pycase Design History

Table of Contents

• 2001-09-16
• 2001-10-08
• 2004-10-24
• 2004-10-25

2001-09-16

Need to generate python models of postgresql database, and thus simplify complex chains of sql statements. Detoured to make tabbedwriter.py. Using resulting pg2py at work on a couple of projects -- seems to save quite a bit of typing.

2001-10-08

Want dia-to-XMI, and then XMI to python code. For that, need to generate python model of XMI/UML. First cut was to hand-craft the xmi.py and uml.py, form the specs. But the better way (given XMI/UML will change) is to autogenerate from the xmi_uml.dtd.

Started with learning to parse DTD's, via parsedtd.py. Now ready to build task-specific xmidtd2py.

...Built it. Also built dia2xmi.py, which reads a dia uml diagram and generates the XMI object tree. Have skeleton in place (enough to drive it). Now struggling with the "to_xml" process, to dump the XMI model into an XML file.

...to_xml is working. Now return to parsing the dia file. Was using xml_objectify, but let's shift to good old DOM.

2004-10-24

Pulled tabbedwriter out as separate package.

Restructured the pcase dirtree, isolating the UML/XMI, postgres, and dia material. I'll come back to those later.

Built "genmodel", "genpython", "gensql" modules and their regression tests.

2004-10-25