More documentation can be found in the doc/ directory or at http://www.recoll.org Recoll user manual Jean-Francois Dockes Copyright (c) 2005 Jean-Francois Dockes This document introduces full text search notions and describes the installation and use of the Recoll application. [ Split HTML / Single HTML ] ---------------------------------------------------------------------- Table of Contents 1. Introduction 1.1. Giving it a try 1.2. Full text search 1.3. Recoll overview 2. Indexation 2.1. Introduction 2.2. The indexation configuration 2.3. Starting indexation 2.4. Using cron to automate indexation 3. Search 3.1. Simple search 3.2. Complex/advanced search 3.3. Document history 3.4. Result list sorting 3.5. Search tips, shortcuts 3.6. Customising the search interface 4. Installation 4.1. Building from source 4.1.1. Prerequisites 4.1.2. Building 4.1.3. Installation 4.2. Installing a prebuilt copy 4.2.1. Installing through a package system 4.2.2. Installing a prebuilt Recoll 4.3. Packages needed for external file types 4.4. Configuration overview 4.4.1. Main configuration file 4.4.2. The mimemap file 4.4.3. The mimeconf file ---------------------------------------------------------------------- Chapter 1. Introduction 1.1. Giving it a try If you do not like reading manuals (who does?) and would like to give Recoll a try, just perform installation and start the recoll user interface, which will index your home directory and let you search it right after. Do not do this if your home has a huge number of documents and you do not want to wait or are very short on disk space. In this case, you may want to edit the configuration file first to restrict the indexed area. Also be aware that you will need to install the appropriate supporting applications for document types that need them (for example antiword for ms-word files). ---------------------------------------------------------------------- 1.2. Full text search Recoll is a full text search application. Full text search applications let you find your data by content rather than by external attributes (like a file name). More specifically, they will let you specify words (terms) that should or should not appear in the text you are looking for, and return a list of matching documents, ordered so that the most relevant documents will appear first. You do not need to remember in what file or email message you stored a given piece of information. You just ask for related terms, and the tool will return a list of documents where those terms are prominent. This mode of operation has been made very familiar by internet search engines. The notion of relevance is a difficult one, as only you, the user, actually know which documents are relevant to your search, and the application can only try a guess. The quality of this guess is probably the most important element for a search application. In many cases, you are looking for all the forms of a word, not for a specific form or spelling. These different forms may include plurals, different tenses for a verb, or terms derived from the same root or stem (exemple: floor, floors, floored, floorings...). Recoll will by default expand queries to all such related terms (words that reduce to the same stem). This expansion can be disabled at search time. Stemming, by itself, does not provide for misspellings or phonetic searches. Recoll currently does not support these. ---------------------------------------------------------------------- 1.3. Recoll overview Recoll uses the Xapian information retrieval library as its storage and retrieval engine. Xapian is a very mature package using a sophisticated probabilistic ranking model. Recoll provides the interface to get data into (indexation) and out (searching) of the system. In practice, Xapian works by remembering where terms appear in your document files. The acquisition process is called indexation. The resulting database can be big (roughly the size of the original document set), but it is not a document archive. Recoll can only display documents that still exist at the place from which they were indexed. (Actually, there is a way to reconstruct a document from the information in the database, but the result is not nice, as all formatting, punctuation and capitalisation are lost). Recoll stores all internal data in Unicode UTF-8 format, and it can index files with different character sets, encodings, and languages into the same database. It has input filters for many document types. Stemming depends on the document language. Recoll stores the unstemmed versions of terms and uses auxiliary databases for term expansion. It can switch stemming languages, or add a language, without reindexing. Storing documents in different languages in the same database is possible, and useful in practice, but does introduce possibilities of confusion. Recoll currently makes no attempt at automatic language recognition. Recoll has many parameters which define exactly what to index, and how to classify and decode the source documents. These are kept in a configuration file. A default configuration is copied into a standard location (usually something like /usr/[local/]share/recoll/examples) during installation. The default parameters from this file may be overriden by values that you set inside your personal configuration, found by default in the .recoll subdirectory of your home directory. The default configuration will index your home directory with default parameters and should be sufficient for giving Recoll a try, but you may want to adjust it later. Indexation is started automatically the first time you execute the recoll search graphical user interface, or by executing the recollindex command. Searches are performed inside the recoll program, which has many options to help