SCM

SCM Repository

[tm] Diff of /pkg/R/corpus.R
ViewVC logotype

Diff of /pkg/R/corpus.R

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

trunk/R/textmin/R/textdoccol.R revision 71, Sun Nov 19 17:30:26 2006 UTC pkg/R/corpus.R revision 1114, Fri Nov 26 14:05:54 2010 UTC
# Line 1  Line 1 
1  # Author: Ingo Feinerer  # Author: Ingo Feinerer
2    
3  # The "..." are additional arguments for the function_generator parser  .PCorpus <- function(x, cmeta, dmeta, dbcontrol) {
4  setGeneric("TextDocCol", function(object, parser = plaintext_parser, ...) standardGeneric("TextDocCol"))      attr(x, "CMetaData") <- cmeta
5  setMethod("TextDocCol",      attr(x, "DMetaData") <- dmeta
6            signature(object = "Source"),      attr(x, "DBControl") <- dbcontrol
7            function(object, parser = plaintext_parser, ...) {      class(x) <- c("PCorpus", "Corpus", "list")
8                if (inherits(parser, "function_generator"))      x
9                    parser <- parser(...)  }
10    DBControl <- function(x) attr(x, "DBControl")
11    
12    PCorpus <- function(x,
13                        readerControl = list(reader = x$DefaultReader, language = "en"),
14                        dbControl = list(dbName = "", dbType = "DB1"),
15                        ...) {
16        readerControl <- prepareReader(readerControl, x$DefaultReader, ...)
17    
18        if (is.function(readerControl$init))
19            readerControl$init()
20    
21        if (is.function(readerControl$exit))
22            on.exit(readerControl$exit())
23    
24        if (!filehash::dbCreate(dbControl$dbName, dbControl$dbType))
25            stop("error in creating database")
26        db <- filehash::dbInit(dbControl$dbName, dbControl$dbType)
27    
28        # Allocate memory in advance if length is known
29        tdl <- if (x$Length > 0)
30            vector("list", as.integer(x$Length))
31        else
32            list()
33    
               tdl <- list()  
34                counter <- 1                counter <- 1
35                while (!eoi(object)) {      while (!eoi(x)) {
36                    object <- step_next(object)          x <- stepNext(x)
37                    elem <- get_elem(object)          elem <- getElem(x)
38                    # If there is no Load on Demand support          doc <- readerControl$reader(elem, readerControl$language, if (is.null(x$Names)) as.character(counter) else x$Names[counter])
39                    # we need to load the corpus into memory at startup          filehash::dbInsert(db, ID(doc), doc)
40                    if (object@LoDSupport)          if (x$Length > 0) tdl[[counter]] <- ID(doc)
41                        load <- object@Load          else tdl <- c(tdl, ID(doc))
                   else  
                       load <- TRUE  
                   tdl <- c(tdl, list(parser(elem, object@LoDSupport, load, as.character(counter))))  
42                    counter <- counter + 1                    counter <- counter + 1
43                }                }
44        names(tdl) <- x$Names
45    
46                dmeta.df <- data.frame(MetaID = rep(0, length(tdl)))      df <- data.frame(MetaID = rep(0, length(tdl)), stringsAsFactors = FALSE)
47                dcmeta.node <- new("MetaDataNode",      filehash::dbInsert(db, "DMetaData", df)
48                              NodeID = 0,      dmeta.df <- data.frame(key = "DMetaData", subset = I(list(NA)))
                             MetaData = list(create_date = date(), creator = Sys.getenv("LOGNAME")),  
                             children = list())  
   
               return(new("TextDocCol", .Data = tdl, DMetaData = dmeta.df, DCMetaData = dcmeta.node))  
           })  
   
 setGeneric("DirSource", function(directory, load = FALSE) standardGeneric("DirSource"))  
 setMethod("DirSource",  
           signature(directory = "character"),  
           function(directory, load = FALSE) {  
               new("DirSource", LoDSupport = TRUE, FileList = dir(directory, full.names = TRUE),  
                   Position = 0, Load = load)  
           })  
   
 setGeneric("CSVSource", function(object) standardGeneric("CSVSource"))  
 setMethod("CSVSource",  
           signature(object = "character"),  
           function(object) {  
               object <- substitute(file(object))  
               con <- eval(object)  
               content <- scan(con, what = "character")  
               close(con)  
               new("CSVSource", LoDSupport = FALSE, URI = object,  
                   Content = content, Position = 0)  
           })  
 setMethod("CSVSource",  
           signature(object = "ANY"),  
           function(object) {  
               object <- substitute(object)  
               con <- eval(object)  
               content <- scan(con, what = "character")  
               close(con)  
               new("CSVSource", LoDSupport = FALSE, URI = object,  
                   Content = content, Position = 0)  
           })  
   
 setGeneric("ReutersSource", function(object) standardGeneric("ReutersSource"))  
 setMethod("ReutersSource",  
           signature(object = "character"),  
           function(object) {  
               object <- substitute(file(object))  
               con <- eval(object)  
               corpus <- paste(readLines(con), "\n", collapse = "")  
               close(con)  
               tree <- xmlTreeParse(corpus, asText = TRUE)  
               content <- xmlRoot(tree)$children  
   
