Provided by: libwww-search-perl_2.51.70-1_all 
NAME
AutoSearch -- a web-search tracking application
SYNOPSIS
AutoSearch [--stats] [--verbose] -n "Query Name" -s "query string" --engine engine [--mail you@where.com] [--options "opt=val"]... [--filter "filter"] [--host host] [--port port] [--userid bbunny --password c4rr0t5] [--ignore_channels KABC,KCBS,KNBC] qid AutoSearch --VERSION AutoSearch --help AutoSearch --man
DESCRIPTION
AutoSearch performs a web-based search and puts the results set in qid/index.html. Subsequent searches
(i.e., the second form above) AutoSearch determine what changes (if any) occured to the results sent
since the last run. These incremental changes are recorded in qid/YYYYMMDD.html.
AutoSearch is amenable to be run as a cron job because all the input parameters are saved in the web
pages. AutoSearch can act as a automated query agent for a particular search. The output files are
designed to be a set of web pages to easily display the results set with a web browser.
Example:
AutoSearch -n 'LSAM Replication'
-s '"lsam replication"'
-e AltaVista
replication_query
This query (which should be all on one line) creates a directory replication_query and fills it with the
fascinating output of the AltaVista query on "lsam replication", with pages titled ``LSAM Replication''.
(Note the quoting: the single quotes in '"lsam replication"' are for the shell, the double quotes are
for AltaVista to search for the phrase rather than the separate words.)
A more complicated example:
AutoSearch -n 'External Links to LSAM'
-s '(link:www.isi.edu/lsam or link:www.isi.edu/~lsam) -url:isi.edu'
-e AltaVista::AdvancedWeb
-o coolness=hot
This query does an advanced AltaVista search and specifies the (hypothetical) ``coolness'' option to the
search engine.
OPTIONS
"qid"
The query identifer specifies the directory in which all the files that relate to this query and
search results will live. It can be an absolute path, or a relative path from cwd. If the directory
does not exist, it will be created and a new search started.
"--stats"
Show search statistics: the query string, number of hits, number of filtered hits, filter string,
number of suspended (deleted) hits, previous set size, current set size, etc.
"-v" or "--verbose"
Verbose: output additional messages and warnings.
"-n" or "--qn" or "--queryname"
Specify the query name. The query name is used as a heading in the web pages, therefore it should be
a 'nice' looking version of the query string.
"-s" or "--qs" or "--querystring"
Specify the query string. The query string is the character string which will be submitted to the
search engine. You may include special characters to group or to qualify the search.
"-e" or "--engine"
Specify the search engine. The query string will be submitted to the user specified search engine.
In many cases there are specialized versions of search engines. For example, AltaVista::AdvancedWeb
and AltaVista::News allow more powerful and Usenet searches. See AltaVista or the man page for your
search engine for details about specialized variations.
"--listnewurls"
In addition to all the normal file maintenance, print all new URLs to STDOUT, one per line.
"-o" or "--options"
Specify the query options. The query options will be submitted to the user search engine with the
query string. This feature permits modification of the query string for a specific search engine or
option. More than one query option may be specified.
Example: "-o what=news" causes AltaVista to search Usenet. Although this works, the preferred
mechanism in this case would be "-e AltaVista::News" or "-e AltaVista::AdvancedNews". Options are
intended for internal or expert use.
"-f" or "--uf" or "--urlfilter"
This option specifies a regular expression which will be compared against the URLs of any results; if
they match the case-insensitive regular expression, they will be removed from the hit set.
Example: "-f '.*\.isi\.edu'" avoids all of ISI's web pages.
"--cleanup i"
Delete all traces of query results from more than i days ago. If --cleanup is given, all other
options other than the qid will be ignored.
"--cmdline"
Reconstruct the complete command line (AutoSearch and all its arguments) that was used to create the
query results. Command line will be shown on STDERR. If --cmdline is given, all other options other
than the qid will be ignored.
