SCM

SCM Repository

[inlinedocs] Diff of /pkg/inlinedocs/R/parsers.R
ViewVC logotype

Diff of /pkg/inlinedocs/R/parsers.R

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 109, Mon Jun 7 08:42:15 2010 UTC revision 204, Fri Feb 18 16:48:22 2011 UTC
# Line 1  Line 1 
1    ### combine lists or character strings
2    combine <- function(x,y)UseMethod("combine")
3    
4    ### combine character strings by pasting them together
5    combine.character <- function(x,y)
6        paste(x,y,sep="\n")
7    
8    ### combine lists by adding elements or adding to existing elements
9    combine.list <- function(x,y){
10      toadd <- if(".overwrite"%in%names(y)){
11        y <- y[names(y)!=".overwrite"]
12        rep(TRUE,length(y))
13      }else{
14        !names(y)%in%names(x)
15      }
16      toup <- names(y)[!toadd]
17      x[names(y)[toadd]] <- y[toadd]
18      for(up in toup)x[[up]] <- combine(x[[up]],y[[up]])
19      x
20    ### A list, same type as x, but with added elements from y.
21    }
22    
23    ### Prefix for code comments used with grep and gsub.
24    prefix <- "^[ \t]*###[ \t]"
25    
26    decomment <- function
27    ### Remove comment prefix and join lines of code to form a
28    ### documentation string.
29    (comments
30    ### Character vector of prefixed comment lines.
31     ){
32      gsub(prefix,"",comments)
33    ### String without prefixes or newlines.
34    }
35    
36    forall <- function
37  ### For each object in the package that satisfies the criterion  ### For each object in the package that satisfies the criterion
38  ### checked by subfun, parse source using FUN and return the resulting  ### checked by subfun, parse source using FUN and return the resulting
39  ### documentation list.  ### documentation list.
 forall <- function  
40  (FUN,  (FUN,
41  ### Function to apply to each element in the package.  ### Function to apply to each element in the package.
42   subfun=function(x)TRUE   subfun=function(x)TRUE
# Line 9  Line 44 
44  ### is.function. subfun(x)==TRUE means FUN will be applied to x and  ### is.function. subfun(x)==TRUE means FUN will be applied to x and
45  ### the result will be returned.  ### the result will be returned.
46   ){   ){
47    function(objs,docs,...){    FUN <- FUN
48      f <- function(objs,docs,...){
49      objs <- objs[sapply(objs,subfun)]      objs <- objs[sapply(objs,subfun)]
50      L <- list()      L <- list()
51      for(N in names(docs)){      on.exit(cat(sprintf("Parser Function failed on %s\n",N)))
52        for(N in union(names(docs),names(objs))){
53        o <- objs[[N]]        o <- objs[[N]]
54        L[[N]] <- FUN(src=attr(o,"source"),        L[[N]] <- FUN(src=attr(o,"source"),
55                      name=N,objs=objs,o=o,docs=docs,doc=docs[[N]],...)                      name=N,objs=objs,o=o,docs=docs,doc=docs[[N]],...)
56      }      }
57        on.exit()## remove warning message
58      L      L
59    }    }
60      class(f) <- c("allfun","function")
61      f
62  ### A Parser Function.  ### A Parser Function.
63  }  }
64    
65    ### Print method for functions constructed using forall.
66    print.allfun <- function(x,...){
67      e <- environment(x)
68      cat("Function to apply to every element.\nselector:")
69      print(e$subfun)
70      cat("processor:")
71      print(e$FUN)
72    }
73    
74  ### For each function in the package, do something.  ### For each function in the package, do something.
75  forfun <- function(FUN)forall(FUN,is.function)  forfun <- function(FUN)forall(FUN,is.function)
76    
77  ### Parsers for each object/function that are constructed  kill.prefix.whitespace <- function
78  ### automatically. This is a named list, each element is a list of 2  ### Figure out what the whitespace preceding the example code is, and
79  ### elements: forfun/forall, then a parser function for an individual  ### then delete that from every line.
80  ### object.  (ex
81  forall.parsers <-  ### character vector of example code lines.
82    list(## Extract lots of info from normal functions.   ){
83         parsefun=list(forfun,function(src,name,...){    tlines <- gsub("\\s*","",ex)
84           extract.docs.fun(src,name)    ##tlines <- gsub("#.*","",tlines)
85         }),    prefixes <- unique(gsub("\\S.*","",ex[tlines!=""]))
86         ## Fill in author from DESCRIPTION and titles.    FIND <- prefixes[which.min(nchar(prefixes))]
87         author.from.description=list(forall,function(desc,...){    ## Eliminate leading tabulations or 2/4 spaces
88           list(author=desc[,"Maintainer"])    sub(FIND, "", ex)
89         }),  ### Character vector of code lines with preceding whitespace removed.
90         ## The format section sometimes causes problems, so erase it.  }
91         erase.format=list(forall,function(...){  
92           list(format="")  examples.after.return <- function
93         }),  ### Get examples from inline definitions after return()
94         ## Convert the function name to a title.  ### PhG: this does not work well! Think of these situations:
95         title.from.name=list(forall,function(name,doc,...){  ### 1) You have multiple return() in the code of your function,
96           if("title"%in%names(doc))list() else  ### 2) You have return() appearing is some example code, ...
97           list(title=gsub("[._]"," ",name))  ### I can hardly propose a hack here. The whole code of the function
98         }),  ### must be parsed, and one must determine which one is the last line
99         # PhG: it is tests/FUN.R!!! I would like more flexibility here  ### of code that is actually executed.