you find what you are looking for. ---------------------------------------------------------------------- Chapter 2. Indexation 2.1. Introduction Indexation is the process by which the set of documents is analyzed and the data entered into the database. Recoll indexation is normally incremental: documents will only be processed if they have been modified. On the first execution, of course, all documents will need processing. A full index build can be forced later on by specifying an option to the indexation command (recollindex -z). Recoll indexation takes place at discrete times. There is currently no interface to real time file modification monitors. The typical usage is to have a nightly indexation run programmed into your cron file. +------------------------------------------------------------------------+ | Side note: there is nothing in Recoll and Xapian that would prevent | | interfacing with a real time file modification monitor, but this would | | tend to consume significant system resources for dubious gain, because | | you rarely need a full text search to find documents you just | | modified. recollindex -i can be used to add individual files to the | | index if you want to play with this, see the manual page. | +------------------------------------------------------------------------+ Recoll knows about quite a few different document types. The parameters for document types recognition and processing are set in configuration files Most file types, like HTML or word processing files, only hold one document. Some file types, like mail folder files can hold many individually indexed documents. Recoll indexation processes plain text, HTML, openoffice and e-mail files internally. Other types (ie: postscript, pdf, ms-word, rtf) need external applications for preprocessing. The list is in the installation section. Without further configuration, Recoll will index all appropriate files from your home directory, with a reasonable set of defaults. ---------------------------------------------------------------------- 2.2. The indexation configuration Values set in the system-wide configuration file (named like /usr/[local/]share/recoll/examples/recoll.conf) can be overriden by those set in the personal one, named $HOME/.recoll/recoll.conf by default or $RECOLL_CONFDIR/recoll.conf if RECOLL_CONFDIR is set. The most accurate documentation for editing the file is given by comments inside the central one. If you want to adjust the configuration before indexation, just click Cancel when the program asks if it should start initial indexation. This will have created a .recoll directory containing empty configuration files. The configuration is also documented inside the installation chapter of this document, or in the recoll.conf(5) man page. ---------------------------------------------------------------------- 2.3. Starting indexation Indexation is performed either by the recollindex program, or by the indexation thread inside the recoll program (use the File menu). If the recoll program finds no database when it starts, it will automatically start indexation (except if cancelled). It is best to avoid interrupting the indexation process, as this may sometimes leave the database in a bad state. This is not a serious problem, as you then just need to clear everything and restart the indexation: the database files are normally stored in the $HOME/.recoll/xapiandb directory, which you can just delete if needed. Alternatively, you can start recollindex -z, which will reset the database before indexation. ---------------------------------------------------------------------- 2.4. Using cron to automate indexation The most common way to set up indexation is to have a cron task execute it every night. For example the following crontab entry would do it every day at 3:30AM (supposing recollindex is in your PATH): 30 3 * * * recollindex > /tmp/recolltrace 2>&1 The usual command to edit your crontab is crontab -e (which will usually start the vi editor to edit the file). You may have more sophisticated tools available on your system. ---------------------------------------------------------------------- Chapter 3. Search The recoll program provides the user interface for searching. It is based on the QT library. ---------------------------------------------------------------------- 3.1. Simple search 1. Start the recoll program. 2. Possibly choose a search mode: Any term or All terms or File name. 3. Enter search term(s) in the text field at the top of the window. 4. Click the Search button or hit the Enter key to start the search. The initial default search mode is Any term. This will look for documents with any of the search terms (the ones with more terms will get better scores). All terms will ensure that only documents with all the terms will be returned. File name will specifically look for file names, and allows using wildcards (*, ? , []). You can use the Tools / Advanced search dialog for more complex searches. After starting a search, a list of results will instantly be displayed in the main list window. Clicking on the Preview link for an entry will open an internal preview window for the document. Clicking the Edit link will attempt to start an external viewer (have a look at the mimeconf configuration file to see how these are configured). By default, the document list is presented in order of relevance (how well the system estimates that the document matches the query). You can specify a different ordering by using the Tools / Sort parameters dialog. The Preview and Edit edit links may not be present for all entries, meaning that Recoll has no configured way to preview a given file type (which was indexed by name only), or no configured external viewer for the file type. This can