DBCacheContrib

Reusable code that treats TWiki forms as if they were table rows in a database.

WARNING: TWiki-4 only. If you want to use this extension with an earlier version of TWiki, please see an earlier version of this topic

Summary of Contents

This module supports structured queries over a database built on the fly from the forms in TWiki topics. It does not support any tags, as it is provided as a service for other plugins that want to treat a TWiki web as a simple database; for example, the TWiki:Plugins/FormQueryPlugin, which supports the display of query results.

The plugin encapsulates code that was formerly in the "engine room" of the FormQueryPlugin^?. It has been abstracted out in the belief that it will be useful to anyone who wants to do simple search operations from a plugin.

Features

• Perform complex queries on the TWiki database
• Cache TWiki topics for rapid queries

How the database gets built

You can think of the database as an array of all the topics in a web. Each array entry is a map (or hash in perl terms) that maps a set of field names to values.

Each topic in the web automatically gets a number of standard fields, generated by reading the metadata from the topic (see TWikiMetaData)

• name - name of the topic
• parent - name of parent topic
• _up - reference to the Map of the parent topic, if it exists
• attachments - array of Maps, each of which contains:
  • _up - reference to the Map for the topic
  • name - attachment name
  • attr - e.g hidden
  • comment - attachment comment
  • path - client path used to upload attachment
  • size - size in Kb
  • user - who uploaded the attachment
  • version - e.g. 1.3
• info - Map containing:
  • _up - reference to the Map for the topic
  • author - most recent author
  • date - date of last change
  • format - topic format version
  • version - topic version number
• moved - Map containing:
  • _up - reference to the Map for the topic
  • by - who moved it
  • date - when they moved it
  • from - where they moved it from
  • to - where they moved it to
• form - form type
• form name - e.g. if a "MyForm" is attached, this will be MyForm. This is a reference to a Map containing a key for each field in the form. Each key maps to the value in the form data for that key. The Map will also have an _up reference to the Map for the topic.
• text - raw text of the topic)

The sub-Maps created for info, form name, moved, and each row in attachments also have a reference back to the topic Map, called _up.

Other fields may be added by subclasses. Refer to the documentation for the plugin that is using the DBCache for more details.

The cache

To achieve best perfomance the plugin caches the database read from the TWiki topics. It creates this cache in a file in the web, _DBCache. If any topic changes in the web, this cache is automatically rebuilt. The cache file can be deleted at any point with no ill effects.

Extending or customising

Extension or customisation is welcome, as long as all extensions are described and code provided back to the author.

The module is shipped with a perl build file, which should be used for installation and testing. Testing is done using CPAN:Test::Unit, and is invoked using the 'test' build target. Writing tests is a useful way of feeding back bugs as well. I can't encourage you enough to maintain and extend the tests!

Detailed Documentation

Clients use the DBCache by defining a subclass of the TWiki::Contrib::DBCache class. The following POD documentation describes the methods of this class and the various other classes provided by the plugin..

class DBCacheContrib

General purpose cache that treats TWiki topics as hashes. Useful for rapid read and search of the database. Only works on one web.

Typical usage:

  use TWiki::Contrib::DBCacheContrib;

  $db = new TWiki::Contrib::DBCacheContrib( $web ); # always done
  $db->load(); # may be always done, or only on demand when a tag is parsed that needs it

  # the DB is a hash of topics keyed on their name
  foreach my $topic ($db->getKeys()) {
     my $attachments = $db->get($topic)->get("attachments");
     # attachments is an array
     foreach my $val ($attachments->getValues()) {
       my $aname = $attachments->get("name");
       my $acomment = $attachments->get("comment");
       my $adate = $attachments->get("date");
       ...
     }
  }

As topics are loaded, the readTopicLine method gives subclasses an opportunity to apply special processing to indivual lines, for example to extract special syntax such as %ACTION lines, or embedded tables in the text. See FormQueryPlugin^? for an example of this.

new($dataDir, $web)

• $dataDir location of cache file
• $web name of web to create the object for.