100             # please, let me choose which dir to use for examples!  ###
101             ## Get examples for FUN from the file tests/FUN.R  ### I make two propositions here
102         examples.from.testfile=list(forfun,function(name,...){  ### 1) to keep the same mechanism that has the advantage of simplicity
103           tsubdir <- getOption("inlinedocs.exdir")  ###    but to use a special tag
104                   if (is.null(tsubdir)) tsubdir <- "tests"       # Default value  ###    ##examples<< or #{{{examples to separate
105                   tfile <- file.path("..",tsubdir,paste(name,".R",sep=""))  ###    function code from examples explicitly, and
106           if(file.exists(tfile))  ### 2) to place the example in an "ex" attribute
107             list(examples=paste(readLines(tfile),collapse="\n"))  ###    attached to the function
108           else list()  ###    (see next parser). That solution will be also interesting for
109         }),  ###    documenting datasets, something not done yet by inlinedocs!
110         ## Get examples from inline definitions after return()  (src,name="",...) {
111             # PhG: this does not work well! Think at these situations:    ## Look for the examples mark
112             # 1) You have multiple return() in the code of your function,    m <- grep("^\\s*(##examples<<|#\\{\\{\\{examples)", src)
            # 2) You have return() appearing is some example code, ...  
            # I can hardly propose a hack here. The whole code of the function  
            # must be parsed, and one must determine which one is the last line  
            # of code that is actually executed.  
            #  
            # I make two propositions here  
            # 1) to keep the same mechanism that has the advantage of simplicity  
            #    but to use a special tag ##examples<< or #{{{examples to separate  
            #    function code from examples explicitly, and  
            # 2) to place the example in an "ex" attribute attached to the function  
            #    (see next parser). That solution will be also interesting for  
            #    documenting datasets, something not done yet by inlinedocs!  
            examples.after.return = list(forfun, function(name, src, ...) {  
                         # Look for the examples mark  
                         m <- grep("##examples<<|#\\{\\{\\{examples", src)  
113                          if (!length(m)) return(list())                          if (!length(m)) return(list())
114                          if (length(m) > 1)                          if (length(m) > 1)
115                                  warning("More than one examples tag for ", name, ". Taking the last one")      warning("More than one examples tag for ", name,
116                ". Taking the last one")
117                          m <- m[length(m)]                          m <- m[length(m)]
118                          # Look for the lines containing return value comments just before    ## Look for the lines containing return value comments just before
119                          r <- grep("\\s*### ", src[1:(m-1)])                          r <- grep("\\s*### ", src[1:(m-1)])
120                          if (!length(r)) {                          if (!length(r)) {
121                                  value <- NULL                                  value <- NULL
122                          } else {                          } else {
123                                  # Only take consecutive lines before the mark        ## Only take consecutive lines before the mark
124                                  keep <- rev((m - rev(r)) == 1:length(r))                                  keep <- rev((m - rev(r)) == 1:length(r))
125                                  if (!any(keep)) {                                  if (!any(keep)) {
126                                          value <- NULL                                          value <- NULL
# Line 93  Line 128 
128                                          value <- decomment(src[r[keep]])                                          value <- decomment(src[r[keep]])
129                                  }                                  }
130                          }                          }
131                          # Collect now the example code beneath the mark    ## Collect now the example code beneath the mark
132                          ex <- src[(m + 1):(length(src) - 1)]                          ex <- src[(m + 1):(length(src) - 1)]
133                          # Possibly eliminate a #}}} tag    ## Possibly eliminate a #}}} tag
134                          ex <- ex[!grepl("#}}}", ex)]                          ex <- ex[!grepl("#}}}", ex)]
135                          # Eliminate leading tabulations or four spaces    ## Eliminate leading tabulations or four spaces
136                          ex <- sub("^\t|    ", "", ex)    ex <- kill.prefix.whitespace(ex)
137                          # Add an empty line before and after example    ## Add an empty line before and after example
                         ex <- c("", ex, "")  
                         # Return examples and value  
                         list(examples = paste(ex, collapse = "\n"), value = value)  
            }),  
            # PhG: here is what I propose for examples code in the 'ex' attribute  
            examples.in.attr = list(forfun, function (name, o, ...) {  
                         ex <- attr(o, "ex")  
                         if (!is.null(ex)) {  
                                 # Special case for code contained in a function  
                                 if (inherits(ex, "function")) {  
                                         # If source is available, start from there  
                                         src <- attr(ex, "source")  
                                         if (!is.null(src)) {  
                                                 ex <- src  
                                         } else { # Use the body of the function  
                                                 ex <- deparse(body(ex))  
                                         }  
                                         # Eliminate leading and trailing code  
                                         ex <- ex[-c(1, length(ex))]  
                                         # Eliminate leading tabulations or four spaces  
                                         ex <- sub("^\t|    ", "", ex)  
                                         # Add an empty line before and after example  
138                                          ex <- c("", ex, "")                                          ex <- c("", ex, "")
139                                  }    ## Return examples and value
140                                  list(examples = paste(ex, collapse = "\n"))    list(examples = ex, value = value)
                         } else list()  
            }))  
   
 ### List of parser functions that operate on single objects. This list  
 ### is useful for testing these functions, ie  
 ### lonely$parsefun(attr(extract.docs.file,"source"),"extract.docs.file")  
 lonely <- sapply(forall.parsers,function(L)L[[2]])  
   
 extra.code.docs <- function # Extract documentation from code chunks  
 ### Parse R code to extract inline documentation from comments around  
 ### each function. These are not able to be retreived simply by  
 ### looking at the "source" attribute. This is a Parser Function that  
 ### can be used in the parser list of package.skeleton.dx(). TODO:  
 ### Modularize this into separate Parsers Functions for S4 classes,  
 ### prefixes, ##<<blocks, etc. Right now it is not very clean!  
 (code,  
 ### Code lines in a character vector containing multiple R objects to  
 ### parse for documentation.  
  objs,  
 ### The objects defined in the code.  
  ...  