sometimes be adjusted simply by tweaking the mimemap and mimeconf configuration files. You can click on the Query details link at the top of the results page to see the query actually performed, after stem expansion and other processing. ---------------------------------------------------------------------- 3.2. Complex/advanced search The advanced search dialog has fields that will allow a more refined search, looking for documents with all given words, a given exact phrase, none of the given words, or a given file name (with wildcard expansion). All relevant fields will be combined by an implicit AND clause. It will let you search for documents of specific mime types (ie: only text/plain, or text/html or application/pdf etc...) It will let you restrict the search results to a subtree of the indexed area. Click on the Start Search button in the advanced search dialog to start the search. The button in the main window always performs a simple search. Click on the Show query details link at the top of the result page to see the query expansion. ---------------------------------------------------------------------- 3.3. Document history Documents that you actually view (with the internal preview or an external tool) are entered into the document history, which is remembered. You can display the history list by using the Tools/Doc History menu entry. ---------------------------------------------------------------------- 3.4. Result list sorting The documents in a result list are normally sorted in order of relevance. It is possible to specify different sort parameters by using the Sort parameters dialog (located in the Tools menu). The tool sorts a specified number of the most relevant documents in the result list, according to specified criteria. The currently available criteria are date and mime type. The sort parameters stay in effect until they are explicitely reset, or the program exits. An activated sort is indicated in the result list header. ---------------------------------------------------------------------- 3.5. Search tips, shortcuts Disabling stem expansion. Entering a capitalized word in any search field will prevent stem expansion (no search for gardening if you enter Garden instead of garden). This is the only case where character case should make a difference for a Recoll search. Phrases. A phrase can be looked for by enclosing it in double quotes. Example: "user manual" will look only for occurrences of user immediately followed by manual. You can use the This exact phrase field of the advanced search dialog to the same effect. Query explanation. You can get an exact description of what the query looked for, including stem expansion, and boolean operators used, by clicking on the result list header. File names. All file name elements (the broken up file path) are entered as terms during indexation, and you can specify them as ordinary terms in normal search fields. Alternatively, you can use specific file name search which will only look for file names and can use wildcard expansion. Quitting. Entering ^Q almost anywhere will close the application. Closing previews. Entering ^W in a preview tab will close it (and, for the last tab, close the preview window). ---------------------------------------------------------------------- 3.6. Customising the search interface It is possible to customise some aspects of the search interface by using Query configuration entry in the Preferences menu. There are two tabs in the dialog, dealing with the interface itself, and with the parameters used for searching and returning results. User interface parameters: * Number of results in a result page * Result list font: There is quite a lot of information shown in the result list, and you may want to customise the font and/or font size. The rest of the fonts used by Recoll are determined by your generic QT config (try the qtconfig command. * Html help browser: this will let you chose your the preferred browser which will be started from the Help menu to read the user manual. You can enter a simple name if the command is in your PATH, or browse for a full pathname. * Show document type icons in result list: icons in the result list can be turned off. They take quite a lot of space and convey relatively little useful information. Search parameters: * Stemming language: stemming obviously depends on the document's language. This listbox will let you chose among the stemming databases which were built during indexing (this is set in the main configuration file), or later added with recollindex -s (See the recollindex manual). Stemming languages which are dynamically added will be deleted at the next indexation pass unless they are also added in the configuration file. * Dynamically build abstracts: this decides if Recoll tries to build document abstracts when displaying the result list. Abstracts are constructed by taking context from the document information, around the search terms. This can slow down result list display significantly for big documents, and you may want to turn it off. * Replace abstracts from documents: this decides if we should synthetize and display an abstract in place of an explicit abstract found within the document itself. ---------------------------------------------------------------------- Chapter 4. Installation 4.1. Building from source 4.1.1. Prerequisites At the very least, you will need to download and install the xapian core package (Recoll currently uses version 0.9.2), and the qt runtime and development packages (Recoll development currently uses version 3.3.5, but any 3.3 version is probably ok). You will most probably be able to find a binary package for qt for your system. You may have to compile Xapian but this is not difficult (if you are using FreeBSD, there