a simple and versatile collection of nodes with attributes, accessed as direct Python objects and automatically transcribed to assorted backing stores (CSV, SQL, GDBM, etc); the CSV backend can be (loosely) shared between multiple clients
A NodeDB is a base class for (currently small) databases of Node objects, native Python objects identified by their .type and .name attributes, and with uppercase attributes.
Node attribute access
As far as the backing store goes, each attribute is a sequence of values. Within Python, an attribute may be directly accessed as .FOO, which returns element 0 if the value sequence and requires the sequence to have exactly one element, or as .FOOs or .FOOes (note the lowercase plural suffix) which returns a view of the whole sequence.
The plural forms return a sequence view which itself accepts .FOO or .FOOs attributes. If the values are all Nodes, .FOOs returns a new view with all the .FOO values from each Node, so one may cascade access through a graph of Nodes, example:
which might return a sequence of email addresses from all the .LIST_MEMBER values from the root Node N.
The Node attributes obey both the sequence API and some of the set API: you can .append to or .extend one, or .add to or .update as with a set:
M = MemberNode("bill") N.LIST_MEMBERs.add(M)
A NodeDB can be backed by a CSV file (beta quality - I use it myself extensively) or SQL or a DBM file (alpha quality, both need some work). The CSV backend allows multiple clients to share the file; they update by appending to the file and monitor the updates of others.