def __init__(self, bcmdict, wanteddict, elemdict, vendict):
self.vendorlist = list() # contains the active list of vendors.
# All matrix keys match this list.
self.elementlist = list() # contains the list of elements.
# All matrix keys match this list.
self.wanted = None # numpy array
self.prices = None # numpy array
self.stock = None # numpy array
self.bcmdict = bcmdict
self.__initialize_lists(self.bcmdict)
self.wantdict = wanteddict
self.elemdict = elemdict
self.vendict = vendict
self.wanted = self.__buildwantedarray(self.wantdict)
self.stock, self.prices = self.__buildvendorarrays()
self.__vs = VendorStats(self)
self.__update()
self.__need_rebuild = False
class BCMData():
""" Contains the core data of pybcm.
It contains several constant references, as well as mutable data, that
is used by other classes in the system.
Attributes:
vendorsortdict(): returns a dictionary that defines sorting
removevendor( vendorid ): remove a vendor from the vendorlist
removevendors( vendorindices ): remove the list of vendor indices
from the vendorlsit
replacevendorlist(): replace the vendor with
avgprices( ): return an array of average element prices
"""
def __init__(self, bcmdict, wanteddict, elemdict, vendict):
self.vendorlist = list() # contains the active list of vendors.
# All matrix keys match this list.
self.elementlist = list() # contains the list of elements.
# All matrix keys match this list.
self.wanted = None # numpy array
self.prices = None # numpy array
self.stock = None # numpy array
self.bcmdict = bcmdict
self.__initialize_lists(self.bcmdict)
self.wantdict = wanteddict
self.elemdict = elemdict
self.vendict = vendict
self.wanted = self.__buildwantedarray(self.wantdict)
self.stock, self.prices = self.__buildvendorarrays()
self.__vs = VendorStats(self)
self.__update()
self.__need_rebuild = False
#self. = None
#self.VENDSORTLIST = self.__createvendsortinglist(self.BCMDICT)
def __update(self):
"""Update the various arrays."""
self.__sortlists()
self.__updatearrays()
self.__vs.update(self)
#self.AVGPRICES = self.avgprices(stockweighted=True)
def __initialize_lists(self, bcmdict):
""" Build the initial elementlist and vendorlist based on bcmdict."""
logging.info("Building BCMData lists")
#k = [ keytuple for keytuple in bcmdict.keys() ]
for keytuple in list(bcmdict.keys()):
(elementid, vendorid) = keytuple
addtolist(self.vendorlist, vendorid)
addtolist(self.elementlist, elementid)
def __updatearrays(self):
"""Build the various arrays."""
logging.info("Forcing array update...")
self.wanted = self.__buildwantedarray(self.wantdict)
self.__updatevendorarrays()
self.__need_rebuild = False
def __buildvendorarrays(self):
""" Iterate over element, vendor in the elementlist and vendorlist to create the numpy
arrays. Get the data from the bcmdict.
"""
shape = (len(self.elementlist), len(self.vendorlist))
pricearray = ma.masked_array(np.zeros(shape, dtype='int'))
stockarray = ma.masked_array(np.zeros(shape, dtype='int'))
#wanted doesnt change
for eindex, element in enumerate(self.elementlist):
for vindex, vendor in enumerate(self.vendorlist):
if (element, vendor) in self.bcmdict:
stockarray[eindex, vindex] = int(self.bcmdict[element, vendor][0])
pricearray[eindex, vindex] = int(self.bcmdict[element, vendor][1] * 100)
# clip the max value of stock to the wanted quantity
stockarray = np.minimum(stockarray, self.wanted.reshape(len(self.elementlist), 1))
mask = stockarray <= 0
stockarray.mask = mask
pricearray.mask = mask
return stockarray, pricearray
def __updatevendorarrays(self):
""" Create new arrays in case elementlist and vendorlits have changed size """
stockarray, pricearray = self.__buildvendorarrays()
self.prices = pricearray
self.stock = stockarray
return stockarray, pricearray
def __buildwantedarray(self, wanteddict): # returns numpy array of WANTED items
""" Create a numpy array of wanted quantities """
logging.info("Building WANTED array...")
