HOW TO START A NEW DATABASE

ACeDB stores its data in a fast-access binary form.  Data is
ordinarily loaded and dumped from a human-readable flat-file format
known as .ace.  To start a new database, you must:

	1) create a database directory containing the following
	subdirectories:

		wspec/		schema and other files
		database/	binary files
		wgf/		DNA analysis files [optional]
		whelp/		help files [optional]
		wscript/	helper scripts [optional]

	It's perfectly all right to ignore the optional directories.
	They are only relevant for the original C. elegans database.

	2) start acedb and allow it to initialize the binary
	files.
	
	3) load one or more .ace files.

Before you begin, you must have a models.wrm file and one or more .ace
files containing the flat-file representation of the data.  For an
example, see the moviedb example database, which is located at
usr/local/apache/htdocs/AcePerl/archive/moviedb.tar.Z.

The recipe for creating a new database from scratch using the
interactive xace tool is this:

	1) create a directory with the database's name
	2) within that directory create a directory named "wspec" (where 
		the schema lives) and another named "database"
	3) populate the wspec subdirectory with the schema files,
		which you can copy from another database, such as
		the C. elegans database
	4) run xace, giving it the database's directory as its
		command-line argument
	5) xace will prompt you to reinitialize the database, say "OK"
	6) using the edit menu, select "read .ace" file.  Say "yes"
	        when prompted for write access
	7) choose "Open ace file" from the dialog box, and locate
		the file you wish to load
	8) select "Read all"
	9) when done, close the window and select "Save..." from the
		main xace window

Read other .ace files in the same way.

Rather than launching xace, you can do it all with tace.  Lines
surrounded by <angle brackets> represent user input:

	% <tace /usr/local/acedb/my_db>
        // Database directory: /usr/local/acedb/my_db
	The file /usr/local/acedb/my_db/database/ACEDB.wrm does not exist, 
	indicating that the database is empty.
	  Should I re-initialise the system? (y or n) <y>

	// 13 objects read with 0 errors
	// 0 objects read with 0 errors
	// Indexing (this may take several minutes)
	// Indexed 742 tags in 452 objects totalling 383 nodes 
	// 13 objects read with 0 errors
	// 0 objects read with 0 errors
	// Indexing 
	// Reindexed 742 tags in 0 classes containing 0 objects totalling 0 nodes 

	**** Program tace,  compiled on: Jul  6 1999 10:58:24 ****
	**** Using  ACEDB Version 4_7i,  compiled on: Jul  6 1999 10:58:14 ****

	Code by: Jean Thierry-Mieg (CNRS, France) mieg@crbm.cnrs-mop.fr
         Richard Durbin (Sanger Centre, UK) rd@sanger.ac.uk
         Simon Kelley (Sanger Centre, UK) srk@sanger.ac.uk

	You may redistribute this program and database subject to the
	conditions in the accompanying copyright file.  Anyone interested in
	maintaining an up-to-date version should contact one of the authors
	at the above email addresses.

	// Type ? for a list of options

	acedb> <parse /usr/local/acedb/my_db/raw/my_content.ace>
	// Parsing file  /usr/local/acedb/raw/my_content.ace
	// 123 objects read with 0 errors
	// 123 Active Objects
	acedb> <save>
	// 123 Active Objects
	acedb> <quit>
	
	// A bientot

To set up an ACeDB server, please see ACEDB.HOWTO and
ACE_SERVER_TRAPS.HOWTO.

Lincoln Stein
lstein@cshl.org
May 12, 2003