Provided by: libwww-search-perl_2.51.90+~cs6.78-2_all bug


       AutoSearch -- a web-search tracking application


       AutoSearch [--stats] [--verbose] -n "Query Name" -s "query string" --engine engine [--mail] [--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


       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) occurred 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.


           AutoSearch -n 'LSAM Replication'
               -s '"lsam replication"'
               -e AltaVista

       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 '( or'
               -e AltaVista::AdvancedWeb
               -o coolness=hot

       This query does an advanced AltaVista search and specifies the (hypothetical) ``coolness''
       option to the search engine.


           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

           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

           In addition to all the normal file maintenance, print all new URLs to STDOUT, one per

       "-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.

           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.


       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 "".

       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.


       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

       A cron entry like:

           0 3 * * 1 /nfs/u1/wls/ /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/ /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:."


       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:

                       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.

                       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

                       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.

                       The text contained between the braces specifies a query options.  Multiple
                       occurrences of this command are allowed to specify multiple options.

                       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.

                       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,

       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

       Examples of these files can be seen in the pages under
       "", or in the output generated by a new


       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.


       For the library, see WWW::Search, for the perl regular expressions, see perlre.


       Wm. L. Scheding

       AutoSearch is a re-implementation of an earlier version written by Kedar Jog.


       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.



       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.

           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


       None known at this time; please inform the maintainer if any crop up.