m = len(self.elementlist) # ensure the size of the array is consistent with the others
wantedarray = np.ndarray(shape=(m,), dtype=np.int)
for eidx, elementid in enumerate(self.elementlist):
wantedarray[eidx] = wanteddict[elementid]
return wantedarray
def __sortlists(self):
self.__elementsort()
self.__vendorsort()
def __elementsort(self, sortweights=None):
logging.info("Sorting Element List...")
if sortweights:
weights = sortweights
else:
weights = self.elementweights()
#resort the elementlist using these weights
self.elementlist = [y for (x, y) in sorted(zip(weights, self.elementlist), reverse=True)]
def __vendorsort(self, sortby='uniqueitems'):
logging.info("Sorting Vendor List...")
if sortby == 'uniqueitems':
weights = self.__vs.itemspervendor
elif sortby == 'totalitems':
weights = self.__vs.totalvendor
else:
return # nothing sorted
self.vendorlist = [y for (x, y) in sorted(zip(weights, self.vendorlist), reverse=True)]
def __createvendsortinglist(self):
""" Return list of tuples (vendor index, element index, price)
The element index is the element of the highest weight that the vendor offers in sufficient qty
"""
factor = 1.0
k = list()
for vidx, vcol in enumerate(self.stock.T):
for eidx, stock in enumerate(vcol):
if stock > self.wanted[eidx] * factor:
price = self.prices[eidx, vidx]
stock = self.stock[eidx, vidx]
k.append((vidx, eidx, price, stock))
break
# now sort this list on eidx=>ascending, price=>descending, stock->satisfies(eidx)
k = sorted(k, key=itemgetter(1, 2))
return k
def vendorsortdict(self):
"""Return a dictionary describing how to sort vendors
k[vendor index] = (elementid to sort on, price, qty)
Important: It's assumed that the arrays are already sorted on the
element weights, meaning the most costly elements are first. This
algorithim uses the first element the vendor stocks in sufficient quantity
"""
#size = len(self.vendorlist)
#k = np.zeros(shape=(size), dtype=(int, float))
k = dict()
for vidx, vcol in enumerate(self.prices.T): # iterate over columns in the price matrix
for eidx, price in enumerate(vcol):
if price > 0:
qty = self.stock[eidx, vidx]
k[vidx] = (eidx, price, qty)
break
return k
def __sufficientqty(self, eidx, vidx, factor=0.5):
return self.stock[eidx, vidx] >= self.wanted[eidx] * factor
def removevendor(self, vendorid):
#doesn't remove it from the .data, only from the list of vendors
assert vendorid in self.vendorlist, "Vendor %r does not exist in vendorlist" % vendorid
self.vendorlist.remove(vendorid)
self.__need_rebuild = True
def removevendors(self, vendorindices):
#Create a new vendorlist that doesn't include the list of id's passed via vendorindices
logging.info("Trying to remove " + str(len(vendorindices)) + " vendors.")
before = len(self.vendorlist)
newlist = [vendor for vendor in self.vendorlist if self.vendorlist.index(vendor) not in vendorindices]
self.replacevendorlist(newlist)
after = len(self.vendorlist)
logging.info("Removed: " + str(before - after) + " vendors.")
self.__update()
return newlist
def replacevendorlist(self, newvendors):
self.vendorlist = newvendors
#remove all items that contain these vendors from the dictionaries?
self.__update()
def avgprices(self, stockweighted=False):
"""Return a masked array of the average price by element"""
p = ma.array(self.prices, mask=self.prices <= 0)
if stockweighted:
s = ma.array(self.stock, mask=self.stock <= 0)
avgprices = ma.average(p, weights=s, axis=1)
else:
#avgprices = p.sum(axis=1)/(p > 0).sum(axis=1) #denominator sums the non-zero values
avgprices = ma.average(p, axis=1)
return avgprices
def elementweights(self):
#generate a weight for each element - basically the avg price for that element * WANTED qty, normalized
weights = self.wanted * self.avgprices()
return weights
def __itemspervendor(self):
s = self.stock
itemspervendor = np.ndarray(s > 0).sum(0)
return itemspervendor
def partcount(self):
"""count the number of unique parts and the total number of parts"""
return len(self.wanted), self.wanted.sum()