               new("ReutersSource", LoDSupport = FALSE, URI = object,  
                   Content = content, Position = 0)  
           })  
 setMethod("ReutersSource",  
           signature(object = "ANY"),  
           function(object) {  
               object <- substitute(object)  
               con <- eval(object)  
               corpus <- paste(readLines(con), "\n", collapse = "")  
               close(con)  
               tree <- xmlTreeParse(corpus, asText = TRUE)  
               content <- xmlRoot(tree)$children  
   
               new("ReutersSource", LoDSupport = FALSE, URI = object,  
                   Content = content, Position = 0)  
           })  
   
 setGeneric("step_next", function(object) standardGeneric("step_next"))  
 setMethod("step_next",  
           signature(object = "DirSource"),  
           function(object) {  
               object@Position <- object@Position + 1  
               object  
           })  
 setMethod("step_next",  
           signature(object = "CSVSource"),  
           function(object) {  
               object@Position <- object@Position + 1  
               object  
           })  
 setMethod("step_next",  
           signature(object = "ReutersSource"),  
           function(object) {  
               object@Position <- object@Position + 1  
               object  
           })  
   
 setGeneric("get_elem", function(object) standardGeneric("get_elem"))  
 setMethod("get_elem",  
           signature(object = "DirSource"),  
           function(object) {  
               filename <- object@FileList[object@Position]  
               list(content = readLines(object@FileList[object@Position]),  
                    uri = substitute(file(filename)))  
           })  
 setMethod("get_elem",  
           signature(object = "CSVSource"),  
           function(object) {  
               list(content = object@Content[object@Position],  
                    uri = object@URI)  
           })  
 setMethod("get_elem",  
           signature(object = "ReutersSource"),  
           function(object) {  
               # Construct a character representation from the XMLNode  
               con <- textConnection("virtual.file", "w")  
               saveXML(object@Content[[object@Position]], con)  
               close(con)  
   
               list(content = virtual.file, uri = object@URI)  
           })  
   
 setGeneric("eoi", function(object) standardGeneric("eoi"))  
 setMethod("eoi",  
           signature(object = "DirSource"),  
           function(object) {  
               if (length(object@FileList) <= object@Position)  
                   return(TRUE)  
               else  
                   return(FALSE)  
           })  
 setMethod("eoi",  
           signature(object = "CSVSource"),  
           function(object) {  
               if (length(object@Content) <= object@Position)  
                   return(TRUE)  
               else  
                   return(FALSE)  
           })  
 setMethod("eoi",  
           signature(object = "ReutersSource"),  
           function(object) {  
               if (length(object@Content) <= object@Position)  
                   return(TRUE)  
               else  
                   return(FALSE)  
           })  
49    
50  plaintext_parser <- function(...) {      .PCorpus(tdl, .MetaDataNode(), dmeta.df, dbControl)
     function(elem, lodsupport, load, id) {  
         if (!lodsupport || (lodsupport && load)) {  
             doc <- new("PlainTextDocument", .Data = elem$content, URI = elem$uri, Cached = TRUE,  
                        Author = "", DateTimeStamp = date(), Description = "", ID = id, Origin = "", Heading = "")  
         }  
         else {  
             doc <- new("PlainTextDocument", URI = elem$uri, Cached = FALSE,  
                        Author = "", DateTimeStamp = date(), Description = "", ID = id, Origin = "", Heading = "")  
51          }          }
52    
53          return(doc)  .VCorpus <- function(x, cmeta, dmeta) {
54      }      attr(x, "CMetaData") <- cmeta
55        attr(x, "DMetaData") <- dmeta
56        class(x) <- c("VCorpus", "Corpus", "list")
57        x
58  }  }
 class(plaintext_parser) <- "function_generator"  
   
 reut21578xml_parser <- function(...) {  
     function(elem, lodsupport, load, id) {  
         corpus <- paste(elem$content, "\n", collapse = "")  
         tree <- xmlTreeParse(corpus, asText = TRUE)  
         node <- xmlRoot(tree)  
   
         # Mask as list to bypass S4 checks  
         class(tree) <- "list"  
59    
60          # The <AUTHOR></AUTHOR> tag is unfortunately NOT obligatory!  # Register S3 corpus classes to be recognized by S4 methods. This is
61          if (!is.null(node[["TEXT"]][["AUTHOR"]]))  # mainly a fix to be compatible with packages which were originally
62              author <- xmlValue(node[["TEXT"]][["AUTHOR"]])  # developed to cooperate with corresponding S4 tm classes. Necessary
63          else  # since tm's class architecture was changed to S3 since tm version 0.5.
64              author <- ""  setOldClass(c("VCorpus", "Corpus", "list"))
   
         datetimestamp <- xmlValue(node[["DATE"]])  
         description <- ""  
         id <- xmlAttrs(node)[["NEWID"]]  
   