 ### ignored  
  ){  
   parsed <- extract.file.parse(code)  
   extract.docs.try <- function(o,on)  
     {  
       ## Note: we could use parsed information here too, but that  
       ## would produce different results for setMethodS3 etc.  
       doc <- list()  
       if ( !is.null(parsed[[on]]) ){  
         if ( !is.na(parsed[[on]]@code[1]) ){ # no code given for generics  
           doc$definition <- paste(parsed[[on]]@code,collapse="\n")  
         }  
         if(!"description"%in%names(doc) && !is.na(parsed[[on]]@description) ){  
           doc$description <- parsed[[on]]@description  
         }  
         if ( "setMethodS3" == parsed[[on]]@created ){  
           # PhG: this may be wrong! It does not catch correctly how the method  
                   # must be splitted in case of methods containing dots. for instance,  
                   # as.data.frame.matrix must be split into: m1 = as.data.frame and  
                   # m2 = matrix... here you got m1 = as, and m2 = data.frame.matrix!!!  
                   pattern <- "^([^\\.]+)\\.(.*)$"  
           doc$s3method=c(m1 <- gsub(pattern,"\\1",on,perl=TRUE),  
               m2 <- gsub(pattern,"\\2",on,perl=TRUE))  
           if ( 0 < length(grep("\\W",m1,perl=TRUE)) ){  
                           m1 <- paste("`",m1,"`",sep="")  
           }  
           cat("S3method(",m1,",",m2,")\n",sep="")  
         }  
       }  
       if("title" %in% names(doc) && !"description" %in% names(doc) ){  
         ## For short functions having both would duplicate, but a  
         ## description is required. Therefore automatically copy title  
         ## across to avoid errors at package build time.  
         doc$description <- doc$title  
       }  
       doc  
     }  
   extract.docs <- function(on){  
     res <- try({o <- objs[[on]]  
                 extract.docs.try(o, on)},FALSE)  
     if(class(res)=="try-error"){  
       cat("Failed to extract docs for: ",on,"\n\n")  
       list()  
     } else if(0 == length(res) && inherits(objs[[on]],"standardGeneric")){  
       NULL  
     } else if(0 == length(res) && "function" %in% class(o)  
               && 1 == length(osource <- attr(o,"source"))  
               && 1 == length(grep(paste("UseMethod(",on,")",sep="\""),osource))  
               ){  
       ## phew - this should only pick up R.oo S3 generic definitions like:  
       ## attr(*, "source")= chr "function(...) UseMethod(\"select\")"  
       NULL  
     } else res  
   }  
   doc.names <- names(objs)  
   res <- sapply(doc.names,extract.docs,simplify=FALSE)  
   ## Special processing for S4 classes as they do not appear in normal ls()  
   for ( nn in names(parsed) ){  
     if ( parsed[[nn]]@created == "setClass" ){  
       S4class.docs <- extract.docs.setClass(parsed[[nn]])  
       docname <- paste(nn,"class",sep="-")  
       if ( is.null(res[[docname]]) ){  
         res[[docname]] <- S4class.docs  
         doc.names <- c(doc.names,docname)  
       } else {  
         stop(nn," appears as both S4 class and some other definition")  
       }  
     }  
   }  
   inherit.docs <- function(on){  
     in.res <- res[[on]]  
     if ( !is.null(parsed[[on]]) ){  
       for ( parent in parsed[[on]]@parent ){  
         if ( !is.na(parent) ){  
           if ( is.null(in.res) ){  
             in.res <- res[[parent]]  
           } else if ( parent %in% names(res) ){  
             parent.docs <- res[[parent]]  
             for ( nn in names(parent.docs) ){  
               if ( !nn %in% names(in.res) ){  
                 in.res[[nn]] <- parent.docs[[nn]]  
               }  
             }  
           }  
         }  
       }  
     }  
     invisible(in.res)  
   }  
   all.done <- FALSE  
   while ( !all.done ){  
     res1 <- sapply(doc.names,inherit.docs,simplify=FALSE)  
     all.done <- identical(res1,res)  
     res <- res1  
   }  
   ## now strip out any generics (which have value NULL in res):  
   res.not.null <- sapply(res,function(x){!is.null(x)})  
   if ( 0 < length(res.not.null) && length(res.not.null) < length(res) ){  
     res <- res[res.not.null]  
   }  
   res  
 ### named list of lists, one for each object to document.  
141  }  }
142    
143  ### List of parsers to use by default with package.skeleton.dx.  prefixed.lines <- structure(function(src,...){
144  default.parsers <-  ### The primary mechanism of inline documentation is via consecutive
145    c(extra.code.docs=extra.code.docs,  ### groups of lines matching the specified prefix regular expression
146      sapply(forall.parsers,function(L)L[[1]](L[[2]])),  ### "\code{^### }" (i.e. lines beginning with "\code{### }") are
147      edit.package.file=function(desc,...){  ### collected as follows into documentation sections:\describe{
148        in.details <- setdiff(colnames(desc),"Description")  ### \item{description}{group starting at line 2 in the code}
149        details <- paste(paste(in.details,": \\tab ",desc[,in.details],"\\cr",  ### \item{arguments}{group following each function argument}
150                               sep=""),collapse="\n")  ### \item{value}{group ending at the penultimate line of the code}}
151        L <-  ### These may be added to by use of the \code{##<<} constructs
152          list(list(title=desc[,"Title"],  ### described below.
