Output

is (naturally) handled by grazer (thus, all the options documented here belong to section grazer), which outputs warnings, progress and query result. Warnings and progress always go to standard output, result can be redirected to a file. Result is prolog, list of documents and epilog. Document is document id (file name and optionally offset in the file), document relevance and list of "clots", i. e. clusters of words found near to each other. Clot contains highlighted passages (found words) and default passages (their neighbors).

general
nametypeimplicit value
output_filestring""
If not empty, names the file into which result is printed.

spot_distanceint8
Maximal distance (in words) of found words which are grouped into one clot; must be between 2 and 1000.
Note that if it's big enough, relative to the size of found document (i. e. the document contains less than 2 * spot_distance words), the whole document is retrieved.

doc_limitint0
Maximum number of displayed documents; 0 means no limit.

max_highlightsint0
Maximal number of highlighted passages (basically found words) displayed in one clot; 0 means no limit.

show_relevanceintno
If set to yes, and if query contains just one soft operator (i. e. no boolean expressions), grazer shows metrics of every displayed document, so that you see why it was found and can tweak your operator.

You can change output formatting with templates.

templates of warning strings
Description Names only

templates of progress strings
Description Names only

templates of result strings
Description Names only

Example 1: settings for HTML output

output_file = "/home/locus/r.html"
prolog_template = "<html>\n<menu>\n"
epilog_template = "</menu>\n</html>\n"
doc_file_template = "<ol> <a href=%0>%0</a>\n\t<menu>\n"
doc_end_template = "</ol>\n"
metric_template = "\t<ol>%0 = %1</ol>\n"
relevance_template = "\t<ol><b>relevance = %0</b></ol>\n\t</menu>\n"
clot_head_template = ""
clot_break_template = ""
clot_tail_template = "\n"
passage_template = "%0%1%2"
highlighted_passage_template = "%0<b>%1</b>%2"
show_relevance = yes



Example 2: HTML with document parts

[locus]
additional_separators = "<>'"
doc_parts = "title"

[grazer]
title_parts = "title"
output_file = "/home/locus/r.html"
prolog_template = "<html>\n"
epilog_template = "</html>\n"

doc_file_template = "<h2><a href=%0 %2/a></h2>\n<UL>\n"
#Well, it may not be correct HTML, but it renders OK,
#at least for me... If you insist on portability,
#postprocess the grazer-generated file with some
#other program.

doc_end_template = "</UL>\n"
clot_head_template = "\t<LI>"
clot_break_template = ""
clot_tail_template = "</p>\n"
passage_template = "%0%1%2"
highlighted_passage_template = "%0<b>%1</b>%2"
show_relevance = no

[title]
push_rx = "<title>"
pop_rx = "</title>"
push_rx_is_case_sensitive = no
pop_rx_is_case_sensitive = no