DragonFly On-Line Manual Pages
AUTOSEARCH(1) User Contributed Perl Documentation AUTOSEARCH(1)
AutoSearch -- a web-search tracking application
AutoSearch [--stats] [--verbose] -n "Query Name" -s "query string"
--engine engine [--mail email@example.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
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
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"'
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'
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 started.
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
"-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.
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.
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
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.
If your outgoing mail server rejects email from certain users, you
can use this argument to set the From: header.
If the search engine requires a login/password (e.g.
Ebay::Completed), use this.
If the search engine requires a login/password (e.g. Ebay::Mature),
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
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
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:
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
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.
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 occurrencs 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
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.
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
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
Wm. L. Scheding
AutoSearch is a re-implementation of an earlier version written by
Copyright (C) 1996-1997 University of Southern California. All rights
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.
These are good ideas that people have suggested.
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 AutoSearch.
None known at this time; please inform the maintainer firstname.lastname@example.org
if any crop up.
perl v5.20.2 2015-08-30 AUTOSEARCH(1)