theory Parsing

imports Base "Package/Simple_Inductive_Package"

begin

chapter {* Parsing *}

text {*

  Isabelle distinguishes between \emph{outer} and \emph{inner} syntax. 

  Commands, such as \isacommand{definition}, \isacommand{inductive} and so

  on, belong to the outer syntax, whereas terms, types and so on belong to the 

  inner syntax. For parsing inner syntax, 

  Isabelle uses a rather general and sophisticated algorithm, which 

  is driven by priority grammars. Parsers for outer syntax are built up by functional

  parsing combinators. These combinators are a well-established technique for parsing, 

  which has, for example, been described in Paulson's classic ML-book \cite{paulson-ml2}.

  Isabelle developers are usually concerned with writing these outer syntax parsers, 

  either for new definitional packages or for calling methods with specific arguments. 

  \begin{readmore}

  The library for writing parser combinators is split up, roughly, into two

  parts. The first part consists of a collection of generic parser combinators

  defined in the structure @{ML_struct Scan} in the file @{ML_file

  "Pure/General/scan.ML"}. The second part of the library consists of

  combinators for dealing with specific token types, which are defined in the

  structure @{ML_struct OuterParse} in the file @{ML_file

  "Pure/Isar/outer_parse.ML"}. Specific parsers for packages are defined

  in @{ML_file "Pure/Isar/spec_parse.ML"}. Parsers for method arguments

  are defined in @{ML_file "Pure/Isar/args.ML"}.

  \end{readmore}

*}

section {* Building Generic Parsers *}

text {*

  \begin{itemize}

  \item @{text "FAIL"} is used to indicate that alternative routes of parsing 

  might be explored. 

  \item @{text "MORE"} indicates that there is not enough input for the parser. For example 

  in @{text "($$ \"h\") []"}.

  \item @{text "ABORT"} is the exception that is raised when a dead end is reached. 

  It is used for example in the function @{ML "!!"} (see below).

  \end{itemize}

  However, note that these exceptions are private to the parser and cannot be accessed

  by the programmer (for example to handle them).

  Slightly more general than the parser @{ML "$$"} is the function @{ML

  Scan.one}, in that it takes a predicate as argument and then parses exactly

  one item from the input list satisfying this predicate. For example the

  following parser either consumes an @{text [quotes] "h"} or a @{text

  [quotes] "w"}:

@{ML_response [display,gray] 

"let 

  val hw = Scan.one (fn x => x = \"h\" orelse x = \"w\")

in

  (hw input1, hw input2)

end"

    "((\"h\", [\"e\", \"l\", \"l\", \"o\"]),(\"w\", [\"o\", \"r\", \"l\", \"d\"]))"}

  Two parsers can be connected in sequence by using the function @{ML "--"}. 

  For example parsing @{text "h"}, @{text "e"} and @{text "l"} (in this 

  order) you can achieve by:

  @{ML_response [display,gray] 

  "(((\"h\", \"e\"), \"l\"), [\"l\", \"o\"])"}

  Note how the result of consumed strings builds up on the left as nested pairs.  

  If, as in the previous example, you want to parse a particular string, 

  then you should use the function @{ML Scan.this_string}:

  @{ML_response [display,gray] 

  "(\"hell\", [\"o\"])"}

  Parsers that explore alternatives can be constructed using the function @{ML

  "||"}. The parser @{ML "(p || q)" for p q} returns the

  result of @{text "p"}, in case it succeeds, otherwise it returns the

  result of @{text "q"}. For example:

@{ML_response [display,gray] 

"let 

  val hw = $$ \"h\" || $$ \"w\"

in

  (hw input1, hw input2)

end"

  "((\"h\", [\"e\", \"l\", \"l\", \"o\"]), (\"w\", [\"o\", \"r\", \"l\", \"d\"]))"}

  The functions @{ML "|--"} and @{ML "--|"} work like the sequencing function 

  for parsers, except that they discard the item being parsed by the first (respectively second)

  parser. For example:

@{ML_response [display,gray]

"let 

  val just_e = $$ \"h\" |-- $$ \"e\" 

  val just_h = $$ \"h\" --| $$ \"e\" 

in 

  (just_e input, just_h input)

end"