is a port). You may also need libiconv. Recoll currently uses version 1.9 (this should not be critical). On Linux systems, the iconv interface is part of libc and you should not need to do anything special. ---------------------------------------------------------------------- 4.1.2. Building Recoll has been built on Linux (redhat7.3, mandriva 2005, Fedora Core 3), FreeBSD and Solaris 8. If you build on another system, I would very much welcome patches. Depending on the qt configuration on your system, you may have to set the QTDIR and QMAKESPECS variables in your environment: * QTDIR should point to the directory above the one that holds the qt include files (ie: qt.h). * QMAKESPECS should be set to the name of one of the qt mkspecs subdirectories (ie: linux-g++). On many Linux systems, QTDIR is set by the login scripts, and QMAKESPECS is not needed because there is a default link in mkspecs/. The Recoll configure script does a better job of checking these variables after release 1.1.1. Before this, unexplained errors will occur during compilation if the environment is not set up. Also, for 1.1.0 the qmake command should be in your PATH (later releases can also find it in $QTDIR/bin). Normal procedure: cd recoll-xxx configure make (practises usual hardship-repelling invocations) There little autoconfiguration. The configure script will mainly link one of the system-specific files in the mk directory to mk/sysconf. If your system is not known yet, it will tell you as much, and you may want to manually copy and modify one of the existing files (the new file name should be the output of uname -s). ---------------------------------------------------------------------- 4.1.3. Installation Either type make install or execute recollinstall prefix, in the root of the source tree. This will copy the commands to prefix/bin and the sample configuration files, scripts and other shared data to prefix/share/recoll. You can then proceed to configuration. ---------------------------------------------------------------------- 4.2. Installing a prebuilt copy 4.2.1. Installing through a package system If you are lucky enough to be using a port system or a prebuilt package (RPM or other), just follow the usual procedure, and have a look at the configuration section. ---------------------------------------------------------------------- 4.2.2. Installing a prebuilt Recoll The unpackaged binary versions are just compressed tar files of a build tree, where only the useful parts were kept (executables and sample configuration). The executable binary files are built with a static link to libxapian and libiconv, to make installation easier (no dependencies). However, this also means that you cannot change the versions which are used. After extracting the tar file, you can proceed with installation as if you had built the package from source. ---------------------------------------------------------------------- 4.3. Packages needed for external file types Recoll uses external applications to index some file types. You need to install them for the file types that you wish to have indexed (these are run-time dependencies. None is needed for building Recoll): * PDF: pdftotext is part of the Xpdf package. * Postscript: pstotext. * MS Word: antiword. * RTF: unrtf * dvi: dvips * djvu: DjVuLibre * MP3: Recoll will use the id3info command from the id3lib package to extract tag information. Without it, only the filenames will be indexed. Text, Html, mail folders and Openoffice files are processed internally. ---------------------------------------------------------------------- 4.4. Configuration overview There are two sets of configuration files. The system-wide files are kept in a directory named like /usr/[local/]share/recoll/examples, they define default values for the system. A parallel set of files exists in the .recoll directory in your home (this can be changed with the RECOLL_CONFDIR environment variable. The database is also kept in .recoll by default, (this can be changed by a configuration parameter). If the .recoll directory does not exist when recoll or recollindex are started, it will be created with a set of empty configuration files. recoll will give you a chance to edit the configuration file before starting indexation. recollindex will proceed immediately. Most of the parameters specific to the recoll GUI are set through the Preferences menu and stored in the standard QT place ($HOME/.qt/recollrc). You probably do not want to edit this by hand. For other options, Recoll uses text configuration files. You will have to edit them by hand for now (there is still some hope for a GUI configuration tool in the future). The most accurate documentation for the configuration parameters is given by comments inside the default files, and we will just give a general overview here. All configuration files share the same format. For exemple, a short extract of the main configuration file might look as follows: # Space-separated list of directories to index. topdirs = ~/docs /usr/share/doc [~/somedirectory-with-utf8-txt-files] defaultcharset = utf-8 There are three kinds of lines: * Comment (starts with #) or empty. * Parameter affectation (name = value). * Section definition ([somedirname]). Section lines allow redefining some parameters for a directory subtree. Some of the parameters used for indexation are looked up hierarchically from the more to the less specific. Not all parameters can be meaningfully redefined, this is specified for each in the next section. The tilde character (~) is expanded in file names to the name of the user's home directory. White space is used for separation inside lists. Elements with embedded spaces can be quoted using double-quotes. ---------------------------------------------------------------------- 