1 112 43.9
0x.=%&1* Figure 32 abc(...32...)
A more circumspect regular expression template might look like this:
^1$ ^2$ ^3$
(The caret (^) and dollar sign ($) are regular expression anchors that indicate the beginning of the record and end of the record respectively.) Refer to regexp (6) for further information regarding regular expressions. Note that the related **pattern** command implements a more powerful regular expression syntax for defining multi-record patterns.
The patt command supports the following options:
-c makes pattern-matching sensitive to comments -e echoes matched patterns in the output -f templatefile use pattern specified in templatefile -h displays a help screen summarizing the command syntax -m invokes collapsed multiple-record matching mode -s regexp skip (ignore) data records containing the defined regular expression -t output_tag generate **patt output spine; tag each occurrence of the pattern with the string output_tag
Options are specified in the command line.
By default, the patt command is insensitive to the presence or absence of Humdrum comments. Pattern searches may be made sensitive to occurrences of comments by specifying the -c option.
In the default operation, patt outputs a Humdrum global comment for each pattern matched in the input. Each comment identifies the line number in the input where the found pattern begins.
With the -e option, each instance of the found pattern is echoed in the output. Each output pattern is preceded by the appropriate exclusive interpretation(s) and followed by appropriate spine-path terminator(s).
Certain types of data records may be ignored in the pattern-search by invoking the -s (skip) option. This option must be accompanied by a user-define regular expression. All input data records matching the regular expression are ignored. This option is useful, for example, in skipping null data tokens, barlines, marked embellishment tones, or other types of data.
The -m option invokes a multiple record matching mode. In this mode, patt attempts to match as many successive regular expressions in the template file as possible for a given input record, before continuing with the next input and template records. In this way, a template file consisting of several records, may possibly match a single input record. (See EXAMPLES below.)
The -t option causes patt to output a single spine of **patt output. The user specifies an output tag (character string) on the command line. Each instance of the found pattern causes the tag to be output in the **patt spine at the position corresponding to the onset of each found pattern. (See EXAMPLES below.) Note that the -t and -e options are mutually exclusive.
Whatever options are invoked, patt always produces output that conforms to the Humdrum syntax.
The following examples illustrate the operation of the patt command. Consider the following target file and pattern template file:
A simple search for the pattern template might use the command:
`` patt -f template target``
!! Pattern found at line 5 of file input !! Pattern found at line 6 of file input
In the first instance, notice that patt is able to identify overlapping patterns. Two instances of the 1-2-3 pattern are identified beginning on the fifth and sixth lines respectively.
Note however, that the first instance of the pattern (beginning at line 2) was not identified due to the interruption of the common system barline in the fourth line. The barlines can be ignored by invoking the -s option, followed by a regular expression that uniquely identifies the records to be skipped – in this case the equals sign. The command:
`` patt -s = -f template target``
!! Pattern found at line 2 of file input !! Pattern found at line 5 of file input !! Pattern found at line 6 of file input
Actual instances of the pattern can be output by invoking the -e (echo) option:
`` patt -e -s = -f template target``
!! Pattern found at line 2 of file input **foo 1 2 =1 1 3 *- !! Pattern found at line 5 of file input **foo 1 3 2 1 3 2 *- !! Pattern found at line 6 of file input **foo 2 1 3 2 3 *-
Notice that each pattern found is output as a self-contained Humdrum output with initial exclusive interpretations and concluding spine-path terminators. If a single continuous output is desired, the rid -u command may be used to eliminate the duplicate interpretations.
Instead of outputting the individual patterns, the -t option may be used to output a spine that marks each instance of the found pattern. In the following command, the beginning of each occurrence of the pattern is labelled in the **patt spine by the tag “one-two-three.”
`` patt -t one-two-three -s = -f template target``
**foo **patt 1 one-two-three 2 . =1 . 1 3 one-two-three 2 1 one-two-three 3 2 . 3 . =2 . 1 . 2 3 . 4 . 2 3 1 . *- *-
For some tasks (such as the identification of tone-rows in 12-tone music), nominally “successive” elements of the pattern may be collapsed within a single sonority or record. The -m option invokes a multiple record matching mode. By way of example, the following command:
`` patt -m -t theme123 -s = -f template target``
**foo **patt 1 theme123 2 . =1 . 1 3 theme123 2 1 theme123 3 2 . 3 . =2 . 1 theme123 2 3 . 4 . 2 3 1 theme123 *- *-
Note that in the above examples, the extensive capabilities provided by patt for defining complex regular expressions have not been illustrated. Refer to regexp (6) for further pertinent information.
DOS 2.0 and up, with the MKS Toolkit. OS/2 with the MKS Toolkit. UNIX systems supporting the Korn shell or Bourne shell command interpreters, and revised awk (1985).
If a comment is present in the template pattern, failing to specify the -c option will make pattern matching a logically impossibility.
The patt command differs from the related pattern command in the following ways: (1) patt always produces output conforming to the Humdrum syntax whereas pattern never produces Humdrum output. (2) pattern allows multi-record `wild cards’ in the template file, and so permits the creation of more sophisticated regular expressions. (3) The pattern command does not directly provide an “echo” option.
- **Pertinent description in the Humdrum User Guide**
- **Index to Humdrum Commands**
- **Table for Contents for Humdrum User Guide**