Construct a new DBCache object.

readTopicLine($topic, $meta, $line, $fh) --> text

• $topic - name of the topic being read
• $meta - reference to the hash object for this topic
• line - the line being read
• $fh - the file handle of the file
• return text to insert in place of line in the text field of the topic

Called when reading a topic that is being cached, this method is invoked on each line in the topic. It is designed to be overridden by subclasses; the default implementation does nothing. The sort of expected activities will be (for example) reading tables and adding them to the hash for the topic.

onReload($topics)

• $topics - perl array of topic names that have just been loaded (or reloaded)

Designed to be overridden by subclasses. Called when one or more topics had to be read from disc rather than from the cache. Passed a list of topic names that have been read.

load()

Load the web into the database. Returns a string containing 3 numbers that give the number of topics read from the cache, the number read from file, and the number of previously cached topics that have been removed.

write($archive)

• $archive - the TWiki::Contrib::DBCacheContrib::Archive being written to

Writes this object to the archive. Archives are used only if Storable is not available. This method must be overridden by subclasses is serialisation of their data fields is required.

read($archive)

• $archive - the TWiki::Contrib::DBCacheContrib::Archive being read from

Reads this object from the archive. Archives are used only if Storable is not available. This method must be overridden by subclasses is serialisation of their data fields is required.

class Search

Search operators work on the fields of a TWiki::Contrib::DBCacheContrib::Map. The fields are given by name, and the values by strings or numbers. Strings should always be surrounded by 'single-quotes'. Strings which are regular expressions (RHS of , ~ operators) use 'perl' re syntax (see =man perlre for help). Numbers can be signed integers or decimals.

Warning single and double quotes are not allowed in values!

The following operators are available:

Operator	$Result*	Meaning
=	Boolean	LHS exactly matches the regular expression on the RHS. The expression must match the whole string.
!=	Boolean	Inverse of =
=~	Boolean	LHS contains RHS i.e. the RHS is found somewhere in the field value.
<	Boolean	Numeric <
>	Boolean	Numeric >
>=	Boolean	Numeric >=
<=	Boolean	Numeric <=
lc	String	Unary lower case
uc	String	unary UPPER CASE
EARLIER_THAN	BOOLEAN	Date is earlier than the given date
LATER_THAN	Boolean	LHS is later than the given date (string containing a date e.g. '1 Apr 2003')
WITHIN_DAYS	Boolean	Date (which must be in the future) is within n working days of todays date
!	Boolean	Unary NOT
AND	Boolean	AND
OR	Boolean	OR
()	any	Bracketed subexpression

Dates for EARLIER_THAN, LATER_THAN and WITHIN_DAYS must be dates in the format expected by Time::ParseDate (like the ActionTrackerPlugin^?). WITHIN_DAYS works out the number of working days (i.e. excluding Saturday and Sunday). Apologies in advance if your weekend is offset ± a day!

A search object implements the "matches" method as its general contract with the rest of the world.

Example

Get a list of attachments that have a date earlier than 1st January 2000

  $db = new TWiki::Contrib::DBCacheContrib::DBCache( $web ); # always done
  $db->load();
  my $search = new TWiki::Contrib::DBCacheContrib::Search("date EARLIER_THAN '1st January 2000'");

  foreach my $topic ($db->getKeys()) {
     my $attachments = $topic->get("attachments");
     foreach my $val ($attachments->getValues()) {
       if ($search->matches($val)) {
          print $val->get("name") . "\n";
       }
     }
  }

new($string)

• $string - string containing an expression to parse

Construct a new search node by parsing the passed expression.

matches($object) -> boolean

• $object - object to test; must implement get

See if object matches the search. $object can actually be any object that provides the method "get" that returns a value given a string key.

toString() -> string

Generates a string representation of the object.

class FileTime^?

Object that handles a file/time tuple for use in Storable and TWiki::Contrib::DBCacheContrib::Archive.

new($file)

• $file - filename

Construct from a file name

uptodate() -< boolean

Check the file time against what is seen on disc. Return 1 if consistent, 0 if inconsistent.

toString() -> string

Generates a string representation of the object.

write()

TWiki::Contrib::DBCacheContrib::Archive hook

read()

TWiki::Contrib::DBCacheContrib::Archive hook

class Array

Generic array object. This is required because perl arrays are not objects, and cannot be subclassed e.g. for serialisation. To avoid lots of horrid code to handle special cases of the different perl data structures, we use this array object instead.

new()

Create a new, empty array object

add($object)

• $object any perl data type

Add an element to the end of the array

find($object) -> integer

• $object datum of the same type as the content of the array

Uses "==" to find the given element in the array and return it's index

remove($index)

• $index - integer index

Remove an entry at an index from the array.

get($key, $root) -> datum

• $k - key
• $root - what # refers to

Subfield syntax

• get("9", $r) where $n is a number will get the 9th entry in the array
• get("[9]", $r) will also get the 9th entry
• get(".9", $r) will also get the 9th entry
• get(".X", $r) will return the sum of the subfield X of each entry
• get("[?search]", $r) will perform the given search over the entries in the array. Always returns an array result, even when there is only one result. For example: [?name='Sam'] will return an array of all the entries that have their subfield name set to Sam.
• # means "reset to root". So get("#[3]", $r) will return the 4th entry of $r (assuming $r is an array!).
• get("[*X]", $r) will get a new array made from subfield X of each entry in this array.