"--mail user@address" or "-m user@address"
After search is complete, send email to that user, listing the NEW results. Email is HTML format.
Requires the Email::Send and related modules. If you send email through an SMTP server, you must set
environment variable SMTPSERVER to your server name or IP address. If your SMTP server requires
password, you must set environment variables SMTPUSERNAME and SMTPPASSWORD. If you send email via
sendmail, you should set environment variable SENDMAIL if the sendmail executable is not in the path.
"--emailfrom user@address"
If your outgoing mail server rejects email from certain users, you can use this argument to set the
From: header.
"--userid bbunny"
If the search engine requires a login/password (e.g. Ebay::Completed), use this.
"--password Carr0t5"
If the search engine requires a login/password (e.g. Ebay::Mature), use this.
DESCRIPTION
AutoSearch submits a query to a search engine, produces HTML pages that reflect the set of 'hits'
(filtered search results) returned by the search engine, and tracks these results over time. The URL and
title are displayed in the qid/index.html, the URL, the title, and description are displayed in the
'weekly' files.
To organize these results, each search result is placed in a query information directory (qid). The
directory becomes the search results 'handle', an easy way to track a set of results. Thus a qid of
"/usr/local/htdocs/lsam/autosearch/load_balancing" might locate the results on your web server at
"http://www.isi.edu/lsam/autosearch/load_balancing".
Inside the qid directory you will find files relating to this query. The primary file is index.html,
which reflects the latest search results. Every not-filtered hit for every search is stored in
index.html. When a hit is no longer found by the search engine it a removed from index.html. As new
results for a search are returned from the search engine they are placed in index.html.
At the bottom of index.html, there is a heading "Weekly Search Results", which is updated each time the
search is submitted (see "AUTOMATED SEARCHING"). The list of search runs is stored in reverse
chronological order. Runs which provide no new information are identified with
No Unique Results found for search on <date>
Runs which contain changes are identified by
Web search results for search on <date>
which will be linked a page detailing the changes from that run.
Detailed search results are noted in weekly files. These files are named YYYYMMDD.html and are stored in
the qid directory. The weekly files include THE URL, title, and a the description (if available). The
title is a link to the original web page.
AUTOMATED SEARCHING
On UNIX-like systems, cron(1) may be used to establish periodic searches and the web pages will be maintained by AutoSearch. To establish the first search, use the first example under SYNOPSIS. You must specify the qid, query name and query string. If any of the items are missing, you will be interactively prompted for the missing item(s). Once the first search is complete you can re-run the search with the second form under SYNOPSIS. A cron entry like: 0 3 * * 1 /nfs/u1/wls/AutoSearch.pl /www/div7/lsam/autosearch/caching might be used to run the search each Monday at 3:00 AM. The query name and query string may be repeated; but they will not be used. This means that with a cron line like: 0 3 * * 1 /nfs/u1/wls/AutoSearch.pl /www/div7/lsam/autosearch/caching -n caching -s caching a whole new search series can be originated by rm -r /www/div7/lsam/autosearch/caching However, the only reason to start a new search series would be to throw away the old weekly files. We don't recommend running searches more than once per day, but if so the per-run files will be updated in-place. Any changes are added to the page with a comment that "Recently Added:"; and deletions are indicated with "Recently Suspended:."
CHANGING THE LOOK OF THE PAGES
The basic format of these two pages is simple and customizable. One requirement is that the basic
structure remain unchanged. HTML comments are used to identify sections of the document. Almost
everything can be changed except for the strings which identify the section starts and ends.
Noteworthy tags and their meaning:
<!--Top-->.*<!--/Top-->
The text contained within this tag is placed at the top of the output page. If the text
contains AutoSearch WEB Searching, then the query name will replace it. If the text does
not contain this magic string and it is the first ever search, the user will be asked for
a query name.
<!--Query{.*}/Query-->
The text contained between the braces is the query string. This is how AutoSearch
maintains the query string. You may edit this string to change the query string; but
only in qid/index.html. The text ask user is special and will force AutoSearch to
request the search string from the user.
