* using log directory 'd:/Rcompile/CRANpkg/local/3.5/diffrprojects.Rcheck' * using R version 3.5.3 (2019-03-11) * using platform: x86_64-w64-mingw32 (64-bit) * using session charset: ISO8859-1 * checking for file 'diffrprojects/DESCRIPTION' ... OK * this is package 'diffrprojects' version '0.1.14' * checking package namespace information ... OK * checking package dependencies ... OK * checking if this is a source package ... OK * checking if there is a namespace ... OK * checking for hidden files and directories ... OK * checking for portable file names ... OK * checking whether package 'diffrprojects' can be installed ... OK * checking installed package size ... OK * checking package directory ... OK * checking DESCRIPTION meta-information ... OK * checking top-level files ... OK * checking for left-over files ... OK * checking index information ... OK * checking package subdirectories ... OK * checking R files for non-ASCII characters ... OK * checking R files for syntax errors ... OK * loading checks for arch 'i386' ** checking whether the package can be loaded ... OK ** checking whether the package can be loaded with stated dependencies ... OK ** checking whether the package can be unloaded cleanly ... OK ** checking whether the namespace can be loaded with stated dependencies ... OK ** checking whether the namespace can be unloaded cleanly ... OK ** checking loading without being on the library search path ... OK ** checking use of S3 registration ... OK * loading checks for arch 'x64' ** checking whether the package can be loaded ... OK ** checking whether the package can be loaded with stated dependencies ... OK ** checking whether the package can be unloaded cleanly ... OK ** checking whether the namespace can be loaded with stated dependencies ... OK ** checking whether the namespace can be unloaded cleanly ... OK ** checking loading without being on the library search path ... OK ** checking use of S3 registration ... OK * checking dependencies in R code ... OK * checking S3 generic/method consistency ... OK * checking replacement functions ... OK * checking foreign function calls ... OK * checking R code for possible problems ... [6s] OK * checking Rd files ... OK * checking Rd metadata ... OK * checking Rd cross-references ... OK * checking for missing documentation entries ... OK * checking for code/documentation mismatches ... OK * checking Rd \usage sections ... OK * checking Rd contents ... OK * checking for unstated dependencies in examples ... OK * checking contents of 'data' directory ... OK * checking data for non-ASCII characters ... OK * checking data for ASCII and uncompressed saves ... OK * checking line endings in C/C++/Fortran sources/headers ... OK * checking pragmas in C/C++ headers and code ... OK * checking compiled code ... OK * checking examples ... ** running examples for arch 'i386' ... ERROR Running examples in 'diffrprojects-Ex.R' failed The error most likely occurred in: > ### Name: diffrproject > ### Title: class for diffrproject > ### Aliases: diffrproject > ### Keywords: data > > ### ** Examples > > > > > > ## Creating a Diffrprojects Instance > > # To create a diffrproject we use the diffrproject creator object - > # its simply an object with an function that knows how to create a project. > > # Creating a project looks like this: > > library(diffrprojects) > dp <- diffrproject$new() > > > # Et viola - we created a first, for now empty, project that we will > # use throughout the tutorial. > > > > > > ## Some Help Please > > # To get a better idea about what this thing called *diffrproject* really is > # you can consult its help page which gives a broad overview over its > # capabilities: > > ?diffrproject diffrproject package:diffrprojects R Documentation _c_l_a_s_s _f_o_r _d_i_f_f_r_p_r_o_j_e_c_t _D_e_s_c_r_i_p_t_i_o_n: class for diffrproject _U_s_a_g_e: diffrproject _F_o_r_m_a_t: 'R6Class' creator object. _V_a_l_u_e: Object of 'diffrproject' _T_h_e _d_i_f_f_r_p_r_o_j_e_c_t_s _c_l_a_s_s _f_a_m_i_l_y: Diffrproject consists of an set of R6 classes that are conencted by inheritance. Each class handles a different set of functionalities that are modular. R6_rtext_extended A class that has nothing to do per se with diffrprojects. It merely adds some basic features to the base R6 class (debugging, hashing, getting fields and handling warnings and messages as well as listing content). This class is imported from rtext package dp_base [inherits from rtext::R6_rtext_extended] This class forms the foundation of all diffrpojects (dp_xxx) classes by implementing data fields for meta data, texts, data on texts, links between texts, alignment of text tokens, and data on the alignment of text tokens. Furthermore it implements methods add, delete, code, and link texts or to aggregate text data on text token