Changeset 388

Timestamp:

01/10/07 08:42:36

Author:

fumanchu

Message:

Doc updates, mostly.

Files:

• trunk/doc/index.html (modified) (2 diffs)
• trunk/doc/modeling.html (modified) (6 diffs)
• trunk/doc/storage.html (modified) (13 diffs)
• trunk/storage/__init__.py (modified) (2 diffs)
• trunk/storage/db.py (modified) (1 diff)
• trunk/storage/storefs.py (modified) (1 diff)
• trunk/units.py (modified) (1 diff)

trunk/doc/index.html

@@ -62 +62 @@
-    <li><a href='modeling.html#schemas'>Managing Schemas</a>
-        <ul>
-
-
+
+
+
+
-        </ul>
-    </li>
@@ -100 +102 @@
-<li><a href='storage.html'>Deployers: Configuring Storage</a>
-    <ul>
-
+
+
-        <ul>
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-        </ul>
-    </li>

trunk/doc/modeling.html

@@ -826 +826 @@
-<a name='schemas'><h3>Managing Schemas</h3></a>
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-Archaeology-Biography application at various libraries around the world.
-After a year, one of the developers wishes to implement a new reporting
@@ -836 +869 @@
-Biography class isn't very informative. It would be better if we could
-rename that to "ArchaeologistID":</p>
+
+<pre>
+arena.rename_property(Biography, "ArchID", "ArchaeologistID")
+</pre>
+
+<p>Assuming we've already made the change to our model, the above example
+renames the property in the persistence layer (the database) using the
+<tt class='def'>rename_property(cls, oldname, newname)</tt> method.
+Additional <tt>arena</tt> methods:</p>
+
+<p>Unit classes (tables):</p>
+<ul>
+<li><tt class='def'>create_storage(cls)</tt></li>
+<li><tt class='def'>has_storage(cls)</tt></li>
+<li><tt class='def'>drop_storage(cls)</tt></li>
+</ul>
+
+<p>Unit properties (columns):</p>
+<ul>
+<li><tt class='def'>add_property(cls, name)</tt></li>
+<li><tt class='def'>has_property(cls, name)</tt></li>
+<li><tt class='def'>drop_property(cls, name)</tt></li>
+</ul>
+
+<p>Unit property (column) indices:</p>
+<ul>
+<li><tt class='def'>add_index(cls, name)</tt></li>
+<li><tt class='def'>has_index(cls, name)</tt></li>
+<li><tt class='def'>drop_index(cls, name)</tt></li>
+</ul>
+
+
+<h4>Upgrading: Schema Objects</h4>
+
+<p>The <tt>Schema</tt> class helps you manage changes to your Dejavu model
+throughout its lifetime. Taking our <tt>rename_property</tt> example from
+above, we can rewrite it in a Schema obejcts like this:</p>
-
-<pre>class ArchBioSchema(dejavu.Schema):
@@ -849 +919 @@
-</pre>
-
-
-
-
-
-
-
-
-
-
+
-If you examine the base Schema class, you will see that it already has an
-<tt>upgrade_to_0</tt> method. The "zeroth" upgrade makes no schema changes;
@@ -863 +925 @@
-1 in the example, just in case I need some setup code in the future ;).</p>
-
- version argument,
-
-
-
-
-
-The Schema superclass
+
+
+
+
+for broken steps) if it comes to 
+
+Each Schema
-While an upgrade is in process, this value will be an int, the same number
-as that of the upgrade method. That is, while upgrade_to_2 is running,
@@ -876 +938 @@
-
-<p>After you run <tt>upgrade</tt>, you can call the
-
-
+
+
+
-If storage already exists for a given class, it is skipped.</p>
-
@@ -889 +952 @@
-and deployed schema).</p>
-
-
+
+
-
-<p>The <tt>Schema</tt> class uses a magic table in the database to keep