         # The <TITLE></TITLE> tag is unfortunately NOT obligatory!  
         if (!is.null(node[["TEXT"]][["TITLE"]]))  
             heading <- xmlValue(node[["TEXT"]][["TITLE"]])  
         else  
             heading <- ""  
   
         topics <- unlist(xmlApply(node[["TOPICS"]], function(x) xmlValue(x)), use.names = FALSE)  
   
         if (!lodsupport || (lodsupport && load)) {  
             doc <- new("XMLTextDocument", .Data = tree, URI = elem$uri, Cached = TRUE, Author = author,  
                        DateTimeStamp = datetimestamp, Description = "", ID = id, Origin = "Reuters-21578 XML",  
                        Heading = heading, LocalMetaData = list(Topics = topics))  
         } else {  
             doc <- new("XMLTextDocument", URI = elem$uri, Cached = FALSE, Author = author,  
                        DateTimeStamp = datetimestamp, Description = "", ID = id, Origin = "Reuters-21578 XML",  
                        Heading = heading, LocalMetaData = list(Topics = topics))  
         }  
65    
66          return(doc)  # The "..." are additional arguments for the FunctionGenerator reader
67      }  VCorpus <- Corpus <- function(x,
68  }                                readerControl = list(reader = x$DefaultReader, language = "en"),
69  class(reut21578xml_parser) <- "function_generator"                                ...) {
70        readerControl <- prepareReader(readerControl, x$DefaultReader, ...)
71    
72  rcv1_parser <- function(...) {      if (is.function(readerControl$init))
73      function(elem, lodsupport, load, id) {          readerControl$init()
         corpus <- paste(elem$content, "\n", collapse = "")  
         tree <- xmlTreeParse(corpus, asText = TRUE)  
         node <- xmlRoot(tree)  
74    
75          # Mask as list to bypass S4 checks      if (is.function(readerControl$exit))
76          class(tree) <- "list"          on.exit(readerControl$exit())
77    
78          datetimestamp <- xmlAttrs(node)[["date"]]      # Allocate memory in advance if length is known
79          id <- xmlAttrs(node)[["itemid"]]      tdl <- if (x$Length > 0)
80          heading <- xmlValue(node[["title"]])          vector("list", as.integer(x$Length))
81        else
82            list()
83    
84          if (!lodsupport || (lodsupport && load)) {      if (x$Vectorized)
85              doc <- new("XMLTextDocument", .Data = tree, URI = elem$uri, Cached = TRUE, Author = "",          tdl <- mapply(function(x, id) readerControl$reader(x, readerControl$language, id),
86                         DateTimeStamp = datetimestamp, Description = "", ID = id, Origin = "Reuters Corpus Volume 1 XML",                        pGetElem(x),
87                         Heading = heading)                        id = if (is.null(x$Names)) as.character(seq_len(x$Length)) else x$Names,
88          } else {                        SIMPLIFY = FALSE)
89              doc <- new("XMLTextDocument", URI = elem$uri, Cached = FALSE, Author = "",      else {
90                         DateTimeStamp = datetimestamp, Description = "", ID = id, Origin = "Reuters Corpus Volume 1 XML",          counter <- 1
91                         Heading = heading)          while (!eoi(x)) {
92                x <- stepNext(x)
93                elem <- getElem(x)
94                doc <- readerControl$reader(elem, readerControl$language, if (is.null(x$Names)) as.character(counter) else x$Names[counter])
95                if (x$Length > 0)
96                    tdl[[counter]] <- doc
97                else
98                    tdl <- c(tdl, list(doc))
99                counter <- counter + 1
100          }          }
   
         return(doc)  
101      }      }
102        names(tdl) <- x$Names
103        df <- data.frame(MetaID = rep(0, length(tdl)), stringsAsFactors = FALSE)
104        .VCorpus(tdl, .MetaDataNode(), df)
105  }  }
 class(rcv1_parser) <- "function_generator"  
   
 newsgroup_parser <- function(...) {  
     function(elem, lodsupport, load, id) {  
         mail <- elem$content  
         author <- gsub("From: ", "", grep("^From:", mail, value = TRUE))  
         datetimestamp <- gsub("Date: ", "", grep("^Date:", mail, value = TRUE))  
         origin <- gsub("Path: ", "", grep("^Path:", mail, value = TRUE))  
         heading <- gsub("Subject: ", "", grep("^Subject:", mail, value = TRUE))  
         newsgroup <- gsub("Newsgroups: ", "", grep("^Newsgroups:", mail, value = TRUE))  
106    
107          if (!lodsupport || (lodsupport && load)) {  `[.PCorpus` <- function(x, i) {
108              # The header is separated from the body by a blank line.      if (missing(i)) return(x)
109              # Reference: \url{http://en.wikipedia.org/wiki/E-mail#Internet_e-mail_format}      index <- attr(x, "DMetaData")[[1 , "subset"]]
110              for (index in seq(along = mail)) {      attr(x, "DMetaData")[[1 , "subset"]] <- if (is.numeric(index)) index[i] else i
111                  if (mail[index] == "")      dmeta <- attr(x, "DMetaData")
112                      break      .PCorpus(NextMethod("["), CMetaData(x), dmeta, DBControl(x))
113              }              }
             content <- mail[(index + 1):length(mail)]  
114    
115              doc <- new("NewsgroupDocument", .Data = content, URI = elem$uri, Cached = TRUE,  `[.VCorpus` <- function(x, i) {
116                         Author = author, DateTimeStamp = datetimestamp,      if (missing(i)) return(x)
117                         Description = "", ID = id, Origin = origin,      .VCorpus(NextMethod("["), CMetaData(x), DMetaData(x)[i, , drop = FALSE])
                        Heading = heading, Newsgroup = newsgroup)  
         } else {  
             doc <- new("NewsgroupDocument", URI = elem$uri, Cached = FALSE, Author = author, DateTimeStamp = datetimestamp,  
                        Description = "", ID = id, Origin = origin, Heading = heading, Newsgroup = newsgroup)  
118          }          }
119    
120          return(doc)  `[<-.PCorpus` <- function(x, i, value) {
121      }      db <- filehash::dbInit(DBControl(x)[["dbName"]], DBControl(x)[["dbType"]])
122        counter <- 1
123        for (id in unclass(x)[i]) {
124            if (identical(length(value), 1L)) db[[id]] <- value
125            else db[[id]] <- value[[counter]]
126            counter <- counter + 1
127  }  }
128  class(newsgroup_parser) <- "function_generator"      x
   