  "((\"e\", [\"l\", \"l\", \"o\"]),(\"h\", [\"l\", \"l\", \"o\"]))"}

  The parser @{ML "Scan.optional p x" for p x} returns the result of the parser 

  @{text "p"}, if it succeeds; otherwise it returns 

  the default value @{text "x"}. For example:

@{ML_response [display,gray]

"let 

  val p = Scan.optional ($$ \"h\") \"x\"

in 

  (p input1, p input2)

end" 

 "((\"h\", [\"e\", \"l\", \"l\", \"o\"]), (\"x\", [\"w\", \"o\", \"r\", \"l\", \"d\"]))"}

  The function @{ML Scan.option} works similarly, except no default value can

  be given. Instead, the result is wrapped as an @{text "option"}-type. For example:

@{ML_response [display,gray]

"let 

  val p = Scan.option ($$ \"h\")

in 

  (p input1, p input2)

end" "((SOME \"h\", [\"e\", \"l\", \"l\", \"o\"]), (NONE, [\"w\", \"o\", \"r\", \"l\", \"d\"]))"}

  The function @{ML "!!"} helps to produce appropriate error messages

  for parsing. For example if you want to parse @{text p} immediately 

  followed by @{text q}, or start a completely different parser @{text r},

  you might write:

  @{ML [display,gray] "(p -- q) || r" for p q r}

  However, this parser is problematic for producing an appropriate error

  message, if the parsing of @{ML "(p -- q)" for p q} fails. Because in that

  case you lose the information that @{text p} should be followed by @{text q}.

  To see this assume that @{text p} is present in the input, but it is not

  followed by @{text q}. That means @{ML "(p -- q)" for p q} will fail and

  hence the alternative parser @{text r} will be tried. However, in many

  circumstances this will be the wrong parser for the input ``@{text "p"}-followed-by-something''

  and therefore will also fail. The error message is then caused by the failure

  of @{text r}, not by the absence of @{text q} in the input. This kind of

  situation can be avoided when using the function @{ML "!!"}.  This function

  aborts the whole process of parsing in case of a failure and prints an error

  message. For example if you invoke the parser

  @{ML [display,gray] "!! (fn _ => \"foo\") ($$ \"h\")"}

  on @{text [quotes] "hello"}, the parsing succeeds

  @{ML_response [display,gray] 

  "(\"h\", [\"e\", \"l\", \"l\", \"o\"])"}

  but if you invoke it on @{text [quotes] "world"}

                               "Exception ABORT raised"}

  then the parsing aborts and the error message @{text "foo"} is printed. In order to

  see the error message properly, you need to prefix the parser with the function 

  @{ML "Scan.error"}. For example:

  @{ML_response_fake [display,gray] 

  "Scan.error (!! (fn _ => \"foo\") ($$ \"h\"))"

  "Exception Error \"foo\" raised"}

  This ``prefixing'' is usually done by wrappers such as @{ML "OuterSyntax.local_theory"} 

  (see Section~\ref{sec:newcommand} which explains this function in more detail). 

  Let us now return to our example of parsing @{ML "(p -- q) || r" for p q

  r}. If you want to generate the correct error message for 

  @{text "p"}-followed-by-@{text "q"}, then you have to write:

*}

ML{*fun p_followed_by_q p q r =

let 

  val err_msg = fn _ => p ^ " is not followed by " ^ q

in

  ($$ p -- (!! err_msg ($$ q))) || ($$ r -- $$ r)

end *}

text {*

  Running this parser with the arguments

  @{text [quotes] "h"}, @{text [quotes] "e"} and @{text [quotes] "w"}, and 

  the input @{text [quotes] "holle"} 

                               "Exception ERROR \"h is not followed by e\" raised"} 

  produces the correct error message. Running it with

                          "((\"w\", \"w\"), [\"o\", \"r\", \"l\", \"d\"])"}

  yields the expected parsing. 

  The function @{ML "Scan.repeat p" for p} will apply a parser @{text p} as 

  often as it succeeds. For example:

                "([\"h\", \"h\", \"h\", \"h\"], [\"e\", \"l\", \"l\", \"o\"])"}

  Note that @{ML "Scan.repeat"} stores the parsed items in a list. The function

  @{ML "Scan.repeat1"} is similar, but requires that the parser @{text "p"} 

  succeeds at least once.