level. dp_loadsave [inherits from dp_base] This class allows for loading and saving diffrprojects from and to Rdata files. dp_export [inherits from dp_loadsave] This class provides methods for exporting and importing to and from RSQLite. dp_align [inherits from dp_export] This is one of the workhorses of diffrprojects. The methods of this class allow for adding, deleting or computing alignments between text tokens (e.g. words or lines or sentences or characters or paragraphs, or some other way to split text into chunks). Furthermore it allows to also assign data to individual alignments (a connection beween two token of text from different text versions). dp_inherit [inherits from dp_align] The text_data_inherit method added by this class allows to copy text data from one token of a text version to another token of another text version channeled through aligments with zero distance. Conflicting codings (a text might have multiple codings stemming from several links and from direct coding of the text) are resolved by the fact that text codings are accompanied by a hierarchy level that defaults to zero and gets decreased by one every time the coding is inherited by a token. diffrproject [inherits from dp_inherit] Just a wrapper inheriting from dp_inherit to have a less technical name at the end of the inheritance chain. _E_x_a_m_p_l_e_s: ## Creating a Diffrprojects Instance # To create a diffrproject we use the diffrproject creator object - # its simply an object with an function that knows how to create a project. # Creating a project looks like this: library(diffrprojects) dp <- diffrproject$new() # Et viola - we created a first, for now empty, project that we will # use throughout the tutorial. ## Some Help Please # To get a better idea about what this thing called *diffrproject* really is # you can consult its help page which gives a broad overview over its # capabilities: ?diffrproject # Another way is to call the ls() method. This will present us with a # data frame listing all fields where data is stored and all the methods # (aka object specific functions) of our diffrprojects instance. # Those methods and fields located in *private* are not for the user # to mess around with while non-private (*self* aka public) data fields # can be read by the user and public methods can be triggered by the # user to manipulate the data or retrieve data in a specific format. dp$ls() # The base R class() function furthermore reveals from which classes the # diffrproject class inherits: class(dp) ## Adding Texts to Projects # Our diffrproject (`dp`) has one method called `text_add()` that allows to # add texts to the project. Basically the method can be used in three # different flavors: adding character vectors, adding texts stored on disk, # or by adding rtext objects (see rtext package: # https://CRAN.R-project.org/package=rtext; rtext objects are the way # individual texts are represented within diffrprojects). # For each of these used cases there is one option: # `text`, `text_file`, `rtext`; respectively. # Below are shown examples using each of these methods: # **adding text files** test_file1 <- stringb:::test_file("rc_1_ch1.txt") test_file2 <- stringb:::test_file("rc_2_ch1.txt") dp$text_add(text_file = c(test_file1, test_file2) ) # **adding rtext objects** test_file <- stringb:::test_file("rc_1_ch1.txt") rt <- rtext$new( text_file = test_file) dp$text_add(rtext = rt) # **adding character vectors** test_file1 <- stringb:::test_file("rc_1_ch1.txt") test_file2 <- stringb:::test_file("rc_2_ch1.txt") cv <- "" cv[1] <- text_read(test_file1, NULL) cv[2] <- text_read(test_file2, NULL) dp$text_add(text = cv) # In the last case make sure to put each text in one separate line. # Functions like readLines() or text_read() read in texts such that # each line corresponds to one element in a character vector. With e.g. # text_read()'s tokenize parameter to NULL the text will be read in as one # long string. ## Piping Methods # Now is a good time to mention a feature of diffrprojects that comes in # handy: All functions that do not explicitly extract data # (those usually have some 'get' as part of their name) do return the # object itself so that one can pipe together a series of method calls. # Consider the following example where we initiate a new diffrprojects # instance and add two texts in just one pipe: dp <- diffrproject$ new()$ text_add(text_version_1, name = "version1")$ text_add(text_version_2, name = "version2") length(dp$text) ## Getting Infos About Texts # If we want to get some general overview about the texts gathered in our # project, we can use the text_meta_data() method to do so. # The method has no parameters and returns a data.frame with several # variables informing us about its source, length, encoding used for # storage, and its