Where the result of a subfield expansion is another object (a Map or an Array) then further subfield expansions can be used. For example,

get("parent.UserTable[?SubTopic='ThisTopic'].UserName", $web);

See also TWiki::Contrib::DBCacheContrib::Map for syntax that applies to maps.

size() -> integer

Get the size of the array

sum($field) -> number

• $field - name of a field in the class of objects stored by this array

Returns the sum of values of the given field in the objects stored in this array.

search($search) -> search result

• $search - TWiki::Contrib::DBCacheContrib::Search object to use in the search

Search the array for matches with the given object. values. Return a TWiki::Contrib::DBCacheContrib::Array of matching entries.

getValues() -> perl array

Get a "perl" array of the values in the array, suitable for use with foreach

toString($limit, $level, $strung) -> string

• $limit - recursion limit for expansion of elements
• $level - currentl recursion level

Generates an HTML string representation of the object.

write($archive)

• $archive - the TWiki::Contrib::DBCacheContrib::Archive being written to

Writes this object to the archive. Archives are used only if Storable is not available. This method must be overridden by subclasses is serialisation of their data fields is required.

read($archive)

• $archive - the TWiki::Contrib::DBCacheContrib::Archive being read from

Reads this object from the archive. Archives are used only if Storable is not available. This method must be overridden by subclasses is serialisation of their data fields is required.

class Map

Generic map object for mapping names to things. A name is defined as name = \w+ | \w+ "." name The . indicates a field reference in a sub-map. Objects in the map are either strings, or other objects that must support toString.

new($string)

• $string - optional attribute string in standard TWiki syntax

Create a new, empty array object. Optionally parse a standard attribute string containing name=value pairs. The value may be a word or a quoted string (no escapes!)

fastget($k) -> datum

• $k - key

Get the value for a key, but without any subfield field expansion

get($k, $root) -> datum

• $k - key
• $root what # refers to

Get the value corresponding to key $k; return undef if not set.

Subfield syntax

• get("X",$r) will get the subfield named X.
• get("X.Y",$r) will get the subfield Y of the subfield named X.
• get("[X]",$r) = will get the subfield named =X (so X[Y] and X.Y are synonymous)..
• # means "reset to root". So get("#.Y", $r) will return the subfield =Y of $r (assuming $r is a map!), as will get("#[Y]".

Where the result of a subfield expansion is another object (a Map or an Array) then further subfield expansions can be used. For example,

get("UserTable[0].Surname", $web);

See also TWiki::Contrib::DBCacheContrib::Array for syntax that applies to arrays.

set($k, $v)

• $k - key
• $v - value

Set the given key, value pair in the map.

size() -> integer

Get the size of the map

remove($index) -> old value

• $index - integer index

Remove an entry at an index from the array. Return the old value.

getKeys() -> perl array

Get a "perl" array of the keys in the map, suitable for use with foreach

getValues() -> perl array

Get a "perl" array of the values in the Map, suitable for use with foreach

search($search) -> search result

• $search - TWiki::Contrib::DBCacheContrib::Search object to use in the search

Search the map for keys that match with the given object. values. Return a TWiki::Contrib::DBCacheContrib::Array of matching keys.

toString($limit, $level, $strung) -> string

• $limit - recursion limit for expansion of elements
• $level - currentl recursion level

Generates an HTML string representation of the object.

write($archive)

• $archive - the TWiki::Contrib::DBCacheContrib::Archive being written to

Writes this object to the archive. Archives are used only if Storable is not available. This method must be overridden by subclasses is serialisation of their data fields is required.

read($archive)

• $archive - the TWiki::Contrib::DBCacheContrib::Archive being read from

Reads this object from the archive. Archives are used only if Storable is not available. This method must be overridden by subclasses is serialisation of their data fields is required.

class Archive

Simple file archive storer and restorer. Handles serialising objects using their "write" and "read" methods. Serialisable objects must have a no-parameters constructor.

