mu-find

Langue: en

Version: 258085 (debian - 07/07/09)

Section: 1 (Commandes utilisateur)

NAME

mu-find - find messages based on search criteria

SYNOPSIS

mu-find [ options ] < search patterns >

DESCRIPTION

mu-find searches the databases of message information for messages matching specific patterns. The database is created and kept up-to-date with mu-index(1)

OPTIONS

(NB: see the CONFIGURATION section as well for using the configuration file)

General options

--home=,-h <dir> set the mu home directory; default is ~/.mu This directory is where the message database is stored, as well as configuration files and logs.

--log-stderr, -s write logging information to standard error instead of to <mu-home-directory>/mu-find.log , which is the default.

--log-append, -a append to the log file instead of overwriting it for every run, which is the default.

--debug, -d add a lot of logging for debugging purposes

Output options

Output options determine how mu-find will display its results. Each of the options can be selected by their full name ( text, sql, links or aggregate ), or by a one-letter shortcut (t,s,l,a, respectively); in other words, '-o s' would be a shorthand for '--output=sql', while '-o a' gets you some aggregate data (statistics) about the matched messages.

--output=text gives a textual display of the messages found; you can use the --format option to influence the way this is done.

--output=links generates a maildir with symbolic links to the messages founds. These can then be easily read with a maildir capable mail client, such as mutt(1). NB: The MUTT-INTEGRATION section of this manpage has some convenience macros for mutt integration.

By default, the maildir that is created is mu-found-<n> , where n is the lowest two-digit hexademical number [00..ff] for which there does not already exist a directory. Alternatively, you can use the --linkdir=<dir> option to generate the maildir with a specific name/path. The maildir should not exist yet.

--output=sql instead of searching for any messages, show the SQL query corresponding with the search criteria (search criteria are discussed later in this manpage). This can be useful for testing. For example,

 $ mu-find --output=sql s:foo f:bar
 
would give us:
 SELECT   m.recipients, m.cc, m.sender, m.subject, m.msg_id, 
          NULL AS dummy2, m.flags, m.mdate, m.msize, m.mpath, 
          m.priority, m.tstamp, m.id
 FROM     message m
 WHERE    m.subject LIKE '%foo%'
 AND      m.sender LIKE '%bar%'
 ORDER BY m.mdate DESC;
 

--output=aggregate show some aggregate information (statistics) about the matching messages, such as the top-5 oldest, newest, biggest messages in the set, and the most frequent sender, recipients. this is only marginally useful, as most of the results give us just the database IDs of the messages/contact found; however, these IDS's can be used to for further processing.

Also, included is the number of messages that match, and the average size. Example output:

 $ mu-find --output=aggregate test
 query statistics
 ================
 number of messages              : 90
 newest messages (msgids)        : 1392 509 501 498 490 
 oldest messages (msgids)        : 346 49 51 52 109 
 biggest messages (msgids)       : 842 1120 1053 683 346 
 avg message size (bytes)        : 199834
 most popular senders (cids)     : 548 456 56 644 181 
 most popular recipients (cids)  : 463 554 2 667 668 
 

Other options

--linkdir=, l <linkdir> set the maildir where mu-find should put the symlinks to any matched messages. Note, this directory must not exist yet.

--format=, -t <format> sets the way textual output is formatted. The format is a string consisting of letters corresponding to message fields. Any character not known is displayed as-is. For example, --text-format=d,f|s would display the date of message, a ','character, the sender (From:) of the message, a pipe-character and finally the subject of the message:

         1972-14-12 12:23,Someone|a star is born
         2008-08-13 17:30,foo@bar.cx|get rich quick
         ...
 
The fields corresponding to certain letters are the same as the ones in search expressions (scroll down for the list).

Note: database fields are not escaped with respect to format characters such as ',', so be careful when using this generate comma-separated files and the like. Also be careful to use alphabet letters are separators, as future versions of mu-find might add more fields.

If you do not specify any format, the default of d,f,s is chosen, or, show the 'date', 'from' and 'subject' fields, separated by commas.

Non-printable characters are displayed as '.'.

--sortfields, -S <sortfields>

and

--sortdir, -D a[scending]|d[escending]

set the way the found messages are sorted. The sortfields parameter sets the fields to sort by, one letter per field (s=subject, d=date and so on, see the section on search expressions to get the full list).

sortdir determines whether we sort ascending or descending.

Example (note: search expressions like 's:hello' are discussed in the next section):

  $ mu-find --sortfields=fz --sortdir=a s:hello
 

would sort messages ascending, with a subject containing 'hello' by sender, then by size.

The default is to sort descending by date. However,if any sortfield is explicitly specified, the default sort order is ascending.

SEARCH EXPRESSIONS

Search expressions are short descriptions that describe the messages to find. Search expresssions consist of a character that specifies some property, followed by a ':' and some value to match. The following fields are recognized:
 s: subject
 f: from (sender)
 t: to (recipient)
 c: cc: (carbon-copy recipient)
 m: message id
 p: message priority
 P: path to the message
 d: date
 z: size
 F: flags
 
 And two special types:
 x: xapian search
 B: bookmarks
 

Option x: for 'xapian search' searches the contents of messages, while the other options search for a certain property ('metadata') of the message. Xapian searching is discussed after the other options.

Option B: refers to bookmarks. You can define 'bookmarks' (shortcuts) for search expression. Bookmarks are discussed later in this document.