  Also note that the parser would have aborted with the exception @{text MORE}, if

  you had run it only on just @{text [quotes] "hhhh"}. This can be avoided by using

  the wrapper @{ML Scan.finite} and the ``stopper-token'' @{ML Symbol.stopper}. With

  them you can write:

                "([\"h\", \"h\", \"h\", \"h\"], [])"}

  @{ML Symbol.stopper} is the ``end-of-input'' indicator for parsing strings;

  other stoppers need to be used when parsing, for example, tokens. However, this kind of 

  manually wrapping is often already done by the surrounding infrastructure. 

  The function @{ML Scan.repeat} can be used with @{ML Scan.one} to read any 

  string as in

  @{ML_response [display,gray] 

"let 

   val p = Scan.repeat (Scan.one Symbol.not_eof)

in

   Scan.finite Symbol.stopper p input

end" 

"([\"f\", \"o\", \"o\", \" \", \"b\", \"a\", \"r\", \" \", \"f\", \"o\", \"o\"], [])"}

  where the function @{ML Symbol.not_eof} ensures that we do not read beyond the 

  end of the input string (i.e.~stopper symbol).

  The function @{ML "Scan.unless p q" for p q} takes two parsers: if the first one can 

  parse the input, then the whole parser fails; if not, then the second is tried. Therefore

                               "Exception FAIL raised"}

  fails, while

                          "(\"w\",[\"o\", \"r\", \"l\", \"d\"])"}

  succeeds. 

  The functions @{ML Scan.repeat} and @{ML Scan.unless} can be combined to read any

  input until a certain marker symbol is reached. In the example below the marker

  symbol is a @{text [quotes] "*"}.

  @{ML_response [display,gray]

"let 

  val p = Scan.repeat (Scan.unless ($$ \"*\") (Scan.one Symbol.not_eof))

in

  (Scan.finite Symbol.stopper p input1, 

   Scan.finite Symbol.stopper p input2)

end"

"(([\"f\", \"o\", \"o\", \"o\", \"o\", \"o\"], []),

 ([\"f\", \"o\", \"o\"], [\"*\", \"o\", \"o\", \"o\"]))"}

  After parsing is done, you almost always want to apply a function to the parsed 

  items. One way to do this is the function @{ML "(p >> f)" for p f}, which runs 

  first the parser @{text p} and upon successful completion applies the 

  function @{text f} to the result. For example

@{ML_response [display,gray]

"let 

  fun double (x, y) = (x ^ x, y ^ y) 

in

end"

"((\"hh\", \"ee\"), [\"l\", \"l\", \"o\"])"}

  doubles the two parsed input strings; or

  @{ML_response [display,gray] 

"let 

   val p = Scan.repeat (Scan.one Symbol.not_eof)

in

   Scan.finite Symbol.stopper (p >> implode) input

end" 

"(\"foo bar foo\",[])"}

  where the single-character strings in the parsed output are transformed

  back into one string.

  (FIXME:  move to an earlier place)

  The function @{ML Scan.ahead} parses some input, but leaves the original

  input unchanged. For example:

  @{ML_response [display,gray]

  "(\"foo\", [\"f\", \"o\", \"o\"])"} 

  The function @{ML Scan.lift} takes a parser and a pair as arguments. This function applies

  the given parser to the second component of the pair and leaves the  first component 

  untouched. For example

@{ML_response [display,gray]

"((\"h\", \"e\"), (1, [\"l\", \"l\", \"o\"]))"}

  (FIXME: In which situations is this useful? Give examples.) 

  \begin{exercise}\label{ex:scancmts}

  Write a parser that parses an input string so that any comment enclosed

  within @{text "(*\<dots>*)"} is replaced by the same comment but enclosed within

  @{text "(**\<dots>**)"} in the output string. To enclose a string, you can use the

  function @{ML "enclose s1 s2 s" for s1 s2 s} which produces the string @{ML

  "s1 ^ s ^ s2" for s1 s2 s}. Hint: To simplify the task ignore the proper 

  nesting of comments.