This module is only used if Storable isn't available. Storable is much faster, because it is implemented in C.

new($file, $rw)

• $file - archive file path
• $rw - mode "r" or "w"

Create a new archive, using filename $file and mode $rw which must be "r" or "w". The archive will remain in existence (and the file remain open) until "close" is called. An exclusive lock is taken for write as long as the file is open. Throws an exception if the archive cannot be opened.

close()

Close this archive. MUST be called to close the file.

writeByte($b)

• $b - byte to write

Write a byte to the archive

writeString($s)

• $s - string to write

Write a string to the archive

writeInt($i)

• $i integer to write

Write a 32-bit integer to the archive

writeObject()

Write an object to the archive. An object must implement read($archive) and write($archive), or may be undef or a string. No other types are supported.

readByte() -> byte

Read a byte from the archive

readString() -> string

Read a UTF8 string from the archive

readInt() -> integer

Read a 32-bit integer from the archive

readObject() -> object

Read an object from the archive

Installation Instructions

You do not need to install anything in the browser to use this extension. The following instructions are for the administrator who installs the extension on the server where TWiki is running.

Like many other TWiki extensions, this module is shipped with a fully automatic installer script written using the BuildContrib.

• If you have TWiki 4.1 or later, and Perl 5.8, you can install from the configure interface (Go to Plugins->Find More Extensions)
• If you have a permanent connection to the internet (and Perl 5.8), you are recommended to use the automatic installer script
  • Just download the DBCacheContrib_installer perl script and run it.
  • If the $TWIKI_PACKAGES environment variable is set to point to a directory, the installer will try to get archives from there. Otherwise it will try to download from twiki.org or cpan.org, as appropriate.
  • (Developers only: the script will look for twikiplugins/DBCacheContrib/DBCacheContrib.tgz before downloading from TWiki.org)
• If you don't have a permanent connection, you can still use the automatic installer, by downloading all required TWiki archives to a local directory.
  • Point the environment variable $TWIKI_PACKAGES to this directory, and the installer script will look there first for required TWiki packages.
    • $TWIKI_PACKAGES is actually a path; you can list several directories separated by :
  • If you are behind a firewall that blocks access to CPAN, you can build a local CPAN mini-mirror, as described at http://twiki.org/cgi-bin/view/Codev/BuildingDakar#CPAN_local_minimirror
• The installer script will:
  • Automatically resolve dependencies,
  • Copy files into the right places in your local install even if you have renamed data directories,
  • check in new versions of any installed files that have existing RCS histories files in your existing install.
• If you don't want to use the script, or have problems on your platform (e.g. you don't have Perl 5.8), then you can still install manually:
  1. Download and unpack one of the .zip or .tgz archives to a temporary directory.
  2. Manually copy the contents across to the relevant places in your TWiki installation.
  3. Check in any installed files that have existing ,v files in your existing install (take care not to lock the files when you check in)
  4. Manually edit LocalSite.cfg to set any configuration variables.
  5. Run configure and enable the module, if it is a plugin.
  6. Repeat from step 1 for any missing dependencies.

Contrib Info

Author:	TWiki:Main/CrawfordCurrie
Copyright ©:	This code is based on an original development of Motorola Inc. and is protected by the following copyrights: Copyright @copy 2002-2003 Motorola Inc. All Rights Reserved. Portions copyright © 2004. Crawford Currie http://www.c-dot.co.uk
License:	As required for the publication of all extensions to TWiki, this software is published under the terms of the GNU General Public License. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details, published at http://www.gnu.org/copyleft/gpl.html
Version:	11537
Change History:
11537	Added lc and uc operators for case-insensitive searches
9303	don't die on broken symlinks
8682	Item1580 one-char fix that makes the difference
8110	Item663 formatting and text fixes
7552	Item997 test update
7274	Item719 onReload() is not a static method.
7262	Item719 MichaelDaum?'s patch (almost) to correct parameters to onReload
7260	Item727 made it clean the form name using normaliseWebTopicName
6353	Item380 do as the man says; make all $/ local
5720	Updated tests
5719	Fix for correct handling of parent relations
5229	Small improvement to the way it handles errors from Storable and Archive
5223	Documentation fixes, adding gifs.
5048	Cairo readiness
5036	Split from SharedCode?
5031	Moving to new name
5030	About to rename
5019	Improved topic data model, cleaned up tests
5008	Added extended access syntax, [?], [*] etc.
5006	Doc fixes
5005	Poddified documentation
5003	Initial version
8 Jul 2004	Initial version, split out from FormQueryPlugin?
Dependencies:	Name Version Description Time::ParseDate >=2003.0211 Required. Available from CPAN. Storable >=2.07 Recommended; accelerates cache handling. Available from CPAN
Perl Version:	5.0
Plugin Home:	http://TWiki.org/cgi-bin/view/Plugins/DBCacheContrib
Feedback:	http://TWiki.org/cgi-bin/view/Plugins/DBCacheContribDev
Appraisal:	http://TWiki.org/cgi-bin/view/Plugins/DBCacheContribAppraisal