Computational Concepts Toolbox Data type: values, literals, Higher - PowerPoint PPT Presentation

Computational Concepts Toolbox • Data type: values, literals, • Higher Order Functions operations, – Functions as Values • Expressions, Call – Functions with functions as Abstract Data Types argument expression – Assignment of function • Variables values • Assignment Statement • Higher order function David E. Culler patterns • Sequences: tuple, list – Map, Filter, Reduce CS8 – Computational Structures in Data Science • Data structures • Function factories – create • Tuple assignment http://inst.eecs.berkeley.edu/~cs88 and return functions • Call Expressions • Recursion • Function Definition Lecture 7 Statement – Linear, Tail, Tree March 7, 2016 • Conditional Statement • Iteration: list comp, for, while 2/22/16 2 UCB CS88 Sp16 L4 Universality Administrative Issues • Everything that can be computed, can be • Midterm went well (results on gradescope) computed with what you know now. • Well • March 15 12:30 – 3:00 Study session for • or poorly repeat??? • Lab05 today gets you started on ADTs • Maps project out in lieu of homework – Due Sun 3/20 “before break” – Two-week project 2/22/16 3 2/22/16 4 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4

C.O.R.E concepts Creating an Abtract Data Type • Operations Perform useful computations treating objects abstractly as – Express the behavior of objects, invariants, etc C ompute whole values and operating on – Implemented (abstractly) in terms of Constructors and them. Selectors for the object Abstract Data Type • Representation Provide operations on the O perations abstract components that allow – Constructors & Selectors ease of use – independent of – Implement the structure of the object concrete representation. R epresentation Constructors and selectors that • An abstraction barrier violation occurs when a provide an abstract interface to part of the program that can use the higher level a concrete representation functions uses lower level ones instead Execution on a computing E valuation – At either layer of abstraction machine • Abstraction barriers make programs easier to get right, maintain, and modify Abstraction Barrier – Few changes when representation changes 2/22/16 5 2/22/16 6 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4 Examples You have seen Examples You have seen • Lists • Lists – Constructors : • Tuples » list( … ) • Strings » [ <exps>,… ] – Constructors : » [<exp> for <var> in <list> [ if <exp> ] ] » str( … ) – Selectors : <list> [ <index or slice> ] » “<chars>”, ‘<chars>’ – Operations : in, not in, +, *, len, min, max – Selectors : <str> [ <index or slice> ] » Mutable ones too (but not yet) – Operations : in, not in, +, *, len, min, max • Tuples • Range – Constructors : – Constructors : » tuple( … ) » range(<end>), range(<start>,<end>), » ( <exps>,… ) range(<start>,<end>,<step>) – Selectors : <tuple> [ <index or slice> ] – Selectors : <range> [ <index or slice> ] – Operations : in, not in, +, *, len, min, max – Operations : in, not in, len, min, max 2/22/16 7 2/22/16 8 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4

Example ADT: lookup table (lut) lut ADT • Unordered collection of unique key => value • Constructors bindings – lut() - Return an empty lut – “lookup, i.e., ge t, the value associated with a key” – lut_add(lut, key, value) - Return a lut with new key => value binding • Where does this occur? – lut_del(lut, key) - Return a lut without a binding for key – Phonebook • Selectors – Facebook friends – lut_get(lut, key) - Return value in lut bound to key or – Movie listings None if none exists. application – Restaurant ratings – lut_keys(lut) - Return a list of keys for bindings in lut – Roster – lut_values(lut) - Return a list of values for bindings in lut – … lut operations – lut_items(lut) - Return a list of (key, value) for bindings in lut • Operations lut representation http://cs88-website.github.io/assets/slides/adt/lut.py 2/22/16 9 2/22/16 10 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4 lut ADT The Layered Design Process • Constructors • Build the application based entirely on the ADT interface – lut() , lut_add(lut, key, value), lut_del(lut, key) – Operations, Constructors and Selectors • Selectors • Build the operations entirely in ADT – lut_get(lut, key) , lut_keys(lut), lut_values(lut), lut_items(lut) Constructors and Selectors • Operations – Not the implementation of the representation – lut_with_bindings(bindings) - Return a lut of bindings • Build the constructors and selectors on some – lut_len(lut) - Return the number of bindings in lut. concrete representation – lut_print(lut) - Print a representation of bindings in lut. – lut_map_values(lut, fun) – lut_sorted(lut, fun) – lut_update(lut, key, value) – lut_fuzzy_get(lut, fuzz_key, dist_fun) » Return (key, value) for the key closest to fuzz_key under dist_fun. 2/22/16 11 2/22/16 12 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4