153                    description=desc[,"Description"],    clines <- grep(prefix,src)
154                    `tabular{ll}`=details,    if(length(clines)==0)return(list())
                   author=desc[,"Maintainer"]))  
       names(L) <- paste(desc[,"Package"],"-package",sep="")  
       L  
     })  
   
 extract.docs.fun <- function # Extract documentation from a function  
 ### Given source code of a function, return a list describing inline  
 ### documentation in that source code.  
 (code,  
 ### The function to examine.  
  name.fun  
 ### The name of the function/chunk to use in warning messages.  
  ){  
   res <- list()  
   clines <- grep(prefix,code)  
   if(length(grep("#",code[1]))){  
     res$title <- gsub("[^#]*#\\s*(.*)","\\1",code[1],perl=TRUE)  
   }  
   if(length(clines) > 0){  
     ##details<<  
     ## The primary mechanism is that consecutive groups of lines matching  
     ## the specified prefix regular expression "\code{^### }" (i.e. lines  
     ## beginning with "\code{### }") are collected  
     ## as follows into documentation sections:\describe{  
     ## \item{description}{group starting at line 2 in the code}  
     ## \item{arguments}{group following each function argument}  
     ## \item{value}{group ending at the penultimate line of the code}}  
     ## These may be added to by use of the \code{##<<} constructs described  
     ## below.  
155      bounds <- which(diff(clines)!=1)      bounds <- which(diff(clines)!=1)
156      starts <- c(1,bounds+1)      starts <- c(1,bounds+1)
157      ends <- c(bounds,length(clines))      ends <- c(bounds,length(clines))
158      ## detect body of function using paren matching
159      code <- gsub("#.*","",src)
160      f <- function(ch)cumsum(nchar(gsub(sprintf("[^%s]",ch),"",code)))
161      parens <- f("(")-f(")")
162      body.begin <- which(diff(parens)<0 & parens[-1]==0)+2
163      if(length(body.begin)==0)body.begin <- 1 ## rare cases
164      is.arg <- function(){
165        gres <- grep("^\\s*#",src[start-1],perl=TRUE)
166        0 == length(gres) && start<=body.begin
167      }
168      res <- list()
169      for(i in seq_along(starts)){      for(i in seq_along(starts)){
170        start <- clines[starts[i]]        start <- clines[starts[i]]
171        end <- clines[ends[i]]        end <- clines[ends[i]]
172        lab <- if(end+1==length(code))"value"      lab <- if(all(grepl("^\\s*#",src[end:(length(src)-1)])))"value"
173        else if(start==2)"description"        else if(start==2)"description"
174        else if ( 0 == length(grep("^\\s*#",code[start-1],perl=TRUE)) ){      else if(is.arg()){
175           #arg <- gsub("^[ (]*","",code[start-1])        ##twutz: strip leading white spaces and brackets and ,
176           #arg <- gsub("^([^=,]*)[=,].*","\\1",arg)        arg <- gsub("^[ \t(,]*", "", src[start - 1])
          #arg <- gsub("...","\\dots",arg,fix=TRUE) ##special case for dots  
                  arg <- gsub("^[ \t(,]*", "", code[start - 1])  #twutz: strip leading white spaces and brackets and ,  
177                   arg <- gsub("^([^=,]*)[=,].*", "\\1", arg)                   arg <- gsub("^([^=,]*)[=,].*", "\\1", arg)
178                   arg <- gsub("^([^ \t]*)([ \t]+)$","\\1",arg)   #twutz: remove trailing whitespaces        ##twutz: remove trailing whitespaces
179          arg <- gsub("^([^ \t]*)([ \t]+)$","\\1",arg)
180                   arg <- gsub("...", "\\dots", arg, fix = TRUE)                   arg <- gsub("...", "\\dots", arg, fix = TRUE)
181           paste("item{",arg,"}",sep="")           paste("item{",arg,"}",sep="")
182         } else {         } else {
183           next;           next;
184         }         }
185        res[[lab]] <- decomment(code[start:end])      res[[lab]] <- decomment(src[start:end])
     }  
186    }    }
187      res
188    },ex=function(){
189    test <- function
190    ### the description
191    (x,
192    ### the first argument
193     y ##<< another argument
194     ){
195      5
196    ### the return value
197    ##seealso<< foobar
198    }
199    src <- attr(test,"source")
200    prefixed.lines(src)
201    extract.xxx.chunks(src)
202    })
203    
204    extract.xxx.chunks <- function # Extract documentation from a function
205    ### Given source code of a function, return a list describing inline
206    ### documentation in that source code.
207    (src,
208    ### The source lines of the function to examine, as a character
209    ### vector.
210     name.fun="(unnamed function)",
211    ### The name of the function/chunk to use in warning messages.
212     ...
213    ### ignored.
