Extension counter program [___.ext]

Blitz3D Forums/Blitz3D Programming/Extension counter program [___.ext]

virtlands (Posted 2013) [#1]

Here's some weird code, it's a program to do a report on counts of each extension type within a directory (path). What's unusual about this code is not what it does, but how it does it. It's just an exercise in string logic, and stuff... This version of the program is made flexible to include any extension lengths, (most extensions seem to be of length 3). Examples of exts: BB, bmx, jpeg, png, exe, dat, torrent, decls ... How it works: As the program runs and encounters new extensions, these extensions are appended to a superset string EXT$ which grows longer with each new addition. The clever part is that there is an int array ExtCount() that parallels it. The indexes in the array reference appropriate indexes within EXT$. The index locations in ExtCount() store extension counts, (and extension lengths). Which means that extension types don't need to be known ahead of time; They are appended to EXT$ only if they are not already in it. So I thought of various complications that could happen with this method: What if EXT$ is currently = "JPEG" and a newcomer extension is "JPE", .. well "JPE" is a subset of "JPEG". We don't want the two to be confused, so they shouldn't share the same index, or should they?? The current solution is put them in like "JPEJPEG..." or "JPEGJPE...", so that separate and accurate counts are kept of each. ( There can be intermediate data of course, like "JPExxxJPEGxyxy" .) ;;------------ Another idea which may be an improvement that I haven't tried yet is to create an array of EXT strings like DIM EXT$[7] which allows for the possibility of extension lengths 1 to 7 The reasoning there is that ext lengths are assumed per each EXT array index and I don't need to store lengths. Then an additional possibility pops up when transferring this logic to other things too (that have nothing to do with file extensions): Create a sort of De Bruin sequence or sparse version of that where the substrings to be found within each EXT[] are of equal lengths each and can overlap to create a shortest superstring possible. For example, in the superstring, "ABCDEFxxG", if each substring has a length of 3, then you just stored 7 strings in 9 bytes. The substrings are "ABC", "BCD", "CDE", "DEF", "EFx", "Fxx", "xxG" Another fun fact: If you were to consider extensions of length 3, and you wanted to include all English letters A-Z (26 letters), then the shortest superstring that would include all combinations (taken 3 at a time: ... ABC,ABD,ABE,.....) would be a string of length 17601 bytes (= 26^3 + 25 ). ;================================================================= Here's the B3D demo code: ;; File "extension" counter program ;; ;; This program demonstrates a clever and weird way to report the number of each ;; type of file extension in a directory without even knowing ahead of time what the extensions are. ;; The code is flexible and assumes that the extensions can be of any reasonable length, (from 1 to 10?) ;; ;; Created by Virtlands, Feb 21,2013 ;; ;; We shall estimate that there are not more than 1000 files in a directory... ; Can re-dim later. Global extCmax = 1000 ;; In this INT array, (data shall be stored as SHORT variables) ;; each lower 2 bytes shall record the ext count, (If the count is more than 65535, we're in trouble.) ;; each higher 2 bytes shall record the extension length, (extension Examples: TXT, jpeg, JPG, ICO, BB, torrent,...) ; statistically, extension lengths are usually 2 or 3,... Dim EXTcount( extCmax ) ;; This shall grow into a long superset string that contains all named extensions. *(in lowercase) Global EXT$ ;; Each index of the EXT$ string shall correspond to an index of the EXTcount() array ;; That's the magic. Global sampleDir$ Global CurSave$ Global Appdir$ = SystemProperty$("AppDir") sampleDir$ = "C:\Blitz3d Projects\" ;; <== Enter your own path here if you don't want to use "APPDIR" ;; Let the program begin: If Len(sampleDir) Then sampledir = lastslash(sampledir) If FileType(SampleDir)<>2 ;; (This is in case the user does not input a sample.) SampleDir = Appdir End If curSave = CurrentDir$() If cursave <> SampleDir ChangeDir SampleDir readit = ReadDir(SampleDir) Graphics 1000,600,32,2 file$="" : extlen=0 : ex$ ="" : found=False : s$="" Repeat file = Lower$(NextFile(readit)) ;; DO everything low. L = Len(file) If L=0 CloseDir readit Exit End If ft = FileType(file) If ft =1 lastdot = lastdot(file) If lastdot extlen = L-lastdot ex = Right$(file, extlen) ;; grab the extension Else file = file+"." ;; If the file is missing a dot, then give it one L=L+1 : lastdot=L : extlen=0 ;; We shall also count those with no extension. ex = "" End If If extlen ;; if the file has an extension .. found=False ;; assume at first that it is not found in EXT$ i = Instr(EXT, ex) If i ;; if the ext is found, then make sure that it's not a subset of another ext, ;; like when "TIF" is subset of "TIFF", or "JPE" is subset of "JPEG" While(i) L1 = EXTcount(i) Shr 16 ;; = length of that one possible match already in EXT If L1=extlen EXTcount(i) = EXTcount(i) +1 ;; found a perfect match here, so count it i=0 : found=True ;; ..don't count it again in this loop Else i = Instr(EXT, ex, i+1) ;; search for a new match if possible End If Wend End If If found=False ;; if the extension was not found, then add it to the EXT string EXT = EXT +ex i = Instr(EXT,ex) EXTcount(i) = 1 + (extlen Shl 16) ;; ..store stuff in SHORT variables to save space If extCmax-i <10 ;; expand our memory if necessary... extCmax = extCmax +40 ;; add about 40 more Dim EXTcount(extCmax) End If End If Else EXTcount(0) = EXTcount(0)+1 ;; index 0 is reserved for counting files with no extensions End If End If Forever If cursave <> SampleDir ChangeDir cursave ;; change back to the current dir AppTitle CurrentDate$() Print Print " File extension counter program " Print Print" EXT string = "+EXT Print" len(EXT) = "+Len(EXT) Print ;; The lower 2 bytes contain the count, and the higher 2 bytes contain extension length ;; no_ext = EXTcount(0) And $FFFF Print" There were "+no_ext+" files that had no extension. " Print Print" Here are the rest of the counts: " For z= 1 To extCmax ;; now loop through all of it and display the data i = EXTcount(z) And $FFFF ;; if it has a count, then display it If i extlen = EXTcount(z) Shr 16 s = Mid$(EXT, z, extlen) ;; retrieve the actual extension saved in EXT$ Print " "+i+" : "+s End If Next EXT = "" WaitKey():End ;; This function returns the index of the last dot in a string Function LastDot(s$) Local i,j i = Instr(s,".") While(i) j = i i = Instr(s,".",i+1) Wend Return j End Function Function lastslash$(s$) ; Gives a last slash to any string that is missing one L = Len(s) If L If Mid$(s,L,1) = "\" Return s Else Return s+"\" End If Else Return s End If End Function

_PJ_ (Posted 2013) [#2]

Have you considered just checking the CLASSES registry for existing registered extensions?

virtlands (Posted 2013) [#3]

Hi, _PJ_ I don't yet know how to do registry reads from within BlitZ3D. I heard it's a complicated thing. The point of the above program is not just for reading file extensions, but for somehow encountering any random strings, then smashing them together, then counting them.