JSAM’s basic operations are get, put and remove, with basic support for transactions via begin, commit and rollback. All stored objects are identified by an integer key.
A JSAM database is opened by supplying a URL to jsam.jsurl.parse() and calling the open() method on the returned object. The syntax of URL depends on the particular “access method”.
The following URL types currently exist:
* file: directory plain text JSON files * rcs: directory plain text JSON files under RCS * cvs: directory plain text JSON files under CVS (not tested) * svn: directory plain text JSON files under SVN (not tested) * sqlite: object trees stored in an SQLite3 database * mysql: object trees stored in a MySQL database (lightly tested) * pgsql: object trees stored in a PostgreSQL database (lightly tested)
Query methods include:
* Collections: a jQuery/Django-like method-chained interface
* XPath query language subset (only supported for SQL backends)
* P4J (E4X-like) query language (only supported for SQL backends)
which allows construction of "xpath" queries in Python using
method chaining and operator overload, bridging the lexical divide
between code and query.
- pydoc jsam.<URLTYPE>_jsam
gives information on the URL syntax for <URLTYPE>
- pydoc jsam.jsam.JSAM
gives information on jsam object methods
- pydoc jsam.xpath
gives information on XPath query syntax (SQL backends only)
- pydoc jsam.p4j
gives information on P4J Python native query syntax
NOTE:
* uses Python 2.6+ supplied "json"; requires "simplejson" on Python 2.5 * file-based storage depends on POSIX file semantics * requires "MySQLdb" for MySQL * requires "psycopg2" for PostgreSQL
The name “JSAM” is a joke on IBM Mainframe O/S data “Access Methods” (BDAM, BSAM, VSAM, ISAM, QSAM, BPAM).
The obvious name ‘jsonstore’ was already taken by another package in the Python Package Index (PyPI).
The URL/URI syntax was cribbed from Django, which copied SQLAlchemy.