4.4.1. Main configuration file recoll.conf is the main configuration file. It defines things like what to index (top directories and things to ignore), and the default character set to use for document types which do not specify it internally. The default configuration will index your home directory. If this is not appropriate, use recoll to copy the sample configuration, click Cancel, and edit the configuration file before restarting the command. This will start the initial indexation, which may take some time. Paramers: topdirs Specifies the list of directories or files to index (recursively for directories). The indexer will not follow symbolic links inside the indexed trees. If an entry in the topdirs list is a symbolic link, indexation will not start and will generate an error. skippedNames A space-separated list of patterns for names of files or directories that should be completely ignored. The list defined in the default file is: *~ #* bin CVS Cache caughtspam tmp The list can be redefined for subdirectories, but is only actually changed for the top level ones in topdirs. The top-level directories are not affected by this list (that is, a directory in topdirs might match and would still be indexed). The list in the default configuration does not exclude hidden directories (names beginning with a dot), which means that it may index quite a few things that you do not want. On the other hand, mail user agents like thunderbird usually store messages in hidden directories, and you probably want this indexed. One possible solution is to have .* in skippedNames, and add things like ~/.thunderbird or ~/.evolution in topdirs. loglevel Verbosity level for recoll and recollindex. A value of 4 lists quite a lot of debug/information messages. 2 only lists errors. logfilename Where should the messages go. 'stderr' can be used as a special value. filtersdir A directory to search for the external filter scripts used to index some types of files. The value should not be changed, except if you want to modify one of the default scripts. The value can be redefined for any subdirectory. indexstemminglanguages A list of languages for which the stem expansion databases will be built. See recollindex(1) for possible values. You can add a stem expansion database for a different language by using recollindex -s, but it will be deleted during the next indexation. Only languages listed in the configuration file are permanent. iconsdir The name of the directory where recoll result list icons are stored. You can change this if you want different images. dbdir The name of the Xapian database directory. It will be created if needed when the database is initialized. defaultcharset The name of the character set used for files that do not contain a character set definition (ie: plain text files). This can be redefined for any subdirectory. If it is not set at all, the character set used is the one defined by the nls environment (LC_ALL, LC_CTYPE, LANG), or iso8859-1 if nothing is set. guesscharset Decide if we try to guess the character set of files if no internal value is available (ie: for plain text files). This does not work well in general, and should probably not be used. usesystemfilecommand Decide if we use the file -i system command as a final step for determining the mime type for a file (the main procedure uses suffix associations as defined in the mimemap file). This can be useful for files with suffixless names, but it will also cause the indexation of many bogus "text" files. indexallfilenames Recoll indexes file names in a special section of the database to allow specific file names searches using wild cards. This parameter decides if file name indexing is performed only for files with mime types that would qualify them for full text indexation, or for all files inside the selected subtrees, independant of mime type. ---------------------------------------------------------------------- 4.4.2. The mimemap file mimemap specifies the file name extension to mime type mappings. For file names without an extension, or with an unknown one, the system's file -i command will be executed to determine the mime type (this can be switched off inside the main configuration file). mimemap also has a list of extensions which should be ignored totally (to avoid losing time by executing file for things that certainly should not be indexed). The mappings can be specified on a per-subtree basis, which may be useful in some cases. Example: gaim logs have a .txt extension but should be handled specially, which is possible because they are usually all located in one place. mimemap also has a recoll_noindex variable which is a list of suffixes. Matching files will be skipped (avoids unnecessary decompressions or file executions). This is partially redundant with skippedNames in the main configuration file, with two differences: it will not affect directories, and it can be changed for any subdirectory. ---------------------------------------------------------------------- 4.4.3. The mimeconf file mimeconf specifies how the different mime types are handled for indexation, and for display. Changing the indexation parameters is probably not a good idea except if you are a Recoll developper. You may want to adjust the external viewers defined in (ie: html is either previewed internally or displayed using firefox, but you may prefer mozilla, your openoffice.org program might be named oofice instead of openoffice ...). Look for the [view] section. You can also change the icons which are displayed by recoll in the result lists (the values are the basenames of the png images inside the iconsdir directory (specified in recoll.conf). ----------------------------------------------------------------------