214     ){
215      res <- list()
216    ##details<< For simple functions/arguments, the argument may also be    ##details<< For simple functions/arguments, the argument may also be
217    ## documented by appending \code{##<<} comments on the same line as the    ## documented by appending \code{##<<} comments on the same line as the
218    ## argument name. Mixing this mechanism with \code{###} comment lines for    ## argument name. Mixing this mechanism with \code{###} comment lines for
# Line 386  Line 285 
285          chunk.sep <- "\n\n"          chunk.sep <- "\n\n"
286        }        }
287        chunk.res <- NULL        chunk.res <- NULL
288        if ( 0 == length(grep("^\\s*$",payload,perl=TRUE)) )        if ( !grepl("^\\s*$",payload,perl=TRUE) )
289          chunk.res <-          chunk.res <-
290            if ( is.null(res[[field]]) ) payload            if ( is.null(res[[field]]) ) payload
291            else paste(res[[field]], payload, sep=chunk.sep)            else paste(res[[field]], payload, sep=chunk.sep)
292        invisible(chunk.res)        invisible(chunk.res)
293      }      }
294    while ( k <= length(code) ){    while ( k <= length(src) ){
295      line <- code[k]      line <- src[k]
296      if ( 0 < length(grep(extra.regexp,line,perl=TRUE) ) ){      if ( grepl(extra.regexp,line,perl=TRUE) ){
297        ## we have a new extra chunk - first get field name and any payload        ## we have a new extra chunk - first get field name and any payload
298        new.field <- gsub(extra.regexp,"\\1",line,perl=TRUE)        new.field <- gsub(extra.regexp,"\\1",line,perl=TRUE)
299        new.contents <- gsub(extra.regexp,"\\2",line,perl=TRUE)        new.contents <- gsub(extra.regexp,"\\2",line,perl=TRUE)
# Line 469  Line 368 
368            first.describe <- TRUE;            first.describe <- TRUE;
369          }          }
370        }        }
371      } else if ( in.chunk && 0<length(grep(cont.re,line,perl=TRUE)) ){      } else if ( in.chunk && grepl(cont.re,line,perl=TRUE) ){
372        ## append this line to current chunk        ## append this line to current chunk
373        if ( 0 == length(grep(prefix,line,perl=TRUE)) ){        if ( !grepl(prefix,line,perl=TRUE) ){
374          ##describe<< Any lines with "\code{### }" at the left hand          ##describe<< Any lines with "\code{### }" at the left hand
375          ## margin within the included chunks are handled separately,          ## margin within the included chunks are handled separately,
376          ## so if they appear in the documentation they will appear          ## so if they appear in the documentation they will appear
# Line 484  Line 383 
383            payload <- stripped            payload <- stripped
384          }          }
385        }        }
386      } else if ( 0 < length(grep(arg.pat,line,perl=TRUE)) ){      } else if ( grepl(arg.pat,line,perl=TRUE) ){
387        not.describe <- (0==in.describe && !first.describe)        not.describe <- (0==in.describe && !first.describe)
388        if ( in.chunk && not.describe){        if ( in.chunk && not.describe){
389          res[[cur.field]] <- end.chunk(cur.field,payload)          res[[cur.field]] <- end.chunk(cur.field,payload)
# Line 493  Line 392 
392        arg <- gsub(arg.pat,"\\\\item\\{\\1\\}",line,perl=TRUE)        arg <- gsub(arg.pat,"\\\\item\\{\\1\\}",line,perl=TRUE)
393        in.chunk <- TRUE        in.chunk <- TRUE
394        if ( not.describe ){        if ( not.describe ){
395            ## TDH 2010-06-18 For item{}s in the documentation list names,
396            ## we don't need to have a backslash before, so delete it.
397            arg <- gsub("^[\\]+","",arg)
398          cur.field <- gsub("...","\\dots",arg,fix=TRUE) ##special case for dots          cur.field <- gsub("...","\\dots",arg,fix=TRUE) ##special case for dots
399          payload <- comment          payload <- comment
400        } else {        } else {
# Line 543  Line 445 
445  ### with the string in this list (implemented in modify.Rd.file).  ### with the string in this list (implemented in modify.Rd.file).
446  }  }
447    
448    ### Parsers for each function that are constructed automatically. This
449    ### is a named list, and each element is a parser function for an
450    ### individual object.
451    forfun.parsers <-
452      list(prefixed.lines=prefixed.lines,
453           examples.after.return=examples.after.return,
454           extract.xxx.chunks=extract.xxx.chunks,
455           ## title from first line of function def
456           title.from.firstline=function(src,...){
457             first <- src[1]
458             if(!is.character(first))return(list())
459             if(!grepl("#",first))return(list())
460             list(title=gsub("[^#]*#\\s*(.*)","\\1",first,perl=TRUE))
461           },
462           ## PhG: it is tests/FUN.R!!! I would like more flexibility here
463           ## please, let me choose which dir to use for examples!
464           ## Get examples for FUN from the file tests/FUN.R
465           examples.from.testfile=function(name,...){
466             tsubdir <- getOption("inlinedocs.exdir")
467             if (is.null(tsubdir)) tsubdir <- "tests"       # Default value
468             tfile <- file.path("..",tsubdir,paste(name,".R",sep=""))
469             if(file.exists(tfile))
470               list(examples=readLines(tfile))
471             else list()
472           },
473           definition.from.source=function(doc,src,...){
474             def <- doc$definition
475             is.empty <- function(x)is.null(x)||x==""
476             if(is.empty(def) && !is.empty(src))
477               list(definition=src)
478             else list()
479           })
480    
481    ### List of Parser Functions that can be applied to any object.
