Author Topic: LIKE + (Read 21675 times)

John · « **Reply #30 on:** November 05, 2018, 11:35:33 AM »

I think I should be effective with regex before trying another cryptic tool like 8th.

AIR · « **Reply #31 on:** November 05, 2018, 01:35:40 PM »

Well, it's not so much about if the language can do it, but about how would you implement it.

Python makes it simple too....

AIR.

AIR · « **Reply #32 on:** November 05, 2018, 01:49:49 PM »

Quote from: John on November 05, 2018, 11:35:33 AM

I think I should be effective with regex before trying another cryptic tool like 8th.

Regex is one of those things that will drive you crazy until you have an "AHA!" moment and it makes sense.

Perl is a good way to practice; PCRE was created with that syntax in mind...If I recall, the SB "RE" module is posix-based, meaning it uses the posix syntax, which is an entirely different way of doing stuff.

John · « **Reply #33 on:** November 05, 2018, 01:51:33 PM »

Keep in mind that the split BY can consist of more than one character.

Quote

If I recall, the SB "RE" module is posix-based, meaning it uses the posix syntax, which is an entirely different way of doing stuff.

It now makes sense why none of my testing of the RE module worked.

John · « **Reply #34 on:** November 05, 2018, 09:14:00 PM »

Another feature I'm planning on adding to the Script BASIC EXTRACT is if the user passes a negative value in the optional third argument, EXTRACT will act LIKE a function and return the MATCH (just the value MATCH[-x,0]) for the place holder position rather than the default occurance count. If no optional argument is passed, EXTRACT emulates SB native LIKE returning TRUE / FALSE.

Curious if BaCon can return various types? (numeric or a string)

AIR · « **Reply #35 on:** November 05, 2018, 09:58:26 PM »

I don't think you can do that in Bacon with an array/associate array, but you should be able to do it using an array of RECORDs (struct, in C parlance). But the number of entries per RECORD is fixed and typed so if you set up a string as a field, you can't directly place a number in there. You may be able to just set up longs and pass the address of the variables, but then you get into some C-pointer like territory...

SB finagles this by essentially treating everything as a string, so you're able to do this in your arrays....

John · « **Reply #36 on:** November 05, 2018, 11:25:08 PM »

SB doesn't allow returning arrays. You have to pass them back and forth as byref (default) arguments. SB variables are in the form of a variant.

The MATCH array is created in the EXTRACT function but it's a global variable not a local array.

AIR · « **Reply #37 on:** November 09, 2018, 10:06:06 PM »

Hey John, did you ever finish the SB version?

John · « **Reply #38 on:** November 10, 2018, 07:53:00 AM »

Close but got sidetracked playing in the sandbox. I hope to have some time this weekend to finish it.

AIR · « **Reply #39 on:** November 10, 2018, 11:43:49 AM »

I figured as much. I'm really interested in what you come up with, especially the lower level module bits.

John · « **Reply #40 on:** November 10, 2018, 08:18:26 PM »

This is a very early release that gets the patterns MATCH items but not the other data that surrounds it (non-relevent jokers)

Code: ScriptBasic

' EXTRACT (LIKE +) Code Challenge - Script BASIC / JRS
 
html = """<!DOCTYPE html>
<html>
  <head>
    <title>AllBASIC.INFO Forum LIKE Code Challenge</title>
  </head>
  <body>
LIKE it or don't.
  </body>
</html>"""
 
FUNCTION EXTRACT(base, mask)
  UNDEF MATCH
  start = 1
  match_idx = 2
  SPLITA mask BY "*" TO patterns
  FOR idx = 0 TO UBOUND(patterns) STEP 2
    MATCH[match_idx, 1] = INSTR(base, patterns[idx], start)
    start = MATCH[match_idx, 1]
    MATCH[match_idx, 2] = INSTR(base, patterns[idx + 1], start)
    start = MATCH[match_idx, 2]
    MATCH[match_idx, 1] = MATCH[match_idx, 1] + LEN(patterns[idx])
    MATCH[match_idx, 2] = MATCH[match_idx, 2] - MATCH[match_idx, 1]
    MATCH[match_idx, 0] = MID(base, MATCH[match_idx, 1], MATCH[match_idx, 2])
    match_idx += 2
  NEXT
END FUNCTION
 
