Ruby, irb and spotlight - possibly some AppleScript

Argghhh, apologies. I kept automatically doing command S, which posts the above message so that it has multiple edits. For some reason I can now no longer edit the message. My last and what I thought was the final edit with tags set etc. was refused. Hopefully I can resist the automatic command S now.

I wanted to add the following:

The result of command 1 is a ruby array whose first element is the image width, and the second the image height.

The result of command 2 is a list of the image files in my Pictures folder which have a width of 2272 pixels and a height of 1704 pixels.

The result of command 3 is a list of every image file on my computer with the width greater than 2272 pixels and height greater than 1705 pixels.

The result of command 4 is a list of image files on my computer that have a content creation date more recently than 3 months ago, but older than 2 months ago.

The result of command 5 is a list of every image file whose content was created in the last 30 days.

Commands 2 to 5 can produce huge amounts of output. I was bit concerned the pipe would fill and block the command but everything appears to work as it should.

Thanks red menace. Here is my next attempt:

#!/usr/bin/env ruby

require 'Open3'

module SpotlightCommand

# returns a two item array, [ width, height ]

def self.get_imagewidthheight(imageFilePath)

finalResult = [ -1, -1 ] # width, height

resultStr, exitVal = Open3.capture2("mdls", "-name", "kMDItemPixelWidth",

"-name", "kMDItemPixelHeight", imageFilePath)

return finalResult unless exitVal.exitstatus.zero?

resultArray = resultStr.split("\n")

if (resultArray.length.eql? 2)

tempArray = resultArray[0].split(" ")

if (tempArray[0].eql? "kMDItemPixelHeight")

finalResult[1] = tempArray[2].to_i

elsif (tempArray[0].eql? "kMDItemPixelWidth")

finalResult[0] = tempArray[2].to_i

end

tempArray = resultArray[1].split(" ")

if (tempArray[0].eql? "kMDItemPixelWidth")

finalResult[0] = tempArray[2].to_i

elsif (tempArray[0].eql? "kMDItemPixelHeight")

finalResult[1] = tempArray[2].to_i

end

end

return finalResult

end

# essentially a private module method, though I've not found a easy solution def self.make_contenttypepartofquery(fileType)

contentTypeQueryPart = nil

typesHash = { :"public.jpeg" => "public.jpeg",

:"public.png" => "public.png",

:"public.tiff" => "public.tiff",

:"com.compuserve.gif" => "com.compuserve.gif" }

fileType = typesHash[fileType.intern] unless fileType.nil?

contentTypeQueryPart = if fileType.nil?

"kMDItemContentTypeTree == public.image"

else

"kMDItemContentType == " + fileType

end

return contentTypeQueryPart

end

# essentially a private module method, though I've not found a easy solution

def self.runquerycommand(theCommand)

theOutput = ""

IO.popen(theCommand) do |io|

theOutput = io.read

end

theOutput = theOutput.split("\n")

return theOutput

end

def self.find_imagefiles(width: 800, height: 600,

fileType: "public.image", onlyInDirPath: nil)

theCommand = [ "mdfind" ]

theCommand.push('-onlyin', onlyInDirPath) unless onlyInDirPath.nil?

query = self.make_contenttypepartofquery(fileType) + " && "

query += "kMDItemPixelWidth == " + width.to_s + " && "

query += "kMDItemPixelHeight == " + height.to_s

theCommand.push(query)

return self.runquerycommand(theCommand)

end

def self.find_imagefiles_largerthan(width: 800, height: 600,

fileType: nil, onlyInDirPath: nil)

theCommand = [ "mdfind" ]

theCommand.push('-onlyin', onlyInDirPath) unless onlyInDirPath.nil?

query = self.make_contenttypepartofquery(fileType) + " && "

query += "kMDItemPixelWidth >= " + width.to_s + " && "

query += "kMDItemPixelHeight >= " + height.to_s

theCommand.push(query)

return self.runquerycommand(theCommand)

end

def self.find_imagefilescreated(monthsAgo: 3, fileType: nil,

onlyInDirPath: nil)

monthsAgo =(-( monthsAgo.to_i )).to_s

monthsAgoPlus1 = (monthsAgo.to_i + 1).to_s

theCommand = [ "mdfind" ]

theCommand.push('-onlyin', onlyInDirPath) unless onlyInDirPath.nil?

