repository: the system

Previous: xmlobj: XML-based record object ] [ Top:  ] [ Next: list: a data source ]

The repository object describes the system we're working with. A repository consists in large part of a series of lists, or data sources, each of which contains a series of entries, often referred to as "objects" despite the confusion in an object-oriented context.

Lists are named with space-free, punctuation-free strings, and list operations are conducted against the repository itself. For convenience, a list object is also defined, but it is not required.
class repository:
   """Implements a wftk repository.  Define a repository using either a wftk.xml object or a
      string indicating a filename to read XML from."""
Initialization for the repository object calls either repos_open or repos_open_file, depending on what it's given. The open_file alternative is the default; this method of opening also makes sure the current working directory is set to the repository's directory; lots of repository actions use various filesystem things, and so the repmgr's default mode is filesystem-based. However, you can also build your own repository description and pass it in as an XMLAPI structure, in which case you're on your own as far as the current working directory is concerned.
   def __init__ (self, defn='system.defn'):
      """Create and optionally open repository object."""
      self.repos = None
      if defn:
         if type(defn) == type(''): # string = filename
            self.repos = repmgr.open_file (defn)
         elif defn.__class__ == xml:
            self.repos = defn._xml
A great deal of what the repmgr does is to organize datasources into named lists. A list may be stored as lines in a text file, as files in a directory, as records in a table in an RDBMS, or anything else you want to write or request an adaptor for. Most of the list manipulation routines are still vaporware, even in the repmgr C API library, but you can pretty much see what will be possible at some point.

The repmgr defines a number of pseudolists whose names start with underscores (in general, names starting with underscores are reserved for use by the repmgr; this isn't enforced, but you should probably avoid calling things names starting with underscores to be on the safe side.) The most interesting is _lists, which is the list of lists defined in the repository.

There are two ways of querying a list; the first is the "list" method, which returns a list object, which can be used to iterate around in the results. It corresponds roughly to a Recordset object in Microsoft's DAO schema. The second way of querying a list is the "keys" method, which simply returns a Python list of keys of entries. This list can be used to retrieve entries as required.
   # --------------------------------------------------------------
   # List manipulation.
   # --------------------------------------------------------------
   def list (self, list='_lists', query=None, order=None):
      if self.repos==None: return []
      return repmgr.list (self.repos, list)
   def keys (self, list='_lists', query=None, order=None):
      if self.repos==None: return []
      return repmgr.list (self.repos, list)
   def defn (self, list='_lists'):
      if self.repos==None: return None
      x = repmgr.defn (self.repos, list)
      if not x: return None
      ret = xml()
      ret._xml = x
      return ret

   #def create
   #def drop
   #def define
For entry manipulation, there are two routes to take: you can talk to the repository object directly (which you may want to do, for instance, when loading an object from a file) or you can create an entry object and talk to it. The "get" method creates an entry object for us when called; if there is no entry with the key supplied, it will return None.
   # --------------------------------------------------------------
   # Entry manipulation.
   # --------------------------------------------------------------
   def get (self, list, key):
      ret = entry(self, list, key)
      ret.load ()
      if ret.is_element(): return ret
      return None
   def add (self, list, xml):
      ret = entry(self, list)
      ret.parse (xml)
      return ret
   #def mod
   def delete (self, list, key):
      return repmgr.delete (self.repos, list, key)

   #def merge
   #def getkey
   #def form
   #def edit
   #def display
   #def getvalue
   #def setvalue

   # --------------------------------------------------------------
   # Publishing
   # --------------------------------------------------------------
   #def publish_all
   #def publish_list
   #def publish_obj
   #def publish_pages
   #def publish_page
   #def get_layout

   # --------------------------------------------------------------
   # Change management.
   # --------------------------------------------------------------
   #def changes
   #def snapshot
   #def push
   #def push_all
   #def pull
   #def pull_all
   #def synch
   #def attach (self,:
   #def retrieve

   # --------------------------------------------------------------
   # Etc.
   # --------------------------------------------------------------
   def user_auth (self, user, password):
      if self.repos==None: return
      return repmgr.user_auth (self.repos, user, password)
Previous: xmlobj: XML-based record object ] [ Top:  ] [ Next: list: a data source ]

This code and documentation are released under the terms of the GNU license. They are additionally copyright (c) 2001, Vivtek. All rights reserved except those explicitly granted under the terms of the GNU license.