EXTRACT html, "*<title>*</title>*<body>*</body>*"
 
PRINT "Title:  ", MATCH[2, 0],"\n"
PRINT "Start:  ", MATCH[2, 1],"\n"
PRINT "Length: ", MATCH[2, 2],"\n"
PRINTNL
PRINT "Body:   ", MATCH[4, 0],"\n"
PRINT "Start:  ", MATCH[4, 1],"\n"
PRINT "Length: ", MATCH[4, 2],"\n"
 

$ time scriba extract.sb Title: AllBASIC.INFO Forum LIKE Code Challenge Start: 44 Length: 39 Body: LIKE it or don't. Start: 110 Length: 21 real 0m0.007s user 0m0.008s sys 0m0.000s $

BTW: Firefox for Windows is now at 63.0.1.

John · « **Reply #41 on:** November 11, 2018, 08:47:19 AM »

The EXTRACT function seems to work fine with the Firefox download challenge replacing LIKE. One issue I have discovered with this version of EXTRACT is a pattern needs to be more than one character.

Code: ScriptBasic

' Firefox Download Challenge (Script BASIC Version) by JRS.
 
INCLUDE curl.bas
 
FUNCTION EXTRACT(base, mask)
  UNDEF MATCH
  start = 1
  match_idx = 2
  SPLITA mask BY "*" TO patterns
  FOR idx = 0 TO UBOUND(patterns) STEP 2
    MATCH[match_idx, 1] = INSTR(base, patterns[idx], start)
    start = MATCH[match_idx, 1]
    MATCH[match_idx, 2] = INSTR(base, patterns[idx + 1], start)
    start = MATCH[match_idx, 2]
    MATCH[match_idx, 1] = MATCH[match_idx, 1] + LEN(patterns[idx])
    MATCH[match_idx, 2] = MATCH[match_idx, 2] - MATCH[match_idx, 1]
    MATCH[match_idx, 0] = MID(base, MATCH[match_idx, 1], MATCH[match_idx, 2])
    match_idx += 2
  NEXT
  EXTRACT = UBOUND(patterns) + 1 / 2
END FUNCTION
 
 
ch = curl::init()
curl::option(ch,"URL","https://www.mozilla.org/en-US/firefox/new/")
wp = curl::perform(ch)
 
IF EXTRACT(wp, """*data-latest-firefox="*" data-esr-versions*<div id="other-platforms">*<li class="os_win64">*<a href="*"
            class*""") THEN
  PRINT "Downloading Latest 64Bit Firefox (",MATCH[2, 0],") for Windows.\n"
  curl::option(ch,"FOLLOWLOCATION", 1)
  curl::option(ch,"NOPROGRESS",0)
  curl::option(ch,"FILE","Firefox_Setup-" & MATCH[2, 0] & ".exe")
  curl::option(ch,"URL", MATCH[6, 0])
  curl::perform(ch)
  PRINTNL
  PRINT "Firefox_Setup-" & MATCH[2, 0] & ".exe Downloaded ",FORMAT("%~##,###,###~ Bytes",curl::info(ch,"SIZE_DOWNLOAD")), _
  " at ",FORMAT("%~##,###,###~ Bytes/Second",curl::info(ch,"SPEED_DOWNLOAD")),".\n"
ELSE
  PRINT "<< ERROR >>\n"
END IF
 
curl::finish(ch)
 

$ time scriba ff_extract.sb Downloading Latest 64Bit Firefox (63.0.1) for Windows. % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 134 100 134 0 0 611 0 --:--:-- --:--:-- --:--:-- 611 100 42.3M 100 42.3M 0 0 22.1M 0 0:00:01 0:00:01 --:--:-- 34.3M Firefox_Setup-63.0.1.exe Downloaded 44,396,144 Bytes at 23,219,740 Bytes/Second. real 0m2.201s user 0m1.075s sys 0m0.214s $

John · « **Reply #42 on:** November 12, 2018, 05:07:15 PM »

I will be posting another update to EXTRACT with better error handling and an optional argument to specify which MATCH item you want returned by EXTRACT rather than the total MATCH count. Multiple occurrences gets unmanageable deciphering which MATCH index to select for a given occurrence.