<!--SearchEngine{.*}/SearchEngine-->
The text contained between the braces is the search engine. Other engines supported are
HotBot and Lycos. You may edit this string to change the engine used; but only in
qid/index.html. The text ask user is special and will force AutoSearch to to request the
search string from the user.
<!--QueryOptions{.*}/QueryOptions-->
The text contained between the braces specifies a query options. Multiple occurrencs of
this command are allowed to specify multiple options.
<!--URLFilter{.*}/URLFilter-->
The text contained between the braces is the URL filter. This is how AutoSearch
maintains the filter. Again you may edit this string to change the query string; but
only in qid/index.html. The text ask user is special and will force AutoSearch to ask
the user (STDIN) for the query string. When setting up the first search, you must edit
first_index.html, not qid/index.html. The URL filter is a standard perl5 regular
expression. URLs which do not match will be kept.
<!--Bottom-->.*<!--/Bottom-->
The text contained within this tag is placed at the bottom of the output page. This is a
good place to put navigation, page owner information, etc.
The remainder of the tags fall into a triplet of ~Heading, ~Template, and ~, where ~ is Summary, Weekly,
Appended, and Suspended. The sub-sections appear in the order given. To produce a section AutoSearch
outputs the heading, the template, the section, n copies of the formatted data, and an /section. The
tags and their function are:
~Heading The heading tag identifies the heading for a section of the output file. The
SummaryHeading is for the summary portion, etc. The section may be empty (e.g.,
Suspended) and thus no heading is output.
~Template The template tag identifies how each item is to be formatted. Simple text replacement is
used to change the template into the actual output text. The text to be replaced is
noted in ALLCAPS.
~ This tag is used to locate the section (Summary, Weekly, etc.). This section represents
the actual n-items of data.
You can edit these values in the qid/index.html page of an existing search. The file first_index.html
(in the directory above qid) will be used as a default template for new queries.
Examples of these files can be seen in the pages under "http://www.isi.edu/lsam/tools/autosearch/", or in
the output generated by a new AutoSearch.
FILES
first_index.html optional file to determine the default format of the index.html file of a new query.
first_date.html optional file to determine the default format of the YYYYMMDD.html file for a new
query.
qid/index.html (automatically created) latest search results, and reverse chronological list of
periodic searches.
qid/date.html file used as a template for the YYYYMMDD.html files.
qid/YYYYMMDD.html (automatically created) summary of changes for a particular date (AKA 'Weekly' file).
Optional files first_index.html and first_date.html are used for the initial search as a template for
qid/index.html and date.html, respectively. If either of these files does not exist; a default-default
template is stored within the AutoSearch source. The intention of these two files is to permit a user to
establish a framework for a group of search sets which have a common format. By leaving the default
query name and query string alone, they will be overridden by command line inputs.
SEE ALSO
For the library, see WWW::Search, for the perl regular expressions, see perlre.
AUTHORS
Wm. L. Scheding
AutoSearch is a re-implementation of an earlier version written by Kedar Jog.
COPYRIGHT
Copyright (C) 1996-1997 University of Southern California. All rights reserved.
Redistribution and use in source and binary forms are permitted provided that the above copyright notice
and this paragraph are duplicated in all such forms and that any documentation, advertising materials,
and other materials related to such distribution and use acknowledge that the software was developed by
the University of Southern California, Information Sciences Institute. The name of the University may
not be used to endorse or promote products derived from this software without specific prior written
permission.
THIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, WITHOUT
LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
DESIRED FEATURES
These are good ideas that people have suggested.
URL validation.
Validate the status of each URL (with HTTP HEAD requests) and indicate this status in the output.
Multi-search.
It should be possible to merge the results of searches from two search-engines. If this merger were
done as a new search engine, this operation would be transparent to AutoSearch.
BUGS
None known at this time; please inform the maintainer mthurn@cpan.org if any crop up.