482    forall.parsers <-
483      list(## Fill in author from DESCRIPTION and titles.
484           author.from.description=function(desc,...){
485             list(author=desc[,"Maintainer"])
486           },
487           ## The format section sometimes causes problems, so erase it.
488           erase.format=function(...){
489             list(format="")
490           },
491           ## Convert the function name to a title.
492           title.from.name=function(name,doc,...){
493             if("title"%in%names(doc))list() else
494             list(title=gsub("[._]"," ",name))
495           },
496           ## PhG: here is what I propose for examples code in the 'ex' attribute
497           examples.in.attr =  function (name, o, ...) {
498             ex <- attr(o, "ex")
499             if (!is.null(ex)) {
500               ## Special case for code contained in a function
501               if (inherits(ex, "function")) {
502                 ## If source is available, start from there
503                 src <- attr(ex, "source")
504                 if (!is.null(src)) {
505                   ex <- src
506                 } else { ## Use the body of the function
507                   ex <- deparse(body(ex))
508                 }
509                 ## Eliminate leading and trailing code
510                 ex <- ex[-c(1, length(ex))]
511                 ## all the prefixes
512                 ex <- kill.prefix.whitespace(ex)
513                 ## Add an empty line before and after example
514                 ex <- c("", ex, "")
515               }
516               list(examples = ex)
517             } else list()
518           },collapse=function(doc,...){
519             L <- lapply(doc,paste,collapse="\n")
520             L$.overwrite <- TRUE
521             L
522           },tag.s3methods=function(name,env,...){
523             parts <- strsplit(name, ".", fixed = TRUE)[[1]]
524             l <- length(parts)
525             if (l > 1) {
526               for (i in 1:(l - 1)) {
527                 ## Look for a generic function (known by the system or defined
528                 ## in the package) that matches that part of the function name
529                 generic <- paste(parts[1:i], collapse = ".")
530                 if (any(generic %in% utils:::getKnownS3generics()) ||
531                     utils:::findGeneric(generic, env) != "") {
532                   object <- paste(parts[(i + 1):l], collapse = ".")
533                   return(list(.s3method=c(generic, object)))
534                 }
535               }
536             }
537             list()
538           }
539           )
540    
541    ### List of parser functions that operate on single objects. This list
542    ### is useful for testing these functions.
543    lonely <- structure(c(forall.parsers,forfun.parsers),ex=function(){
544      f <- function # title
545    ### description
546      (x, ##<< arg x
547       y
548    ### arg y
549       ){
550        ##value<< a list with elements
551        list(x=x, ##<< original x value
552             y=y, ##<< original y value
553             sum=x+y) ##<< their sum
554        ##end<<
555      }
556      src <- attr(f,"source")
557      lonely$extract.xxx.chunks(src)
558      lonely$prefixed.lines(src)
559    })
560    
561    extra.code.docs <- function # Extract documentation from code chunks
562    ### Parse R code to extract inline documentation from comments around
563    ### each function. These are not able to be retreived simply by
564    ### looking at the "source" attribute. This is a Parser Function that
565    ### can be used in the parser list of package.skeleton.dx(). TODO:
566    ### Modularize this into separate Parsers Functions for S4 classes,
567    ### prefixes, ##<<blocks, etc. Right now it is not very clean!
568    (code,
569    ### Code lines in a character vector containing multiple R objects to
570    ### parse for documentation.
571     objs,
572    ### The objects defined in the code.
573     ...
574    ### ignored
575     ){
576      parsed <- extract.file.parse(code)
577      extract.docs.try <- function(o,on)
578        {
579          ## Note: we could use parsed information here too, but that
580          ## would produce different results for setMethodS3 etc.
581          doc <- list()
582          if ( !is.null(parsed[[on]]) ){
583            if ( !is.na(parsed[[on]]@code[1]) ){ # no code given for generics
584              doc$definition <- paste(parsed[[on]]@code)
585            }
586            if(!"description"%in%names(doc) && !is.na(parsed[[on]]@description) ){
587              doc$description <- parsed[[on]]@description
588            }
589            if ( "setMethodS3" == parsed[[on]]@created ){
590              # PhG: this may be wrong! It does not catch correctly how the method
591              # must be splitted in case of methods containing dots. for instance,
592              # as.data.frame.matrix must be split into: m1 = as.data.frame and
593              # m2 = matrix... here you got m1 = as, and m2 = data.frame.matrix!!!
594                      pattern <- "^([^\\.]+)\\.(.*)$"
595              doc$.s3method=c(m1 <- gsub(pattern,"\\1",on,perl=TRUE),
596                  m2 <- gsub(pattern,"\\2",on,perl=TRUE))
597              if ( grepl("\\W",m1,perl=TRUE) ){
598                              m1 <- paste("`",m1,"`",sep="")
599              }
600              cat("S3method(",m1,",",m2,")\n",sep="")
601            }
602          }
603          if("title" %in% names(doc) && !"description" %in% names(doc) ){
604            ## For short functions having both would duplicate, but a
605            ## description is required. Therefore automatically copy title
606            ## across to avoid errors at package build time.