Words

Now, let's discuss search expressions where a certain message property contains a word. We have following word or textual types:
 s: subject
 f: from (sender)
 t: to (recipient)
 c: Cc: (carbon-copy recipient)
 m: message id
 p: message priority
 P: path to the message
 

For example, the search expression: s:rain matches all messages which have 'rain' in their subject (s:) (including training, brain, terrain etc.)

f: matches all messages sent by (From:), as well as all messages by. Search expressions can be combined - we can say f:bob 's:new york' to get all message from Bob about New York (search expressions with spaces need quotes).

Sometimes, we might want to match messages that don't contain as certain word in some property. To get all mails that do not contain 'banana' in their subject we can use the ^ prefix: s:^banana

And sometimes we might want to match a word exactly ie, we don't want a property to contain the search string, but we want it to be exactly equal. For that we can use the '=': s:=hello to match exactly that string (e.g., it won't match 'Hello World').

Please note that '=' and '^' only work for textual fields, and '^' only for textual fields and flags.

Numbers and intervals

Certain properties of messages are better expressed as numbers or intervals than as words, such as the message size or the message date:
 d: date
 z: size
 

Message size

First, let's look at the message size (the on-disk file size of the message). We could search for a message of size 500K with: z:500k ('k' for kilobyte and 'M' megabyte are supported). In many cases, an interval
  is probably more useful, so we can write: z:250k-750k to get messages between 250k and 750k, and z:1M- or z:-2M to get messages that are either bigger than 1M and messages that are smaller than 2M, respectively.

Message date

Another useful interval type is the message date. Dates can be specified in many different ways, and may have different meanings based on your locale. The recommended format therefore is YYYYMMDD, although other formats may work as well - provided the dates do not contain '-'-characters.

To get all mails in January 2005, we can write: d:20050101-20050131 and we can use the same interval notation as for the message size and write d:20080801- to get all mails received since August 2008, and d:-20033112 to get all mails from 2003 and before.

Single dates (such as d:20080202 ) are interpreted as an interval from the beginning of that day until the end of the day.

mu-find also recognizes a number followed by 'd', 'w', 'f', 'm', 'y' to mean some number of resp. days, weeks, fortnights, months or years ago since the present. Thus, d:3d- matches all messages from the last three days, while d:2y-1m matches all messages that are between 2 years and one month old.

Note that months are assumed to be 30 days, and years 365 days.

Message priority

The message priority is one of 'L' (low), 'N' (normal) or 'H' (high). To match all messages that are have either normal or high importance, the expression would be: p:NH and match only messages of low importance: p:L

Note, messages that do not explicitly specify their importance are assumed to be of normal priority.                 

Flags

The last type of search expression are message flags. Message either describe some property of the email file (file flags), or about the contents (content flags).

You can search message by the following file flags: new ('N'), replied ('R'), flagged ('F'), passed ('P'), seen ('S'), or combinations thereof. For example, to get all new messages: F:N or the get all flagged message you replied to: F:RF New ('N') messages don't have any other flags, the other flags can be combined.

Then, there are content flags: has-attachment ('a'), signed ('s') and encrypted ('x'). For example, you could search for all signed messages with an attachment with: F:as

You can also search for the absence of a flag, by prefixing the flag with '^' character. So for example, to get all encrypted messages without attachments, you can use: F:x^a

Xapian

mu uses the Xapian indexing library to search message contents. For the end-user, the use of Xapian is mostly a technical detail, except for the fact that Xapian searches follow the Xapian syntax: http://xapian.org/docs/queryparser.html

The syntax is very similar to what WWW search engines offer. For example, to get all the messages about either apples or bananas, the search expression would be:

x:apples OR bananas (and note that on the command-line you would need to quote that as: mu-find 'x:apples OR bananas' )

That would get all messages with either apples or bananas in the content -- that is, the message body, the subject field and the sender and recipient fields. More complex queries are possible, say, x:apple NEAR pear OR NOT (orange XOR coconut)

Xapian searches can be combined with other search options, so to get all message from Alice (alice@example.com) regarding unicorns, use: f:alice@example.com x:unicorns

NOTE: only up to 1000 documents are returned, and if you combine Xapian search with other search options, the result will be a subset of that. In other words, if you search query contains an 'x:' option, the max number of results is 1000.

BOOKMARKS

You can also define 'bookmarks' for often-used search expressions. In a file called bookmarks in your Mu home directory, you can specify them:
 [important]
 expr=p:H t:me@example.com f:boss@example.com
 
 [recentdrink]
 expr=d:1w- x:'wine OR beer OR whisky OR gin OR vodka'                   
 

You can then write search queries with the B: prefix, e.g. mu-find -B:important d:1m-

You can combine bookmarks with other search expressions; however, bookmarks cannot contain other bookmark expressions. This might change in the future.

CONFIGURATION

Instead of specifying the options on the command line, you can also specify them in the mu.conf configuration file, in the mu-home directory (by default, ~/.mu ). The General options go in the section [mu] while the mu-find specific options go under [mu-find]. For example, your configuration file could look something like this:
 [mu]
 debug=false
 
 [mu-info]
 format=f,d
 

Note that command line arguments take precedence over the configuration file.

MUTT INTEGRATION

mu-find can be integrated with the mutt(1) email client using some macros:
 macro index <F8> "<shell-escape>rm -rf ~/.mumatch; mu-find -o l -l ~/.mumatch " "mu-find"
 macro index <F9> "<change-folder-readonly>~/.mumatch\n" "display mu-find results"
 

Now, <F8> will perform a search, and <F9> will change to the folder with the results.

BUGS

Please report bugs when you find them: http://code.google.com/p/mu0/issues/list

AUTHOR

Dirk-Jan C. Binnema <djcb@djcbsoftware.nl>

SEE ALSO

mu-index(1), sqlite3(1), mutt(1)