 # Parse a <newsitem></newsitem> element from a well-formed RCV1 XML file  
 rcv1_to_plain <- function(node, ...) {  
     datetimestamp <- xmlAttrs(node)[["date"]]  
     id <- xmlAttrs(node)[["itemid"]]  
     origin <- "Reuters Corpus Volume 1 XML"  
     corpus <- unlist(xmlApply(node[["text"]], xmlValue), use.names = FALSE)  
     heading <- xmlValue(node[["title"]])  
   
     new("PlainTextDocument", .Data = corpus, Cached = TRUE, URI = "", Author = "", DateTimeStamp = datetimestamp,  
         Description = "", ID = id, Origin = "Reuters Corpus Volume 1 XML", Heading = heading)  
129  }  }
130    
131  # Parse a <REUTERS></REUTERS> element from a well-formed Reuters-21578 XML file  .map_name_index <- function(x, i) {
132  reut21578xml_to_plain <- function(node, ...) {      if (is.character(i)) {
133      # The <AUTHOR></AUTHOR> tag is unfortunately NOT obligatory!          if (is.null(names(x)))
134      if (!is.null(node[["TEXT"]][["AUTHOR"]]))              match(i, meta(x, "ID", type = "local"))
         author <- xmlValue(node[["TEXT"]][["AUTHOR"]])  
135      else      else
136          author <- ""              match(i, names(x))
137        }
138      datetimestamp <- xmlValue(node[["DATE"]])      i
     description <- ""  
     id <- xmlAttrs(node)[["NEWID"]]  
   
     origin <- "Reuters-21578 XML"  
   
     # The <BODY></BODY> tag is unfortunately NOT obligatory!  
     if (!is.null(node[["TEXT"]][["BODY"]]))  
         corpus <- xmlValue(node[["TEXT"]][["BODY"]])  
     else  
         corpus <- ""  
   
     # The <TITLE></TITLE> tag is unfortunately NOT obligatory!  
     if (!is.null(node[["TEXT"]][["TITLE"]]))  
         heading <- xmlValue(node[["TEXT"]][["TITLE"]])  
     else  
         heading <- ""  
   
     topics <- unlist(xmlApply(node[["TOPICS"]], function(x) xmlValue(x)), use.names = FALSE)  
   
     new("PlainTextDocument", .Data = corpus, Cached = TRUE, URI = "", Author = author, DateTimeStamp = datetimestamp,  
         Description = description, ID = id, Origin = origin, Heading = heading, LocalMetaData = list(Topics = topics))  
139  }  }
140    
141  setGeneric("load_doc", function(object, ...) standardGeneric("load_doc"))  `[[.PCorpus` <-  function(x, i) {
142  setMethod("load_doc",      i <- .map_name_index(x, i)
143            signature(object = "PlainTextDocument"),      db <- filehash::dbInit(DBControl(x)[["dbName"]], DBControl(x)[["dbType"]])
144            function(object, ...) {      filehash::dbFetch(db, NextMethod("[["))
145                if (!Cached(object)) {  }
146                    con <- eval(URI(object))  `[[.VCorpus` <-  function(x, i) {
147                    corpus <- readLines(con)      i <- .map_name_index(x, i)
148                    close(con)      lazyTmMap <- meta(x, tag = "lazyTmMap", type = "corpus")
149                    Corpus(object) <- corpus      if (!is.null(lazyTmMap))
150                    Cached(object) <- TRUE          .Call("copyCorpus", x, materialize(x, i))
151                    return(object)      NextMethod("[[")
152                } else {  }
153                    return(object)  
154                }  `[[<-.PCorpus` <-  function(x, i, value) {
155            })      i <- .map_name_index(x, i)
156  setMethod("load_doc",      db <- filehash::dbInit(DBControl(x)[["dbName"]], DBControl(x)[["dbType"]])
157            signature(object =  "XMLTextDocument"),      index <- unclass(x)[[i]]
158            function(object, ...) {      db[[index]] <- value
159                if (!Cached(object)) {      x
160                    con <- eval(URI(object))  }
161                    corpus <- paste(readLines(con), "\n", collapse = "")  `[[<-.VCorpus` <-  function(x, i, value) {
162                    close(con)      i <- .map_name_index(x, i)
163                    doc <- xmlTreeParse(corpus, asText = TRUE)      # Mark new objects as not active for lazy mapping
164                    class(doc) <- "list"      lazyTmMap <- meta(x, tag = "lazyTmMap", type = "corpus")
165                    Corpus(object) <- doc      if (!is.null(lazyTmMap)) {
166                    Cached(object) <- TRUE          lazyTmMap$index[i] <- FALSE
167                    return(object)          meta(x, tag = "lazyTmMap", type = "corpus") <- lazyTmMap
168                } else {      }
169                    return(object)      # Set the value
170                }      cl <- class(x)
171            })      y <- NextMethod("[[<-")
172  setMethod("load_doc",      class(y) <- cl
173            signature(object = "NewsgroupDocument"),      y
174            function(object, ...) {  }
175                if (!Cached(object)) {  
176                    con <- eval(URI(object))  # Update NodeIDs of a CMetaData tree
177                    mail <- readLines(con)  .update_id <- function(x, id = 0, mapping = NULL, left.mapping = NULL, level = 0) {
178                    close(con)      # Traversal of (binary) CMetaData tree with setup of NodeIDs
179                    Cached(object) <- TRUE      set_id <- function(x) {
180                    for (index in seq(along = mail)) {          x$NodeID <- id
                       if (mail[index] == "")  
                           break  
                   }  
                   Corpus(object) <- mail[(index + 1):length(mail)]  
                   return(object)  
               } else {  
                   return(object)  
               }  
           })  
   
