Parsing

Fall 2005 Costas Buch - RPI 1

 Machine Code
Program Add v,v,0
v = 5; cmp v,5
if (v>5) jmplt ELSE
x = 12 + v; THEN:
while (x !=3) { Compiler
add x, 12,v
x = x - 3; ELSE:
v = 10; WHILE:
} cmp x,3
...... ...
Fall 2005 Costas Buch - RPI 2
 Compiler

Lexical
parser
analyzer

input output

machine
program
code
Fall 2005 Costas Buch - RPI 3
 A parser knows the grammar
of the programming language

Fall 2005 Costas Buch - RPI 4

 Parser
PROGRAM  STMT_LIST
STMT_LIST  STMT; STMT_LIST | STMT;
STMT  EXPR | IF_STMT | WHILE_STMT
| { STMT_LIST }

EXPR  EXPR + EXPR | EXPR - EXPR | ID

IF_STMT  if (EXPR) then STMT
| if (EXPR) then STMT else STMT
WHILE_STMT while (EXPR) do STMT

Fall 2005 Costas Buch - RPI 5

The parser finds the derivation
of a particular input

derivation
Parser
E => E + E
input
E -> E + E => E + E * E
10 + 2 * 5 |E*E => 10 + E*E
| INT => 10 + 2 * E
=> 10 + 2 * 5

Fall 2005 Costas Buch - RPI 6

 derivation derivation tree
E
E => E + E
=> E + E * E E + E

=> 10 + E*E 10
E
* E
=> 10 + 2 * E
=> 10 + 2 * 5 2 5

machine code
mult a, 2, 5
add b, 10, a
Fall 2005 Costas Buch - RPI 7
 A simple parser

Fall 2005 Costas Buch - RPI 8

 We will build an exhaustive search parser
that examines all possible derivations

Exhaustive Parser
input
grammar derivation
string

Fall 2005 Costas Buch - RPI 9

 Example:

Find derivation of string aabb

Exhaustive Parser
S  SS derivation
input
S  aSb
aabb ?
S  bSa
S 

Fall 2005 Costas Buch - RPI 10

 Exhaustive Search

S  SS | aSb | bSa | 

Phase 1: S  SS Find derivation

of aabb
S  aSb
S  bSa
S 
All possible derivations of length 1
Fall 2005 Costas Buch - RPI 11
 S  SS Find derivation
of aabb
S  aSb
S  bSa
S 

Fall 2005 Costas Buch - RPI 12

 Phase 2 S  SS | aSb | bSa | 
S  SS  SSS
S  SS  aSbS
Phase 1 S  SS  bSaS
Find derivation
S  SS S  SS  S of aabb
S  aSb S  aSb  aSSb
S  aSb  aaSbb
S  aSb  abSab
Fall 2005
S  aSb  ab
Costas Buch - RPI 13
 S  SS | aSb | bSa | 
Phase 2
S  SS  SSS
S  SS  aSbS Find derivation
of aabb
S  SS  S

S  aSb  aSSb
S  aSb  aaSbb
Phase 3
S  aSb  aaSbb  aabb
Fall 2005 Costas Buch - RPI 14
 Final result of exhaustive search

Exhaustive Parser
S  SS
input
S  aSb
aabb
S  bSa
S 
derivation
S  aSb  aaSbb  aabb
Fall 2005 Costas Buch - RPI 15
 The time complexity of exhaustive search

Suppose there are no productions of the form

A (  productions)
A B (unit productions)

Number of phases for string w : at most 2 | w |

Since in every phase: either a new variable is inserted
or a variable changes to terminal
Fall 2005 Costas Buch - RPI 16
For grammar with k productions

Derivation steps for phase 1: at most k

since there are at most k possible derivations

Fall 2005 Costas Buch - RPI 17

 2
Steps for phase 2: at most k  k  k

Derivations Number of
from phase 1 Productions

In General
(i 1) i
Steps for phase i: at most k k  k

Derivations Number of
from phase i-1 Productions
Fall 2005 Costas Buch - RPI 18
Total steps needed for string w:

Exponential to the string length

2 2|w|
k  k  k

phase 1 phase 2 phase 2|w|

Extremely bad!!!
Fall 2005 Costas Buch - RPI 19
 Faster Parsers

Fall 2005 Costas Buch - RPI 20

There exist faster parsing algorithms
for specialized grammars

S-grammar: A  av
Symbol String of variables

Each pair of variable, terminal ( X , )

appears once in a production X  w

(a restricted version of Greinbach Normal form)

Fall 2005 Costas Buch - RPI 21

 S-grammar example: S  aS
S  bSS
S c

Each string has a unique derivation

S  aS  abSS  abcS  abcc

Fall 2005 Costas Buch - RPI 22

 For S-grammars:

In the exhaustive search parsing

there is only one choice in each phase

Steps for a phase: 1

Total steps for parsing string w : | w|

Fall 2005 Costas Buch - RPI 23

 For general context-free grammars:

There exists a parsing algorithm

that parses a string w in O (| w |3 )

(CYK parser, described next)

Fall 2005 Costas Buch - RPI 24

 The CYK Parsing Algorithm

Input: • Arbitrary Grammar G

in Chomsky Normal Form
• String w

Output: Determine if w L(G )

Number of Steps: 3
O (| w | )

Can be easily converted to a Parser

Fall 2005 Costas Buch - RPI 25
 Basic Idea
Consider a grammar G
In Chomsky Normal Form

Denote by F (w) the set of variables

that generate a string w

*
X  F (w) if X w
Fall 2005 Costas Buch - RPI 26
 F (w) can be computed recursively:
prefix suffix

Write w  uv
If X  F (u ) and Y  F (v)
* *
( X  u) (Y  v)

and there is production Z  XY

Then Z  F (w)
* *
( Z  XY  uY  uv  w)
Fall 2005 Costas Buch - RPI 27
 Compute F (w) by taking the union
all possible decompositions of w
Length Set of Variables
1 That generates w
w  u1v1 H1
2
w  u2v2 H2

|w|-1
w  u|w|1v|w|1 H|w|1

Result: F ( w)  H1  H 2    H|w|1
Fall 2005 Costas Buch - RPI 28
 At the basis of the recursion
we have strings of length 1

F ( )  {Variables that generate symbol  }

symbol X 

Fall 2005 Costas Buch - RPI 29

 In order to determine if w L(G )
we only need to examine if S  F (w)
*
Start variable ( S  w)

Remark:
The whole algorithm can be implemented
with dynamic programming

Fall 2005 Costas Buch - RPI 30

Example:
• Grammar G : S  AB
A  BB | a
B  AB | b

• Determine if w  aabbb  L(G )

Fall 2005 Costas Buch - RPI 31
 Decompose the string aabbb
to all possible substrings
Length
1 a a b b b

2 aa ab bb bb

3 aab abb bbb

4 aabb abbb

5 aabbb
Fall 2005 Costas Buch - RPI 32
 S  AB, A  BB | a, B  AB | b

a a b b b
F ( ) {A} {A} {B} {B} {B}
aa ab bb bb

aab abb bbb

aabb abbb

aabbb
Fall 2005 Costas Buch - RPI 33
 S  AB, A  BB | a, B  AB | b

a a b b b
F ( )  {A} {A} {B} {B} {B}
aa ab bb bb
F ()  {} {S,B} {A} {A}
aab abb bbb

aabb abbb

aabbb
Fall 2005 Costas Buch - RPI 34
 S  AB, A  BB | a, B  AB | b
F (aa ) prefix aa suffix
F (a )  { A} F (a )  { A}
There is no production of form X  AA
Thus, F (aa )  {}

F (ab) prefix ab suffix

F (a )  { A} F (b)  {B}
There are two productions of form X  AB
S  AB, B  AB
Thus, F (ab)  {S , B}
Fall 2005 Costas Buch - RPI 35
 S  AB, A  BB | a, B  AB | b
a a b b b
{A} {A} {B} {B} {B}
Since aa ab bb bb
S  F (w) {} {S,B} {A} {A}
aab abb bbb
{S,B} {A} {S,B}
aabbb L(G )
aabb abbb
{A} {S,B}
aabbb
F (aabbb) 
Fall 2005
{S,B}
Costas Buch - RPI 36
 Approximate number of steps:
2 3
O(| w |  | w |)  O(| w | )

Number of Number of
strings Prefix-suffix
decompositions
for a string

Fall 2005 Costas Buch - RPI 37