Handling lists

[Michael Spencer]

superprad at gmail.com wrote:
...
> list = [[10,11,12,13,14,78,79,80,81,300,301,308]]
> how do I convert it so that I arrange them into bins  .
> so If i hvae a set of consecutive numbers i would like to represent
> them as a range in the list with max and min val of the range alone.
> I shd get something like 
> list = [[10,14],[78,81],[300,308]]

> 
Mage:
> Maybe:
> 
> list = [10,11,12,13,14,78,79,80,81,300,301,308]
> 
> new_list = []
> start = 0
> for i in range(1,len(list) + 1):
>     if i == len(list) or list[i] - list[i-1] <> 1:
>         new_list.append([list[start],list[i-1]])
>         start = i
> 
> print new_list

superprad at gmail.com wrote:
> yes that makes sense.But the problem I am facing is  if list=
> [300,301,303,305] I want to consider it as one cluster and include the
> range as [300,305] so this is where I am missing the ranges.
> so If the list has l = [300,301,302,308,401,402,403,408] i want to
> include it as [[300,308],[401,408]].


Mage's solution meets the requirements that you initially stated of treating 
*consecutive* numbers as a group.  Now you also want to consider 
[300,301,303,305] as a cluster.

You need to specify your desired clustering rule, or alternatively specify ho 
many bins you want to create, but as an example, here is a naive approach, that 
could be adapted easily to other clustering rules and (a bit less easily) to 
target a certain number of bins

def lstcluster(lst):
     # Separate neighbors that differ by more than the mean difference
     lst.sort()
     diffs = [(b-a, (a, b)) for a, b in zip(lst,lst[1:])]
     mean_diff = sum(diff[0] for diff in diffs)/len(diffs)
     breaks = [breaks for diff, breaks in diffs if diff > mean_diff]
     groups = [lst[0]] + [i for x in breaks for i in x] + [lst[-1]]
     igroups = iter(groups) # Pairing mechanism due to James Stroud
     return [[i, igroups.next()] for i in igroups]

Note this is quite inefficient due to creating several intermediate lists.  But 
it's not worth optimizing yet, since I'm only guessing at your actual requirement.

lst0 = [10,11,12,13,14,78,79,80,81,300,301,308]
lst1 = [10,12,16,24,26,27,54,55,80,100, 105]
lst3 = [1,5,100,1000,1005,1009,10000, 10010,10019]

  >>> lst0 = [10,11,12,13,14,78,79,80,81,300,301,308]
  >>> lst1 = [10,12,16,24,26,27,54,55,80,100, 105]
  >>> lst2 = [1,5,100,1000,1005,1009,10000, 10010,10019]
  >>> lstcluster(lst0)
  [[10, 14], [78, 81], [300, 308]]
  >>> lstcluster(lst1)
  [[10, 27], [54, 55], [80, 80], [100, 105]]
  >>> lstcluster(lst2)
  [[1, 1009], [10000, 10019]]
  >>>


Michael