 setGeneric("tm_transform", function(object, FUN, ...) standardGeneric("tm_transform"))  
 setMethod("tm_transform",  
           signature(object = "TextDocCol", FUN = "function"),  
           function(object, FUN, ...) {  
               result <- as(lapply(object, FUN, ..., DMetaData = DMetaData(object)), "TextDocCol")  
               result@DMetaData <- DMetaData(object)  
               return(result)  
           })  
   
 setGeneric("as.plaintext_doc", function(object, FUN, ...) standardGeneric("as.plaintext_doc"))  
 setMethod("as.plaintext_doc",  
           signature(object = "PlainTextDocument"),  
           function(object, FUN, ...) {  
               return(object)  
           })  
 setMethod("as.plaintext_doc",  
           signature(object = "XMLTextDocument", FUN = "function"),  
           function(object, FUN, ...) {  
               if (!Cached(object))  
                   object <- load_doc(object)  
   
               corpus <- Corpus(object)  
   
               # As XMLDocument is no native S4 class, restore valid information  
               class(corpus) <- "XMLDocument"  
               names(corpus) <- c("doc","dtd")  
   
               return(FUN(xmlRoot(corpus), ...))  
           })  
   
 setGeneric("tm_tolower", function(object, ...) standardGeneric("tm_tolower"))  
 setMethod("tm_tolower",  
           signature(object = "PlainTextDocument"),  
           function(object, ...) {  
               if (!Cached(object))  
                   object <- load_doc(object)  
   
               Corpus(object) <- tolower(object)  
               return(object)  
           })  
   
 setGeneric("strip_whitespace", function(object, ...) standardGeneric("strip_whitespace"))  
 setMethod("strip_whitespace",  
           signature(object = "PlainTextDocument"),  
           function(object, ...) {  
               if (!Cached(object))  
                   object <- load_doc(object)  
   
               Corpus(object) <- gsub("[[:space:]]+", " ", object)  
               return(object)  
           })  
   
 setGeneric("stem_doc", function(object, ...) standardGeneric("stem_doc"))  
 setMethod("stem_doc",  
           signature(object = "PlainTextDocument"),  
           function(object, ...) {  
               if (!Cached(object))  
                   object <- load_doc(object)  
   
               require(Rstem)  
               splittedCorpus <- unlist(strsplit(object, " ", fixed = TRUE))  
               stemmedCorpus <- wordStem(splittedCorpus)  
               Corpus(object) <- paste(stemmedCorpus, collapse = " ")  
               return(object)  
           })  
   
 setGeneric("remove_words", function(object, stopwords, ...) standardGeneric("remove_words"))  
 setMethod("remove_words",  
           signature(object = "PlainTextDocument", stopwords = "character"),  
           function(object, stopwords, ...) {  
               if (!Cached(object))  
                   object <- load_doc(object)  
   
               require(Rstem)  
               splittedCorpus <- unlist(strsplit(object, " ", fixed = TRUE))  
               noStopwordsCorpus <- splittedCorpus[!splittedCorpus %in% stopwords]  
               Corpus(object) <- paste(noStopwordsCorpus, collapse = " ")  
               return(object)  
           })  
   
 setGeneric("tm_filter", function(object, ..., FUN = s_filter) standardGeneric("tm_filter"))  
 setMethod("tm_filter",  
           signature(object = "TextDocCol"),  
           function(object, ..., FUN = s_filter) {  
               indices <- sapply(object, FUN, ..., DMetaData = DMetaData(object))  
               object[indices]  
           })  
   
 setGeneric("tm_index", function(object, ..., FUN = s_filter) standardGeneric("tm_index"))  
 setMethod("tm_index",  
           signature(object = "TextDocCol"),  
           function(object, ..., FUN = s_filter) {  
               sapply(object, FUN, ..., DMetaData = DMetaData(object))  
           })  
   