John · « **Reply #43 on:** November 12, 2018, 06:08:37 PM »

Script BASIC

Note: Only JOKERS surrounded by match patterns are returned. Others are ignored. If you need them, use LIKE. I have REM'ed the UNDEF MATCH in this example to show both options EXTRACT offers. I tested the EXTRACT function with the Firefox download page and it returned the correct number of JOKERS and returned the data requested using the optional argument.

Code: ScriptBasic

'' EXTRACT (LIKE +) Code Challenge - Script BASIC / JRS
 
html = """<!DOCTYPE html>
<html>
  <head>
    <title>AllBASIC.INFO Forum LIKE Code Challenge</title>
  </head>
  <body>
LIKE it or don't.
  </body>
</html>"""
 
FUNCTION EXTRACT(base, mask, select)
  UNDEF MATCH
  start = 1
  match_idx = 2
  SPLITA mask BY "*" TO patterns
  FOR idx = 0 TO UBOUND(patterns) STEP 2
    MATCH[match_idx, 1] = INSTR(base, patterns[idx], start)
    IF MATCH[match_idx, 1] <> undef THEN
      start = MATCH[match_idx, 1]
    ELSE
      GOTO MATCH_ERROR
    END IF
    MATCH[match_idx, 2] = INSTR(base, patterns[idx + 1], start)
    IF MATCH[match_idx, 2] <> undef THEN
      start = MATCH[match_idx, 2]
    ELSE
      GOTO MATCH_ERROR
    END IF
    MATCH[match_idx, 1] = MATCH[match_idx, 1] + LEN(patterns[idx])
    MATCH[match_idx, 2] = MATCH[match_idx, 2] - MATCH[match_idx, 1]
    MATCH[match_idx, 0] = MID(base, MATCH[match_idx, 1], MATCH[match_idx, 2])
    match_idx += 2
  NEXT
  IF select <> undef AND select <= UBOUND(patterns) + 1 THEN
    EXTRACT = MATCH[select, 0]
    ' UNDEF MATCH
  ELSE
    EXTRACT = UBOUND(patterns) + 2
  END IF
  GOTO DONE
 
MATCH_ERROR:
  EXTRACT = 0
 
DONE:
END FUNCTION
 
PRINT EXTRACT(html, "*<title>*</title>*<body>*</body>*", 2), "\n"
PRINTNL
PRINT "Title:  ", MATCH[2, 0],"\n"
PRINT "Start:  ", MATCH[2, 1],"\n"
PRINT "Length: ", MATCH[2, 2],"\n"
PRINTNL
PRINT "Body:   ", MATCH[4, 0],"\n"
PRINT "Start:  ", MATCH[4, 1],"\n"
PRINT "Length: ", MATCH[4, 2],"\n"
 

$ time scriba extract.sb AllBASIC.INFO Forum LIKE Code Challenge Title: AllBASIC.INFO Forum LIKE Code Challenge Start: 44 Length: 39 Body: LIKE it or don't. Start: 110 Length: 21 real 0m0.009s user 0m0.005s sys 0m0.004s $

John · « **Reply #44 on:** November 13, 2018, 02:59:58 PM »

Is there anyone besides me willing to post a working EXTRACT / MATCH submission for this code challenge?

AllBASIC Forum / ScriptBasic Blog

News:

Author Topic: LIKE + (Read 21675 times)

John

Re: LIKE +

AIR

Re: LIKE +

AIR

Re: LIKE +

John

Re: LIKE +

John

Re: LIKE +

AIR

Re: LIKE +

John

Re: LIKE +

AIR

Re: LIKE +

John

Re: LIKE +

AIR

Re: LIKE +

John

Re: LIKE +

John

Re: LIKE +

John

Re: LIKE +

John

Re: LIKE +

John

Re: LIKE +