notes

Surrounded by cats and trying to get my invoices together. What I really want to listen to is Bola Sete, but as I'm scanning through my library of Brazilian tunes, I go to the Getz/Gilberto album and click the genius button.

A hundred bossa nova and tropicalia tunes present themselves. Rather, 99 brazilian tunes + High and Dry by Radiohead. That's odd.

I click on High and Dry and redo the genius playlist. Same thing. 1 Radiohead tune and 99 brazilian tunes.

wtf? But wait. I start up High and Dry, and sure enough it *is* a bossa nova tune. I'd just never noticed before. The melody, the chord structure, everything. How many other tunes has Thom Yorke written that are actually bossa tunes?

And, what does iTunes Genius know that I don't?

OK-- successfully upgraded this blog to run svn/trunk version of QuillsEnabled & RemoteBlogging. Current egg version is 1.7.0b3 Just tested and outbound rss seems to be working fine.

You may now subscribe to me. Gently, please.

    re.compile()
        useful when there will be multiple results or searching over iteration
        here's the pattern:
        -assign re.compile(r'REGEX') to an OBJECT
            patternobj = re.compile(r'REGEX')
        -use OBJECT.search(VAR)
            for ITERATOR in SOMEFILE.readlines():
                if patternobj.search(ITERATOR):
                    print ITERATOR
                    
    re.search()
        useful for one-off matched
        here's the pattern:
            if re.search(r'REGEX', THING TO LOOK IN): do something
    
    re.match()
        always matches from the start of the string
        here's the pattern:
            if re.match(r'REGEX', THING TO LOOK IN): do something
    
    re.compile() produces a PATTERNOBJ
    re.search() and re.match() produce a MATCHOBJ
    
    MATCHOBJ.group()
        gets the text matched by the group
        here's the pattern:
            string1 = "Find REGEXTHING in this sentence"
            matchobj = re.search(r'(REGEX)', string1) #Note parens
            id = matchobj.group(1) #Num refers to first paren group
        or:
            OBJECT = re.compile(r'REGEX')
            matchobj = OBJECT.search(THING TO LOOK IN)
            if (matchobj):
                print "the matched text was "" + matchobj.group(1) +"
    MATCHOBJ.groups()
        returns all parens groups in a tuple
        
    re.findall()
        finds all matches (returns a list?)
        here's the pattern:
            OBJECT = re.findall(r'REGEX', THING TO LOOK IN)
        as in
            text = "High: 33, low: 17"
            temp_tuples = re.findall(r'(\w+):\s+(\d+)', text)
            print temp_tuples #[('High', '33'), ('low', '17')]
    
    PATTERNOBJ.sub()
        does search and replace on text returned by re.compile
        here's the pattern:
            OBJECT = ("some", "tuple", "of", "strings")
            OBJ2 = re.compile(r'REGEX')
            for ITERATOR in OBJ2:
                ITERATOR = OBJ2.sub('THING TO SWAP IN', ITERATOR)
                print ITERATOR
        options:
            ITERATOR = OBJ2.sub('THING TO SWAP IN', ITERATOR, count=n) 
                #limits number of substitutions in a found term
            
    PATTERNOBJ.subn()
        like sub(), but returns a 2-element tuple containing the subbed text 
        and # of substitutions made
        (THING TO LOOK IN, NUM) = PATTERNOBJ.subn('THING TO SWAP IN', THING TO LOOK IN)
            

**
        
    Flags, classes, qualifiers, and metachars
        ^ = beginning of line
        $ = eol
        \A = beginning of line
        \Z = end of line
        (?i) = case insensitive; put before string
        re.I = case insensitive alt; put after THING TO LOOK IN
        re.MULTILINE = allows ^ and $ to match on start or end of lines
        re.DOTALL = splits a string allowing regex as the splitter
        \d = digit 0-9
        \w = word char: letters, nums or underscores; contains \d
        \s = any whitespace char
        \D = not in \d
        \W = not in \w
        \S = not in \s
        [abcdef] = custom wildcard class
        [^] = negates custom char class
        string.hexdigits = list of hex digits
        string.ascii_letters
        string.ascii_lowercase
        string.ascii_uppercase
        string.digits
        string.punctuation
        string.uppercase
        string.whitespace
        string.printable
        . = wildcard
        \ = toggle/escape the char's function
        * = 0 or more
        + = 1 or more
        ? = 0 or 1
        {3,10} = between 3 and 10 (inclusive)
        {3,} = 3 or more
        () = grouping
        () \1 = backreference
        \b = boundary char, finds word boundary (whitespace, punctuation)
        
**

    Working with files
    
    FILEOBJ.read()
        the idea is to open the while file into a string and then apply the regex to it
        here's the pattern:
            FILE = open('FILENAME')
            TEXT = FILE.read()
            if re.search(r'REGEX', TEXT, re.I):
                print "I found this: ", text

        

Backdrop is a freeware app for Mac that hides everything but the app you're working in.  It's different from Finder's 'Hide Others' command in that it also hides the desktop, and replaces it with a single color that you can tweak.  For scatterbrained compulsives like me, it's great.

Combine that with Megazoomer, specially tweaked for TextMate.  By default, Megazoomer expands the current window to fill the screen. I've got a tweaked version that only expands the window vertically, which makes more sense when you're writing something and want to keep the page normal.

Components^[+] overview

What's a component?

• Some actor that performs part of a "complex action"
• In Zope, some kind of object  that clearly defines what it provides and what it expects from other objects
• That two-way proviso is called a contract
• In Zope, contracts are expressed via interfaces
• Therefore, a component is an object that has an interface

What kind of components are there?

• Three categories usually: Model, View and Controller
• Model = content
• View = presentation
• Controller = "processes data in order to change the model or the view"

How Zope Component Architecture handles MVC:

• Adapters^[+] = implement controllers and views
• Utilities^[+] = "fulfilling a certain task without the context of another component
• what about the Model? = content components = see below

Interfaces^[+]

• Describe what a component is meant to do, but don't describe how it does it
• Not core Python, so you have to import from zope package zope.interface

Content Components

Content components = Python classes

Sounds like maybe a confusing naming convention, since apparently content components "do not have to implement a special interface." Does that mean they don't have to declare any interface?

Assume this is the role of adapters...? To handle interfaces for components that may not declare them themselves...?