code/train/Python/0001972_drillcluster.py

#!/usr/bin/env python

"""
Drill clustering routines to reduce total number of drills and remap 
drilling commands to the new reduced drill set.

--------------------------------------------------------------------

This program is licensed under the GNU General Public License (GPL)
Version 3.  See http://www.fsf.org for details of the license.

Rugged Circuits LLC
http://ruggedcircuits.com/gerbmerge
"""

_STATUS = True ## indicates status messages should be shown
_DEBUG = False ## indicates debug and status messages should be shown

def cluster(drills, tolerance, debug = _DEBUG):
    """
        Take a dictionary of drill names and sizes and cluster them
        A tolerance of 0 will effectively disable clustering
        
        Returns clustered drill dictionary
    """
    
    global _DEBUG
    _DEBUG = debug
    
    clusters = []

    debug_print("\n  " + str( len(drills) ) + " Original drills:")
    debug_print( drillsToString(drills) )
    debug_print("Clustering drill sizes ...", True)

    # Loop through all drill sizes
    sizes = drills.keys()
    sizes.sort()
    for size in sizes:
        
        match = False
        
        # See if size fits into any current clusters, else make new cluster
        for index in range( len(clusters) ):
            c = clusters[index]
            if not len(c):
                break
            mn = min(c)
            mx = max(c)
            
            ##debug_print( "Validating " + str_d(size) + " in " + str_d(c) ) 
            ##debug_print( "Possible cluster range = " + str_d(mx - 2 * tolerance) + " to " + str_d(mn + 2 * tolerance) )
            
            if (size >= mx - 2 * tolerance) and (size <= mn + 2 * tolerance):
               
               debug_print( str_d(size) + " belongs with " + str_d(c) )
               
               clusters[index].append(size)
               match = True
               break
            
        if not match:
            debug_print(str_d(size) + " belongs in a new cluster")
            clusters.append( [size] )
            
    debug_print("\n  Creating new drill dictionary ...")    
    
    new_drills = {}
    tool_num = 0

    # Create new dictionary of clustered drills
    for c in clusters:
        tool_num += 1
        new_drill = "T%02d" % tool_num
        c.sort()
        new_size = ( min(c) + max(c) ) / 2.0
        new_drills[new_size] = new_drill
        
        debug_print(str_d(c) + " will be represented by " + new_drill + " (" + str_d(new_size) + ")")
        
    debug_print("\n  " + str( len(new_drills) ) + " Clustered Drills:")
    debug_print( drillsToString(new_drills) )
    debug_print("Drill count reduced from " + str( len(drills) ) + " to " + str( len(new_drills) ), True)
    
    return new_drills

def remap(jobs, globalToolMap, debug = _DEBUG):
    """
        Remap tools and commands in all jobs to match new tool map
        
        Returns None
    """
    
    # Set global variables from parameters
    global _DEBUG
    _DEBUG = debug
    
    debug_print("Remapping tools and commands ...", True)
    
    for job in jobs:
        job = job.job ##access job inside job layout
        debug_print("\n  Job name: " + job.name)
        debug_print("\n  Original job tools:")
        debug_print( str(job.xdiam) )
        debug_print("\n  Original commands:")
        debug_print( str(job.xcommands) )
        new_tools = {}
        new_commands = {}
        for tool, diam in job.xdiam.items():
        
            ##debug_print("\n  Current tool: " + tool + " (" + str_d(diam) + ")")
        
            # Search for best matching tool
            best_diam, best_tool = globalToolMap[0]          

            for glob_diam, glob_tool in globalToolMap:
                if abs(glob_diam - diam) < abs(best_diam - diam):
                    best_tool = glob_tool
                    best_diam = glob_diam
            ##debug_print("Best match: " + best_tool + " (" + str_d(best_diam) + ")")
            new_tools[best_tool] = best_diam
            ##debug_print(best_tool + " will replace " + tool)

            # Append commands to existing commands if they exist
            if best_tool in new_commands:
                ##debug_print( "Current commands: " + str( new_commands[best_tool] ) )
                temp = new_commands[best_tool]
                temp.extend( job.xcommands[tool] )
                new_commands[best_tool] = temp
                ##debug_print( "All commands: " + str( new_commands[best_tool] ) )
            else:
                new_commands[best_tool] = job.xcommands[tool]
            
        debug_print("\n  New job tools:")
        debug_print( str(new_tools) )
        debug_print("\n  New commands:")
        debug_print( str(new_commands) )
        job.xdiam = new_tools 
        job.xcommands = new_commands
    
def debug_print(text, status = False, newLine = True):
    """
        Print debugging statemetns
        
        Returs None, Printts text
    """
    
    if _DEBUG or (status and _STATUS):
        if newLine:
            print " ", text
        else:
            print " ", text,
            
def str_d(drills):
    """
        Format drill sizes for printing debug and status messages
        
        Returns drills as formatted string
    """
    
    string = ""
    
    try:
        len(drills)
    except:
        string =  "%.4f" % drills
    else:
        string = "["
        for drill in drills:
            string += ( "%.4f" % drill + ", ")
        string = string[:len(string) - 2] + "]"
        
    return string
    
def drillsToString(drills):
    """
        Format drill dictionary for printing debug and status messages
        
        Returns drills as formatted string
    """
    string = ""
    
    drills = drills.items()
    drills.sort()
    for size, drill in drills:
        string += drill + " = " + str_d(size) + "\n  "

    return string
    
""" 
    The following code runs test drill clusterings with random drill sets.
"""

if __name__=="__main__":
  import random
  
  print "  Clustering random drills..."

  old = {}
  tool_num = 0
  while len(old) < 99:
      rand_size = round(random.uniform(.02, .04), 4)
      if rand_size in old:
          continue
      tool_num += 1
      old[rand_size] = "T%02d" % tool_num
      
  new = cluster(old, .0003, True)