 s_filter <- function(object, s, ..., DMetaData) {  
     b <- TRUE  
     for (tag in names(s)) {  
         if (tag %in% names(LocalMetaData(object))) {  
             b <- b && any(grep(s[[tag]], LocalMetaData(object)[[tag]]))  
         } else if (tag %in% names(DMetaData)){  
             b <- b && any(grep(s[[tag]], DMetaData[[tag]]))  
         } else {  
             b <- b && any(grep(s[[tag]], eval(call(tag, object))))  
         }  
     }  
     return(b)  
 }  
   
 setGeneric("fulltext_search_filter", function(object, pattern, ...) standardGeneric("fulltext_search_filter"))  
 setMethod("fulltext_search_filter",  
           signature(object = "PlainTextDocument", pattern = "character"),  
           function(object, pattern, ...) {  
               if (!Cached(object))  
                   object <- load_doc(object)  
   
               return(any(grep(pattern, Corpus(object))))  
           })  
   
 setGeneric("attach_data", function(object, data) standardGeneric("attach_data"))  
 setGeneric("attach_metadata", function(object, name, metadata) standardGeneric("attach_metadata"))  
   
 setGeneric("append_doc", function(object, data, meta) standardGeneric("append_doc"))  
 setMethod("append_doc",  
           signature(object = "TextDocCol", data = "TextDocument", meta = "list"),  
           function(object, data, meta) {  
               object@.Data <- c(object@.Data, list(data))  
               object@DMetaData <- rbind(object@DMetaData, c(MetaID = DCMetaData(object)@NodeID, meta))  
               return(object)  
           })  
   
 setGeneric("append_meta", function(object, dcmeta, dmeta) standardGeneric("append_meta"))  
 setMethod("append_meta",  
           signature(object = "TextDocCol", dcmeta = "list", dmeta = "list"),  
           function(object, dcmeta, dmeta) {  
               object@DCMetaData@MetaData <- c(object@DCMetaData@MetaData, dcmeta)  
               object@DMetaData <- cbind(object@DMetaData, dmeta)  
               return(object)  
           })  
   
 setGeneric("remove_metadata", function(object, name) standardGeneric("remove_metadata"))  
 #setMethod("remove_metadata",  
 #          signature(object = "TextDocCol"),  
 #          function(object, name) {  
 #              object@DMetaData <- DMetaData(object)[names(DMetaData(object)) != name]  
 #              return(object)  
 #          })  
   
 setGeneric("modify_metadata", function(object, name, metadata) standardGeneric("modify_metadata"))  
 #setMethod("modify_metadata",  
 #          signature(object = "TextDocCol"),  
 #          function(object, name, metadata) {  
 #              object@DMetaData[[name]] <- metadata  
 #              return(object)  
 #          })  
   
 setMethod("[",  
           signature(x = "TextDocCol", i = "ANY", j = "ANY", drop = "ANY"),  
           function(x, i, j, ... , drop) {  
               if(missing(i))  
                   return(x)  
   
               object <- x  
               object@.Data <- x@.Data[i, ..., drop = FALSE]  
               object@DMetaData <- DMetaData(object)[i, ]  
               return(object)  
           })  
   
 setMethod("[<-",  
           signature(x = "TextDocCol", i = "ANY", j = "ANY", value = "ANY"),  
           function(x, i, j, ... , value) {  
               object <- x  
               object@.Data[i, ...] <- value  
               return(object)  
           })  
   
 setMethod("[[",  
           signature(x = "TextDocCol", i = "ANY", j = "ANY"),  
           function(x, i, j, ...) {  
               return(x@.Data[[i, ...]])  
           })  
   
 setMethod("[[<-",  
           signature(x = "TextDocCol", i = "ANY", j = "ANY", value = "ANY"),  
           function(x, i, j, ..., value) {  
               object <- x  
               object@.Data[[i, ...]] <- value  
               return(object)  
           })  
   
 # Update \code{NodeID}s of a DCMetaData tree  
 # TODO: Avoid global variables outside of update_id function  
 update_id <- function(object) {  
     id <<- 0  
     mapping <<- left.mapping <<- NULL  
     level <<- 0  
     return(list(root = set_id(object), left.mapping = left.mapping, right.mapping = mapping))  
 }  
   
 # Traversal of (binary) DCMetaData tree with setup of \code{NodeID}s  
 set_id <- function(object) {  
     object@NodeID <- id  
181      id <<- id + 1      id <<- id + 1
182      level <<- level + 1      level <<- level + 1
183            if (length(x$Children) > 0) {
184      if (length(object@children) > 0) {              mapping <<- cbind(mapping, c(x$Children[[1]]$NodeID, id))
185          mapping <<- cbind(mapping, c(object@children[[1]]@NodeID, id))              left <- set_id(x$Children[[1]])
         left <- set_id(object@children[[1]])  
186          if (level == 1) {          if (level == 1) {
187              left.mapping <<- mapping              left.mapping <<- mapping
188              mapping <<- NULL              mapping <<- NULL
189          }          }
190          mapping <<- cbind(mapping, c(object@children[[2]]@NodeID, id))              mapping <<- cbind(mapping, c(x$Children[[2]]$NodeID, id))
191          right <- set_id(object@children[[2]])              right <- set_id(x$Children[[2]])
192    
193          object@children <- list(left, right)              x$Children <- list(left, right)
194      }      }
195      level <<- level - 1      level <<- level - 1
196            x
197      return(object)      }
198        list(root = set_id(x), left.mapping = left.mapping, right.mapping = mapping)
199  }  }
200    
201  setMethod("c",  # Find indices to be updated for a CMetaData tree
202            signature(x = "TextDocCol"),  .find_indices <- function(x) {
           function(x, y, ..., meta = list(merge_date = date(), merger = Sys.getenv("LOGNAME")), recursive = TRUE) {  
               if (!inherits(y, "TextDocCol"))  
                   stop("invalid argument")  
   
               object <- x  
               # Concatenate data slots  
               object@.Data <- c(as(x, "list"), as(y, "list"))  
   