An lut application (lut_app.py) Apps (cont) from lut import * New_book = lut_update(phone_book, "Jack Chow", "805-962-0936") phone_book_data = [ ("Christine Strauch", "510-842-9235"), lut_sorted(new_phone_book, lambda k,v:v) ("Frances Catal Buloan", "932-567-3241"), ("Jack Chow", "617-547-0923"), ("Joy De Rosario", "310-912-6483"), ("Casey Casem", "415-432-9292"), ("Lydia Lu", "707-341-1254")] phone_book = lut_with_bindings(phone_book_data) lut_print(phone_book) print("Jack Chows's Number: ", lut_get(phone_book, "Jack Chow")) print("Area codes") http://cs88-website.github.io/assets/slides/adt/lut_app.py area_codes = lut_map_values(phone_book, lambda x:x[0:3]) lut_print(area_codes) 2/22/16 13 2/22/16 14 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4 Apps (cont) Friends App def name_dist(name1, name2): count = max(len(name1),len(name2)) – friend_data = [ ("Christine Strauch", "Jack Chow"), min(len(name1),len(name2)) ("Christine Strauch", "Lydia Lu"), for i in range(min(len(name1), len(name2))): ("Jack Chow", "Christine Strauch"), if (name1[i] != name2[i]): ("Casey Casem", "Christine Strauch"), count += 1 ("Casey Casem", "Jack Chow"), ("Casey Casem", "Frances Catal Buloan"), return count ("Casey Casem", "Joy De Rosario"), ("Casey Casem", "Casey Casem"), ("Frances Catal Buloan", "Jack Chow"), lut_fuzzy_get(phone_book, "Jack", name_dist)) ("Jack Chow", "Frances Catal Buloan"), ("Joy De Rosario", "Lydia Lu"), ("Joy De Lydia", "Jack Chow") ] 2/22/16 15 2/22/16 16 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4

More Friends Above Abstraction Barrier – lut.py def lut_with_bindings(bindings): def make_friends(friends): """Construct lookup table with (key,val) bindings.""” friend_lut = lut() for (der, dee) in friends: new_lut = lut() old_friends = lut_get (friend_lut, der) for k,v in bindings: new_fr = old_friends + [dee] if old_friends is not None new_lut = lut_add(new_lut, k, v) else [dee] return new_lut friend_lut = lut_update (friend_lut, der, new_fr) return friend_lut 2/22/16 17 2/22/16 18 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4 Aside: lambda Lambda Examples • Function expression >>> msort([1,2,3,4,5], lambda x: x ) [1, 2, 3, 4, 5] – “anonymous” function creation – Expression, not a statement, no return or any other statement >>> msort([1,2,3,4,5], lambda x: -x ) [5, 4, 3, 2, 1] lambda <arg or arg_tuple> : <expression using args> >>> msort([(2, "hi"), (1, "how"), (5, "goes"), (7, "I")], lambda x:x[0] ) [(1, 'how'), (2, 'hi'), (5, 'goes'), (7, 'I')] inc = lambda v : v + 1 def inc(v): >>> msort([(2, "hi"), (1, "how"), (5, "goes"), (7, "I")], return v + 1 lambda x:x[1] ) [(7, 'I'), (5, 'goes'), (2, 'hi'), (1, 'how')] >>> msort([(2,"hi"),(1,"how"),(5,"goes"),(7,"I")], lambda x: len(x[1] )) [(7, 'I'), (2, 'hi'), (1, 'how'), (5, 'goes')] http://cs88-website.github.io/assets/slides/adt/mersort.py 2/22/16 19 2/22/16 20 UCB CS88 Sp16 L4 UCB CS88 Sp16 L4