query = self.make_contenttypepartofquery(fileType) + " && "

query += "kMDItemContentCreationDate > $time.this_month(" + monthsAgo+")" +

" && kMDItemContentCreationDate < $time.this_month("+monthsAgoPlus1+")"

theCommand.push(query)

return self.runquerycommand(theCommand)

end

def self.find_imagefilescreated_since(daysAgo: 20, fileType: nil,

onlyInDirPath: nil)

daysAgo = (-(daysAgo.to_i)).to_s

theCommand = [ "mdfind" ]

theCommand.push('-onlyin', onlyInDirPath) unless onlyInDirPath.nil?

query = self.make_contenttypepartofquery(fileType) + " && "

query += "kMDItemContentCreationDate >= $time.today(" + daysAgo + ")"

theCommand.push(query)

return self.runquerycommand(theCommand)

end

end

I've been looking at the encoding issue. First my configuation. Mavericks 10.9.3. Standard ruby and irb.

I put an image file in a folder called temp with the ß character in the name and a few characters with accents like ü, îéå etc.

If using irb I paste in the full path into:

result = SpotlightCommand.get_imagewidthheight("/Users/ktam/Pictures/temp/DSCßüéåî_0196. JPG")

It doesn't work. If I however use:

pathToFile = SpotlightCommand.find_imagefiles(width: 1280, height: 1024, fileType: "public.jpeg", onlyInDirPath: "/Users/ktam/Pictures/temp")[0]

and then do:

result = SpotlightCommand.get_imagewidthheight(pathToFile)

It works.

I can see when I paste into the irb session that the path isn't right. If I exit irb and paste into the same terminal window on the command line the pasted file path is correct. The string handling of the routines themselves seems to be find so I did some stackoverflowing and that suggests that there is a problem with irb and encodings on OS X with no one actually specifying solutions.

Starting irb with "irb -E UTF-8:UTF-8" doesn't help with typing in non roman paths, or even pasting in non roman paths.

If you are, that is "red menace" (excellent name by the way) getting the same behaviour as me then that is something that seems to a problem with irb and can be worked around. If you are seeing extra added broken behaviour let me know.

Thanks again

Kevin

I am running the script straight from BBEdit instead of irb, but adding UTF-8 encoding when runing the query command works for me. I'm not sure what the default encoding is, but it looks like maybe ASCII - I don't think I had much more than an ellipse or something in there somewhere.

When running scripts directly from the command line things works. I'm surprised BBEdit doesn't default to UTF-8.

Ruby statements usually return their results, so you normally don't need an explicit return statement at the end of a method, and if appropriate the last expression result can even be used

The return for me is basically making it clear what the method is going to return, not just a particular value, but a clue as to the type of object as well. I've never liked things that you get one result as a side effect of another action. (Make an assignment, get a return value), so it is a personal style thing

I also noticed a couple of other odd things such as converting a string to a symbol so that you could look up the symbol in a hash to get a string (??).

I put the code in their to try and sanitise the input for the IO.popen method, though in the end I learnt that using the array form of the popen method reduces the need to sanitise input as each item in the array after the first is past as arguments (argv) to the first item in the array without expansion. So the security risk is already mostly dealt with. In the end I left it in because I was using the hash like a set, where if the uti file type passed in wasn't in the set then the hash returned nil and in that case I could fall back to assigning to the default of "public.image"

I like that you've used a block (lambda) for the split string. I think this is probably the biggest thing I'm missing when writing ruby is by default looking for the block solution. They're not something that comes naturally to me yet.

I'm not sure why the "#{width}" is so popular. Why is it preferred to just concatenating things together? Is there a reason or is it just that cool kids use it? Certainly that form is common on code I've seen posted on stack overflow.

Your solution for specifying the encoding in the IO.popen method doesn't fix things for entering paths when using irb, there is definitely something about hinky about irb's input when it comes to non roman characters. Your fix stays though because I don't know where input strings might come from.

I noticed that snake case is what is expected for method names, so I found your mixing my over verbose snake case with camel case a bit confusing. Nevertheless a big thank you for your input, I've learnt a lot.