Your company here ā€” click to reach over 10,000 unique daily visitors

recollindex - Man Page

indexing command for the Recoll full text search system


recollindex -h
recollindex [ -z|-Z ] [ -k ] [ --nopurge ] [ -P ] [ --diagsfile <diagpath> ]
recollindex -m [ -w <secs>] [ -D ] [ -x ] [ -C ] [ -n|-k ]
recollindex -i [ -Z -k -f -P ] [<path [path ...]>]
recollindex -r [ -Z -K -e -f ] [ -p pattern ] <dirpath>
recollindex -e [<path [path ...]>]
recollindex -l|-S|-E
recollindex -s <lang>
recollindex --webcache-compact
recollindex --webcache-burst <destdir>
recollindex --notindexed [path [path ...]]


Create or update a Recoll index.

There are several modes of operation. All modes support an optional -c <cfgdir> option to specify the configuration directory name, overriding the default or $RECOLL_CONFDIR (or $HOME/.recoll by default).

The normal mode will index the set of files described in the configuration. This will incrementally update the index with files that changed since the last run. If option -z is given, the index will be erased before starting. If option -Z is given, the index will not be reset, but all files will be considered as needing reindexing (in place reset).

recollindex does not process again files which previously failed to index (for example because of a missing helper program). If option -k is given, recollindex will try again to process all failed files. Please note that recollindex may also decide to retry failed files if the auxiliary checking script defined by the "checkneedretryindexscript" configuration variable indicates that this should happen.

The --nopurge option will disable the normal erasure of deleted documents from the index. This can be useful in special cases (when it is known that part of the document set is temporarily not accessible).

The -P option will force the purge pass. This is useful only if the idxnoautopurge parameter is set in the configuration file.

If the option --diagsfile is given, the path given as parameter will be truncated and indexing diagnostics will be written to it. Each line in the file will have a diagnostic type (reason for the file not to be indexed), the file path, and a possible additional piece of information, which can be the MIME type or the archive internal path depending on the issue. The following diagnostic types are currently defined:

Skipped : the path matches an element of skippedPaths or skippedNames.

NoContentSuffix : the file name suffix is found in the noContentSuffixes list.

MissingHelper : a helper program is missing.

Error : general error (see the log).

NoHandler: no handler is defined for the MIME type.

ExcludedMime : the MIME type is part of the excludedmimetypes list.

NotIncludedMime : the onlymimetypes list is not empty and the the MIME type is not in it.

Option -R allows specifying a temporary file for writing error messages in case indexing fails. The same messages go to stderr, and the option is mostly of use to the GUI, for displaying an error popup.

If option -m is given, recollindex is started for real time monitoring, using the file system monitoring package it was configured for (either fam, gamin, or inotify). This mode must have been explicitly configured when building the package, it is not available by default. The program will normally detach from the controlling terminal and become a daemon. If option -D is given, it will stay in the foreground. Option -w <seconds> can be used to specify that the program should sleep for the specified time before indexing begins. The default value is 60. The daemon normally monitors the X11 session and exits when it is reset. Option -x disables this X11 session monitoring (daemon will stay alive even if it cannot connect to the X11 server). You need to use this too if you use the daemon without an X11 context. You can use option -n to skip the initial incrementing pass which is normally performed before monitoring starts. Once monitoring is started, the daemon normally monitors the configuration and restarts from scratch if a change is made. You can disable this with option -C

recollindex -i will index individual files into the index. The stem expansion and aspell databases will not be updated. The skippedPaths and skippedNames configuration variables will be used, so that some files may be skipped. You can tell recollindex to ignore skippedPaths and skippedNames by setting the -f option. This allows fully custom file selection for a given subtree, for which you would add the top directory to skippedPaths, and use any custom tool to generate the file list (ie: a tool from a source code control system). When run this way, the indexer normally does not perform the deleted files purge pass, because it cannot be sure to have seen all the existing files. You can force a purge pass with -P.

recollindex -e will erase data for individual files from the index. The stem expansion databases will not be updated.

Options -i and -e can be combined. This will first perform the purge, then the indexing.

With options -i or -e , if no file names are given on the command line, they will be read from stdin, so that you could for example run:

find /path/to/dir -print | recollindex -e -i

to force the reindexing of a directory tree (which has to exist inside the file system area defined by topdirs in recoll.conf). You could mostly accomplish the same thing with

find /path/to/dir -print | recollindex -Z -i

The latter will perform a less thorough job of purging stale sub-documents though.

recollindex -r mostly works like -i , but the parameter is a single directory, which will be recursively updated. This mostly does nothing more than find topdir | recollindex -i but it may be more convenient to use when started from another program. This retries failed files by default, use option -K to change. One or multiple -p options can be used to set shell-type selection patterns (e.g.: *.pdf).

recollindex -l will list the names of available language stemmers.

recollindex -s will build the stem expansion database for a given language, which may or may not be part of the list in the configuration file. If the language is not part of the configuration, the stem expansion database will be deleted at the end of the next normal indexing run. You can get the list of stemmer names from the recollindex -l command. Note that this is mostly for experimental use, the normal way to add a stemming language is to set it in the configuration, either by editing "recoll.conf" or by using the GUI indexing configuration dialog.
At the time of this writing, the following languages are recognized (out of Xapian's stem.h):

recollindex -S will rebuild the phonetic/orthographic index. This feature uses the aspell package, which must be installed on the system.

recollindex -E will check the configuration file for topdirs and other relevant paths existence (to help catch typos).

recollindex --webcache-compact will recover the space wasted by erased page instances inside the Web cache. It may temporarily need to use twice the disk space used by the Web cache.

recollindex --webcache-burst <destdir> will extract all entries from the Web cache to files created inside <destdir>. Each cache entry is extracted as two files, for the data and metadata.

recollindex --notindexed [path [path ...]] will check each path and print out those which are absent from the index (with an "ABSENT" prefix), or caused an indexing error (with an "ERROR" prefix). If no paths are given on the command line, the command will read them, one per line, from stdin.

Interrupting the command: as indexing can sometimes take a long time, the command can be interrupted by sending an interrupt (Ctrl-C, SIGINT) or terminate (SIGTERM) signal. Some time may elapse before the process exits, because it needs to properly flush and close the index. This can also be done from the recoll GUI (menu entry: File/Stop_Indexing). After such an interruption, the index will be somewhat inconsistent because some operations which are normally performed at the end of the indexing pass will have been skipped (for example, the stemming and spelling databases will be inexistent or out of date). You just need to restart indexing at a later time to restore consistency. The indexing will restart at the interruption point (the full file tree will be traversed, but files that were indexed up to the interruption and for which the index is still up to date will not need to be reindexed).

See Also

recoll(1) recoll.conf(5)

Referenced By

recoll(1), recoll.conf(5), recollq(1).

8 January 2006