               # Update the DCMetaData tree  
               dcmeta <- new("MetaDataNode", NodeID = 0, MetaData = meta, children = list(DCMetaData(x), DCMetaData(y)))  
               update.struct <- update_id(dcmeta)  
               object@DCMetaData <- update.struct$root  
   
               # Find indices to be updated for the left tree  
203                indices.mapping <- NULL                indices.mapping <- NULL
204                for (m in levels(as.factor(DMetaData(x)$MetaID))) {                for (m in levels(as.factor(DMetaData(x)$MetaID))) {
205                    indices <- (DMetaData(x)$MetaID == m)                    indices <- (DMetaData(x)$MetaID == m)
206                    indices.mapping <- c(indices.mapping, list(m = indices))                    indices.mapping <- c(indices.mapping, list(m = indices))
207                    names(indices.mapping)[length(indices.mapping)] <- m                    names(indices.mapping)[length(indices.mapping)] <- m
208                }                }
209        indices.mapping
210    }
211    
212    c2 <- function(x, y, ...) {
213        # Update the CMetaData tree
214        cmeta <- .MetaDataNode(0, list(merge_date = as.POSIXlt(Sys.time(), tz = "GMT"), merger = Sys.getenv("LOGNAME")), list(CMetaData(x), CMetaData(y)))
215        update.struct <- .update_id(cmeta)
216    
217        new <- .VCorpus(c(unclass(x), unclass(y)), update.struct$root, NULL)
218    
219        # Find indices to be updated for the left tree
220        indices.mapping <- .find_indices(x)
221    
222                # Update the DMetaData data frames for the left tree                # Update the DMetaData data frames for the left tree
223                for (i in 1:ncol(update.struct$left.mapping)) {                for (i in 1:ncol(update.struct$left.mapping)) {
224                    map <- update.struct$left.mapping[,i]                    map <- update.struct$left.mapping[,i]
225                    x@DMetaData$MetaID <- replace(DMetaData(x)$MetaID, indices.mapping[[as.character(map[1])]], map[2])          DMetaData(x)$MetaID <- replace(DMetaData(x)$MetaID, indices.mapping[[as.character(map[1])]], map[2])
226                }                }
227    
228                # Find indices to be updated for the right tree                # Find indices to be updated for the right tree
229                indices.mapping <- NULL      indices.mapping <- .find_indices(y)
               for (m in levels(as.factor(DMetaData(y)$MetaID))) {  
                   indices <- (DMetaData(y)$MetaID == m)  
                   indices.mapping <- c(indices.mapping, list(m = indices))  
                   names(indices.mapping)[length(indices.mapping)] <- m  
               }  
230    
231                # Update the DMetaData data frames for the right tree                # Update the DMetaData data frames for the right tree
232                for (i in 1:ncol(update.struct$right.mapping)) {                for (i in 1:ncol(update.struct$right.mapping)) {
233                    map <- update.struct$right.mapping[,i]                    map <- update.struct$right.mapping[,i]
234                    y@DMetaData$MetaID <- replace(DMetaData(y)$MetaID, indices.mapping[[as.character(map[1])]], map[2])          DMetaData(y)$MetaID <- replace(DMetaData(y)$MetaID, indices.mapping[[as.character(map[1])]], map[2])
235                }                }
236    
237                # Merge the DMetaData data frames                # Merge the DMetaData data frames
# Line 645  Line 241 
241                labels <- setdiff(names(DMetaData(x)), names(DMetaData(y)))                labels <- setdiff(names(DMetaData(x)), names(DMetaData(y)))
242                na.matrix <- matrix(NA, nrow = nrow(DMetaData(y)), ncol = length(labels), dimnames = list(row.names(DMetaData(y)), labels))                na.matrix <- matrix(NA, nrow = nrow(DMetaData(y)), ncol = length(labels), dimnames = list(row.names(DMetaData(y)), labels))
243                y.dmeta.aug <- cbind(DMetaData(y), na.matrix)                y.dmeta.aug <- cbind(DMetaData(y), na.matrix)
244                object@DMetaData <- rbind(x.dmeta.aug, y.dmeta.aug)      DMetaData(new) <- rbind(x.dmeta.aug, y.dmeta.aug)
245    
246        new
247    }
248    
249    c.Corpus <-
250    function(x, ..., recursive = FALSE)
251    {
252        args <- list(...)
253    
254        if (identical(length(args), 0L))
255            return(x)
256    
257                return(object)      if (!all(unlist(lapply(args, inherits, class(x)))))
258      })          stop("not all arguments are of the same corpus type")
259  #setMethod("c",  
260  #          signature(x = "TextDocument"),      if (inherits(x, "PCorpus"))
261  #          function(x, ..., recursive = TRUE){          stop("concatenation of corpora with underlying databases is not supported")
262  #              args <- list(...)  
263  #              if(length(args) == 0)      l <- base::c(list(x), args)
264  #                  return(x)      if (recursive)