trunk/doc/storage.html

@@ -12 +12 @@
-
-<h2>Deployers: Configuring Storage</h2>
-
-<h3>Storage Managers</h3>
-
-<p>Storage Managers insulate an application developer from the specifics of
@@ -23 +21 @@
-which you can simply "plug and play". But if you <i>need</i> more control
-over your data storage, you have it, without becoming a programmer.</p>
+
+
+<a name='configuration'><h3>Configuration Files</h3></a>
-
-<p>When you deploy an app built with Dejavu, you must specify Storage
@@ -83 +84 @@
-
-
-h4>Database Storage Managers</h4
-
-5>Microsoft SQL Server / Microsoft Access (Jet)</h5
+a name='databases'><h3>Database Storage Managers</h3></a
+
+4>Microsoft SQL Server / Microsoft Access (Jet)</h4
-<p>This module was developed against ADO 2.7 and 2.8,
-    using MSDE, SQL Server 2000, and Access 2000.
@@ -97 +98 @@
-</ul>
-
-5>PostgreSQL (pyPgSQL)</h5
+4>PostgreSQL (pyPgSQL)</h4
-<p>This class was developed against
-    PostgreSQL 8.0.0 rc-1 on Win2k,
@@ -111 +112 @@
-</ul>
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
-<p>This class was developed against
-    mysql  Ver 14.7 Distrib 4.1.8, for Win95/Win98 (i32),
@@ -125 +138 @@
-</ul>
-
-5>SQLite (pysqlite/sqlite3)</h5
+4>SQLite (pysqlite/sqlite3)</h4
-<p>This class was developed against
-    sqlite 3.0.8 (pysqlite-1.1.6.win32-py2.3),
@@ -142 +155 @@
-
-
-5>Firebird (kinterbasdb)</h5
+4>Firebird (kinterbasdb)</h4
-<p>This class was developed against
-    KInterbasDB Version: (3, 2, 0, 'alpha', 1) and 
@@ -159 +172 @@
-Patches welcome.</b></p>
-
-5>Common Database Configuration Entries</h5
+4>Common Database Configuration Entries</h4
-<p>In addition to the above, Storage Managers for databases (probably)
-accept these additional options:</p>
@@ -216 +229 @@
-
-
-h4>Other Storage Managers</h4
-
-5>RAM</h5
+a name='other'><h3>Other Storage Managers</h3></a
+
+4>RAM</h4
-<p>Persists Units in RAM; all Units are lost when the process exits.</p>
-
-5>Shelve</h5
+4>Shelve</h4
-<p>Persists Units to shelve-type files. Extremely simple implementation;
-everything is pickled. Querying will be slow--every Unit is sucked in
@@ -241 +254 @@
-Configuration entries:</p>
-<ul>
-
+
+
-    <li><b>Path:</b> The file path (directory) in which to place db files.
-        Each Unit subclass will get its own file, of the same name as the
@@ -248 +262 @@
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-
-<p>Some Storage Managers act as "middleware", and can be chained together
@@ -260 +323 @@
-times, and address other integration concerns on their own systems.</p>
-
-5>Caching Proxy</h5
+4>Caching Proxy</h4
-<p>Use this class to persist Units in memory between client connections.
-It must proxy another Storage Manager. Configuration entries:</p>
@@ -286 +349 @@
-
-
-5>Burned Proxy</h5
+4>Burned Proxy</h4
-<p>Use this class to persist Units in memory between client connections.
-It needs another Storage Manager to proxy. Unlike the Caching Proxy above,

trunk/storage/__init__.py

@@ -88 +88 @@
-                                  % self.__class__)
-    
+    def has_property(self, cls, name):
+        raise NotImplementedError("%s has no has_property method."
+                                  % self.__class__)
+    
-    def drop_property(self, cls, name):
-        raise NotImplementedError("%s has no drop_property method."
@@ -95 +99 @@
-        raise NotImplementedError("%s has no rename_property method."
-                                  % self.__class__)
+    
+    def add_index(self, cls, name):
+        raise NotImplementedError("%s has no add_index method."
+                                  % self.__class__)
+    
+    def has_index(self, cls, name):
+        raise NotImplementedError("%s has no has_index method."
+                                  % self.__class__)
+    
+    def drop_index(self, cls, name):
+        raise NotImplementedError("%s has no drop_index method."
+                                  % self.__class__)
+    
-    
-    #                            Transactions                             #

trunk/storage/db.py

@@ -606 +606 @@
-        
-        t.rename(oldname, newname)
+    
+    def add_index(self, cls, name):
+        i = self.db.make_index(cls.__name__, name)
+        self.db[cls.__name__].indices[name] = i
-    
-    def has_index(self, cls, name):

trunk/storage/storefs.py

@@ -43 +43 @@
-        # Character to use for joining multiple identifiers
-        # into a single folder name.
-
+id
-        
-        # Map of file extensions. Keys should be "clsname.propname"

trunk/units.py

@@ -259 +259 @@
-    
-    def assign(self, unit, sequence):
-
+ 
-        if sequence:
-            m = max(sequence)