  \end{exercise}

*}

section {* Parsing Theory Syntax *}

text {*

  Most of the time, however, Isabelle developers have to deal with parsing

  tokens, not strings.  These token parsers have the type:

*}

ML{*type 'a parser = OuterLex.token list -> 'a * OuterLex.token list*}

text {*

  The reason for using token parsers is that theory syntax, as well as the

  parsers for the arguments of proof methods, use the type @{ML_type

  OuterLex.token}.

  \begin{readmore}

  The parser functions for the theory syntax are contained in the structure

  @{ML_struct OuterParse} defined in the file @{ML_file  "Pure/Isar/outer_parse.ML"}.

  The definition for tokens is in the file @{ML_file "Pure/Isar/outer_lex.ML"}.

  \end{readmore}

  The structure @{ML_struct OuterLex} defines several kinds of tokens (for

  example @{ML "Ident" in OuterLex} for identifiers, @{ML "Keyword" in

  OuterLex} for keywords and @{ML "Command" in OuterLex} for commands). Some

  token parsers take into account the kind of tokens. The first example shows

  how to generate a token list out of a string using the function @{ML

  "OuterSyntax.scan"}. It is given the argument @{ML "Position.none"} since,

  at the moment, we are not interested in generating precise error

  messages. The following code\footnote{Note that because of a possible bug in

  the PolyML runtime system, the result is printed as @{text [quotes] "?"},

  instead of the tokens.}

@{ML_response_fake [display,gray] "OuterSyntax.scan Position.none \"hello world\"" 

"[Token (\<dots>,(Ident, \"hello\"),\<dots>), 

 Token (\<dots>,(Space, \" \"),\<dots>), 

 Token (\<dots>,(Ident, \"world\"),\<dots>)]"}

  produces three tokens where the first and the last are identifiers, since

  @{text [quotes] "hello"} and @{text [quotes] "world"} do not match any

  other syntactic category. The second indicates a space.

  We can easily change what is recognised as a keyword with 

  @{ML OuterKeyword.keyword}. For example calling this function 

*}

ML{*val _ = OuterKeyword.keyword "hello"*}

text {*

  then lexing @{text [quotes] "hello world"} will produce

  @{ML_response_fake [display,gray] "OuterSyntax.scan Position.none \"hello world\"" 

"[Token (\<dots>,(Keyword, \"hello\"),\<dots>), 

 Token (\<dots>,(Space, \" \"),\<dots>), 

 Token (\<dots>,(Ident, \"world\"),\<dots>)]"}

  Many parsing functions later on will require spaces, comments and the like

  to have already been filtered out.  So from now on we are going to use the 

  functions @{ML filter} and @{ML OuterLex.is_proper} to do this. For example:

@{ML_response_fake [display,gray]

"let

   val input = OuterSyntax.scan Position.none \"hello world\"

in

   filter OuterLex.is_proper input

end" 

"[Token (\<dots>,(Ident, \"hello\"), \<dots>), Token (\<dots>,(Ident, \"world\"), \<dots>)]"}

  For convenience we define the function:

*}

ML{*fun filtered_input str = 

  filter OuterLex.is_proper (OuterSyntax.scan Position.none str) *}

text {* 

  If you now parse

@{ML_response_fake [display,gray] 

"filtered_input \"inductive | for\"" 

"[Token (\<dots>,(Command, \"inductive\"),\<dots>), 

 Token (\<dots>,(Keyword, \"|\"),\<dots>), 

 Token (\<dots>,(Keyword, \"for\"),\<dots>)]"}

  you obtain a list consisting of only one command and two keyword tokens.

  If you want to see which keywords and commands are currently known to Isabelle, type in

  the following code (you might have to adjust the @{ML print_depth} in order to

  see the complete list):

@{ML_response_fake [display,gray] 

"let 

  val (keywords, commands) = OuterKeyword.get_lexicons ()

in 

  (Scan.dest_lexicon commands, Scan.dest_lexicon keywords)

end" 

"([\"}\", \"{\", \<dots>], [\"\<rightleftharpoons>\", \"\<leftharpoondown>\", \<dots>])"}

  The parser @{ML "OuterParse.$$$"} parses a single keyword. For example:

@{ML_response [display,gray]

"let 

  val input1 = filtered_input \"where for\"

  val input2 = filtered_input \"| in\"

in 

  (OuterParse.$$$ \"where\" input1, OuterParse.$$$ \"|\" input2)

end"

"((\"where\",\<dots>), (\"|\",\<dots>))"}

  Any non-keyword string can be parsed with the function @{ML OuterParse.reserved}.