265  #              return(new("TextDocCol", .Data = list(x, ...)))          Reduce(c2, l)
266  #    })      else {
267            l <- do.call("c", lapply(l, unclass))
268  setMethod("length",          .VCorpus(l,
269            signature(x = "TextDocCol"),                   cmeta = .MetaDataNode(),
270            function(x){                   dmeta = data.frame(MetaID = rep(0, length(l)), stringsAsFactors = FALSE))
271                return(length(as(x, "list")))      }
272      })  }
273    
274  setMethod("show",  c.TextDocument <- function(x, ..., recursive = FALSE) {
275            signature(object = "TextDocCol"),      args <- list(...)
276            function(object){  
277                cat(sprintf(ngettext(length(object),      if (identical(length(args), 0L))
278                                     "A text document collection with %d text document\n",          return(x)
279                                     "A text document collection with %d text documents\n"),  
280                            length(object)))      if (!all(unlist(lapply(args, inherits, class(x)))))
281      })          stop("not all arguments are text documents")
282    
283  setMethod("summary",      dmeta <- data.frame(MetaID = rep(0, length(list(x, ...))), stringsAsFactors = FALSE)
284            signature(object = "TextDocCol"),      .VCorpus(list(x, ...), .MetaDataNode(), dmeta)
285            function(object){  }
286                show(object)  
287    print.Corpus <- function(x, ...) {
288        cat(sprintf(ngettext(length(x),
289                             "A corpus with %d text document\n",
290                             "A corpus with %d text documents\n"),
291                    length(x)))
292        invisible(x)
293    }
294    
295    summary.Corpus <- function(object, ...) {
296        print(object)
297                if (length(DMetaData(object)) > 0) {                if (length(DMetaData(object)) > 0) {
298                    cat(sprintf(ngettext(length(DMetaData(object)),          cat(sprintf(ngettext(length(attr(CMetaData(object), "MetaData")),
299                                                "\nThe global metadata consists of %d tag-value pair\n",                               "\nThe metadata consists of %d tag-value pair and a data frame\n",
300                                                "\nThe global metadata consists of %d tag-value pairs\n"),                               "\nThe metadata consists of %d tag-value pairs and a data frame\n"),
301                                         length(DMetaData(object))))                      length(CMetaData(object)$MetaData)))
302                    cat("Available tags are:\n")                    cat("Available tags are:\n")
303                    cat(names(DMetaData(object)), "\n")          cat(strwrap(paste(names(CMetaData(object)$MetaData), collapse = " "), indent = 2, exdent = 2), "\n")
304            cat("Available variables in the data frame are:\n")
305            cat(strwrap(paste(names(DMetaData(object)), collapse = " "), indent = 2, exdent = 2), "\n")
306        }
307                }                }
     })  
308    
309  setGeneric("inspect", function(object) standardGeneric("inspect"))  inspect <- function(x) UseMethod("inspect", x)
310  setMethod("inspect",  inspect.PCorpus <- function(x) {
311            signature("TextDocCol"),      summary(x)
           function(object) {  
               summary(object)  
312                cat("\n")                cat("\n")
313                show(as(object, "list"))      db <- filehash::dbInit(DBControl(x)[["dbName"]], DBControl(x)[["dbType"]])
314            })      show(filehash::dbMultiFetch(db, unlist(x)))
315    }
316    inspect.VCorpus <- function(x) {
317        summary(x)
318        cat("\n")
319        print(noquote(lapply(x, identity)))
320    }
321    
322  # No metadata is checked  lapply.PCorpus <- function(X, FUN, ...) {
323  setGeneric("%IN%", function(x, y) standardGeneric("%IN%"))      db <- filehash::dbInit(DBControl(X)[["dbName"]], DBControl(X)[["dbType"]])
324  setMethod("%IN%",      lapply(filehash::dbMultiFetch(db, unlist(X)), FUN, ...)
325            signature(x = "TextDocument", y = "TextDocCol"),  }
326            function(x, y) {  lapply.VCorpus <- function(X, FUN, ...) {
327                x %in% y      lazyTmMap <- meta(X, tag = "lazyTmMap", type = "corpus")
328            })      if (!is.null(lazyTmMap))
329            .Call("copyCorpus", X, materialize(X))
330        base::lapply(X, FUN, ...)
331    }
332    
333    writeCorpus <-  function(x, path = ".", filenames = NULL) {
334        filenames <- file.path(path,
335                               if (is.null(filenames)) unlist(lapply(x, function(x) sprintf("%s.txt", ID(x))))
336                               else filenames)
337        i <- 1
338        for (o in x) {
339            writeLines(as.PlainTextDocument(o), filenames[i])
340            i <- i + 1
341        }
342    }

Legend:
Removed from v.71  
changed lines
  Added in v.1114

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