Re: question for the Arista/Transmageddon groups

[Jeff Cobb <JeffC JBCobb net>]

Totally FWIW, I have found a nice Python-based script for doing MD5
hashes. I need to experiment some to make sure that what it generates
matches the output of the default applet but if so, that would remove
the one platform-specific bit that I need libdvdread for. What I am
trying to say guys is that if this works as a drop-in replacement, I
could have a nice Python module for doing basic lookups this weekend
which should make things easier if you are still interested in this. I
don't know if this list is the right place for detailed coding
discussions but of not I have some technical questions about the basic
requirements for this. I mean, I know what I use the metadata for but I
would like to do a little requirements gathering so I can be sure that
what I deliver to you is what you want/what would truly be useful for
your respective projects...let me know. I can add more tables to the
database, collect different metadata, etc. One way it could work is
this:
1. I prove out the hashing stuff.
2. I upgrade my project to use that, dispensing with the rather onerous
albeit useful "build from source" requirement I have. This would take
maybe 10m.
3. Providing I have your requirements on the metadata, I can extend the
DB as needed. Depending on how little or much the changes are, this
probably would take no more than an hour.
4. I can then get an exported object ready for testing/inclusion in your
projects.
5. I also want to start on a better admin tool; currently I can do all
admin work via command line applets I wrote for the purpose. The idea
behind the GUI/TUI on the admin tool is so that non-programmers can help
by adding data to the DB. I am deathmarching ATM but by the weekend I
want to have a basic walkthru of all of your code, with an eye to
integration strategies. Further, since your apps already sport a GUI, if
I am to do this I want to do it in the same way that you do.

That said, a few general questions:
1. Documentation. This can mean anything from just the output of Pydoc
(essentially just the class/methods are documented from source) to
detailed docs on how the object works and how to integrate it into
another project. Not sure what you are interested in.
2. This is a tricky one...or maybe not. I have almost always coded from
an OOP perspective and so right now, the records for a single DVD entry
might result in a DVD module which will be comprised of some DVD-wide
metadata (disc ID, title, etc) and 1-n track objects in an array.  Each
of these reflect the track number on the DVD, parent ID (for
cross-referencing), track title and type. IOW I tried to make each class
a module and exists in one Python file. So for instance a DVD will be
one DVD object and an array of valid tracks. With the class that houses
all of this there will be three or four Python scripts. I have had good
luck from an organization standpoint but as I learned from recent work
with a colleague, I guess some folks like all code in one big file. What
are your preferences?
3. Also, once all is done I may try it on several platforms including
Windows and various Linuxen. The idea here to is confirm proper
operation of the module as well as confirming the ability of this to
live on different platforms.

Let me know. I am off to do some testing.

Best,
Jeff