607            doc$description <- doc$title
608          }
609          doc
610        }
611      extract.docs <- function(on){
612        res <- try({o <- objs[[on]]
613                    extract.docs.try(o, on)},FALSE)
614        if(class(res)=="try-error"){
615          cat("Failed to extract docs for: ",on,"\n\n")
616          list()
617        } else if(0 == length(res) && inherits(objs[[on]],"standardGeneric")){
618          NULL
619        } else if(0 == length(res) && "function" %in% class(o)
620                  && 1 == length(osource <- attr(o,"source"))
621                  && grepl(paste("UseMethod(",on,")",sep="\""),osource)
622                  ){
623          ## phew - this should only pick up R.oo S3 generic definitions like:
624          ## attr(*, "source")= chr "function(...) UseMethod(\"select\")"
625          NULL
626        } else res
627      }
628      doc.names <- names(objs)
629      res <- sapply(doc.names,extract.docs,simplify=FALSE)
630      ## Special processing for S4 classes as they do not appear in normal ls()
631      for ( nn in names(parsed) ){
632        if ( parsed[[nn]]@created == "setClass" ){
633          S4class.docs <- extract.docs.setClass(parsed[[nn]])
634          docname <- paste(nn,"class",sep="-")
635          if ( is.null(res[[docname]]) ){
636            res[[docname]] <- S4class.docs
637            doc.names <- c(doc.names,docname)
638          } else {
639            stop(nn," appears as both S4 class and some other definition")
640          }
641        }
642      }
643      inherit.docs <- function(on){
644        in.res <- res[[on]]
645        if ( !is.null(parsed[[on]]) ){
646          for ( parent in parsed[[on]]@parent ){
647            if ( !is.na(parent) ){
648              if ( is.null(in.res) ){
649                in.res <- res[[parent]]
650              } else if ( parent %in% names(res) ){
651                parent.docs <- res[[parent]]
652                for ( nn in names(parent.docs) ){
653                  if ( !nn %in% names(in.res) ){
654                    in.res[[nn]] <- parent.docs[[nn]]
655                  }
656                }
657              }
658            }
659          }
660        }
661        invisible(in.res)
662      }
663      all.done <- FALSE
664      while ( !all.done ){
665        res1 <- sapply(doc.names,inherit.docs,simplify=FALSE)
666        all.done <- identical(res1,res)
667        res <- res1
668      }
669      ## now strip out any generics (which have value NULL in res):
670      res.not.null <- sapply(res,function(x){!is.null(x)})
671      if ( 0 < length(res.not.null) && length(res.not.null) < length(res) ){
672        res <- res[res.not.null]
673      }
674      res
675    ### named list of lists, one for each object to document.
676    }
677    
678    ### List of parsers to use by default with package.skeleton.dx.
679    default.parsers <-
680      c(extra.code.docs=extra.code.docs, ## TODO: cleanup!
681        sapply(forfun.parsers,forfun),
682        edit.package.file=function(desc,...){
683          in.details <- setdiff(colnames(desc),"Description")
684          details <- sprintf("%s: \\tab %s\\cr",in.details,desc[,in.details])
685          L <-
686            list(list(title=desc[,"Title"],
687                      description=desc[,"Description"],
688                      `tabular{ll}`=details))
689          names(L) <- paste(desc[,"Package"],"-package",sep="")
690          L
691        },
692        sapply(forall.parsers,forall)
693        )
694    
695  setClass("DocLink", # Link documentation among related functions  setClass("DocLink", # Link documentation among related functions
696  ### The \code{.DocLink} class provides the basis for hooking together  ### The \code{.DocLink} class provides the basis for hooking together
697  ### documentation of related classes/functions/objects. The aim is that  ### documentation of related classes/functions/objects. The aim is that
# Line 581  Line 730 
730      ## "prefix" lines will be used instead.      ## "prefix" lines will be used instead.
731      default.description <- NULL      default.description <- NULL
732      while ( start > last.end+1      while ( start > last.end+1
733             && 1 == length(grep(prefix,code[start-1],perl=TRUE)) ){             && grepl(prefix,code[start-1],perl=TRUE) ){
734        start <- start-1        start <- start-1
735      }      }
736      if ( start < chunks[[k]][1] ){      if ( start < chunks[[k]][1] ){
# Line 608  Line 757 
757        ## If the function definition is not embedded within the call, then        ## If the function definition is not embedded within the call, then
758        ## the parent is that function. Test whether the the third value        ## the parent is that function. Test whether the the third value
759        ## looks like a name and add it to parents if so.        ## looks like a name and add it to parents if so.
760        if ( 1 == length(grep("^[\\._\\w]+$",chars[3],perl=TRUE)) ){        if ( grepl("^[\\._\\w]+$",chars[3],perl=TRUE) ){
761          parent <- chars[3]          parent <- chars[3]
762        }        }
763        res[[object.name]] <- new("DocLink",name=object.name,        res[[object.name]] <- new("DocLink",name=object.name,
# Line 640  Line 789 
789        ## If the function definition is not embedded within the call, then        ## If the function definition is not embedded within the call, then
790        ## the parent is that function. Test whether the the fourth value        ## the parent is that function. Test whether the the fourth value
791        ## looks like a name and add it to parents if so.        ## looks like a name and add it to parents if so.
792        if ( 1 == length(grep("^[\\._\\w]+$",chars[4],perl=TRUE)) ){        if ( grepl("^[\\._\\w]+$",chars[4],perl=TRUE) ){
793          parent <- c(chars[4],parent)          parent <- c(chars[4],parent)
794        }        }
795        res[[object.name]] <- new("DocLink",name=object.name,        res[[object.name]] <- new("DocLink",name=object.name,
# Line 681  Line 830 
830    ## the same line or \code{### } comments at the beginning of the    ## the same line or \code{### } comments at the beginning of the
831    ## following line.    ## following line.
832    f.n <- paste(class.name,"class",sep="-")    f.n <- paste(class.name,"class",sep="-")
833    docs <- extract.docs.fun(chunk.source,f.n)    docs <- extract.xxx.chunks(chunk.source,f.n)
834      ## also apply source parsing functions that I separated out into
835      ## separate functions
836      docs <- combine(docs,lonely$prefixed.lines(chunk.source))
837      docs$title <- lonely$title.from.firstline(chunk.source)
838    ##details<<    ##details<<
839    ## The class definition skeleton includes an \code{Objects from the Class}    ## The class definition skeleton includes an \code{Objects from the Class}
840    ## section, to which any \code{##details<<} documentation chunks are    ## section, to which any \code{##details<<} documentation chunks are
# Line 719  Line 872 
872  ### Additional arguments to pass to Parser Functions.  ### Additional arguments to pass to Parser Functions.
873   ){   ){
874    e <- new.env()    e <- new.env()
875    old <- options(keep.source.pkgs=TRUE)    old <- options(keep.source=TRUE)
876    on.exit(options(old))    on.exit(options(old))
   ##tryCatch({  
877      exprs <- parse(text=code)      exprs <- parse(text=code)
878      for (i in exprs) eval(i, e)    for (i in exprs){
879  ##   },error=function(err){      tryCatch(eval(i, e),error=function(e){
880  ##     print(i)        ##print(e)
881  ##     stop("eval or parse failed with error:\n",err)      })
882  ##   })    }
883    objs <- sapply(ls(e),get,e,simplify=FALSE)    objs <- sapply(ls(e),get,e,simplify=FALSE)
884    
885    docs <- list()    docs <- list()
886    
   # PhG: Automatically determine who is S3 methods  
   for (name in names(objs)) {  
     parts <- strsplit(name, ".", fixed = TRUE)[[1]]  
         l <- length(parts)  
         if (l > 1) {  
         for (i in 1:(l - 1)) {  
             # Look for a generic function (known by the system or defined  
             # in the package) that matches that part of the function name  
             generic <- paste(parts[1:i], collapse = ".")  
             if (any(generic %in% utils:::getKnownS3generics()) ||  
                 utils:::findGeneric(generic, e) != "") {  
                 object <- paste(parts[(i + 1):l], collapse = ".")  
                 docs[[name]]$s3method <- c(generic, object)  
                 break  
             }  
         }  
     }  
   }  
887    ## apply parsers in sequence to code and objs    ## apply parsers in sequence to code and objs
888    for(i in seq_along(parsers)){    for(i in seq_along(parsers)){
889      N <- names(parsers[i])      N <- names(parsers[i])
# Line 760  Line 894 
894      }      }
895      p <- parsers[[i]]      p <- parsers[[i]]
896      ## This is the argument list that each parser receives:      ## This is the argument list that each parser receives:
897      L <- p(code=code,objs=objs,docs=docs,...)      L <- p(code=code,objs=objs,docs=docs,env=e,...)
898      docs <- combine(docs,L)      docs <- combine(docs,L)
899    }    }
900      ## post-process to collapse all character vectors
901      for(i in seq_along(docs)){
902        for(j in seq_along(docs[[i]])){
903          if(names(docs[[i]])[j]!=".s3method")
904          docs[[i]][[j]] <- paste(docs[[i]][[j]],collapse="\n")
905        }
906      }
907    if(verbose)cat("\n")    if(verbose)cat("\n")
908    docs    docs
909  ### A list of extracted documentation from code.  ### A list of extracted documentation from code.
910  }  }
911    
912    ### Names of Parser Functions that operate on the desc arg.
913    descfile.names <- c("author.from.description","edit.package.file")
914    
915    ### Names of Parser Functions that do NOT use the desc arg.
916    non.descfile.names <-
917      names(default.parsers)[!names(default.parsers)%in%descfile.names]
918    
919    ### Parsers that operate only on R code, independently of the
920    ### description file.
921    nondesc.parsers <- default.parsers[non.descfile.names]
922    
923  extract.docs.file <- function  extract.docs.file <- function
924  ### Apply all parsers relevant to extract info from just 1 code file.  ### Apply all parsers relevant to extract info from just 1 code file.
925  (f,  (f,
926  ### File name of R code to read and parse.  ### File name of R code to read and parse.
927   parsers=list(extra.code.docs,default.parsers$parsefun),   parsers=NULL,
928  ### Parser Functions to use to parse the code and extract  ### Parser Functions to use to parse the code and extract
929  ### documentation.  ### documentation.
930   ...   ...
931  ### Other arguments to pass to Parser Functions.  ### Other arguments to pass to Parser Functions.
932   ){   ){
933      if(is.null(parsers))parsers <- nondesc.parsers
934    apply.parsers(readLines(f),parsers,verbose=FALSE,...)    apply.parsers(readLines(f),parsers,verbose=FALSE,...)
935  }  }

Legend:
Removed from v.109  
changed lines
  Added in v.204

root@r-forge.r-project.org
ViewVC Help
Powered by ViewVC 1.0.0  
Thanks to:
Vienna University of Economics and Business University of Wisconsin - Madison Powered By FusionForge