name. dp$text_meta_data() ## Showing Text # If you want to have a look at your texts you may do so by using the # text's own text_show methods. Per default this method only shows the # first 500 characters, but it can be set to higher numbers as well. dp$text$version1$text_show(length=1000) dp$text$version2$text_show(length=1000) ## Getting And Setting Infos About the Project # Similar to the text_meta_data() method we can access the projects # meta data via data fields meta and options. But contrary to the # text_meta_data() method that gathers data from all the texts within the # project and does not allow for manipulation of the data, the data # fields allow reading and writing. # First let us have a look and thereafter turn off the message # notification service: # **getting data fields** dp$options # **setting data fields** dp$options$verbose <- FALSE # (note, ask is deprecated and only remains for compatibility # reasons but has no function anymore) # Now it's time to have a look at the projects meta data. # It tells us when the project was created, which path to use for # SQLite exports, which path to use for saving data as in RData # format and what is the projects id. The id is a hash of a time stamp # as well as session information which should ensure uniqueness across # space and time. # All these values can be manipulated by the user to her liking. dp$meta dp$meta$file_path = "./diffrproject.RData" ## Deleting Texts # Of cause we can not only add texts but delete them from the project as # well. For this purpose there is the text_delete() method. # Let's just add two texts and delete one by providing its index number and # the second by providing its name to the text_delete() method. dp$text_add(text = "nonesense", "n1") dp$text_add(text = "nonesense", "n2") dp$text_delete(3) dp$text_delete("n2") length(dp$text) names(dp$text) ## Defining Relationships Between Texts: Linking # The purpose of diffrprojects is to enable data collection on the # difference of texts. Having filled a project with various texts, # there are endless possibilities to form pairs of text for comparison # and change measurement - where endless actually is equal to: $n^2-n$. # Linking can be done via the text_link method which accepts either # index numbers or text names for its from and to arguments # (a third argument delete will delete a specified link if set to TRUE). dp$text_link(from = 1, to = 2) dp$text_link(from = 1, to = 2, delete = TRUE) # If no arguments are specified, text_link will link the first text to # the second, the third to the fourth, the fourth to the fifths and so on. dp$text_link() # To get an idea of what links are currently specified, we can # directly access the link data field or/and ask R to transform the # list found there into a data.frame. dp$link dp$link %>% as.data.frame() ## Aligning Texts and Measuring Change # At the heart of each diffrproject lies the text_align method. # This method compares two texts and tries to align parts # of one text with parts of the other text. The first two # arguments (`t1` and `t2`) are for specifying which pair # of texts to compare - if left as-is, all text pairs that # are specified within the link data field will be aligned. # Text parts are arbitrary character spans defined by the # `tokenizer` argument. This argument expects a function splitting # text into a token data.frame. If the tokenizer argument # is left as-is, it will default to text_tokenize_lines function # from the stringb package. # Text tokens can be pre-processed before alignment. The `clean` # argument allows to hand over a function tranforming a charactr # vector of text tokens into their clean counterparts. # The `ignore` arguments expects a function that is able to # transform a character vector of tokens into a logical vector # of same length, indicating which tokens to ignore throughout # the alignment process and which to consider. # The next argument - `distance` - specifies which distance # metrics to use to calculate distances between strings. # Since the text_align method basically is a wrapper around # diff_align you can get more information via `?diff_align` # and since again diff_align is a wrapper around stringdist # from the stringdist package `?stringdist::stringdist` and # also ``?stringdist::`stringdist-metrics` `` will provide # further insights about possible metrics and how to use the # rest of the arguments to text_align (these are passed through # to stringdist). # Let's have an example using the Levenshtein distance to # calculate distances between tokens (lines per default). # Furthermore we allow the distance between two aligned tokens # to be as large as 15. Tokens which do not find a partner # below that distance are considered to have been deleted # or respectively inserted. Tokens which find a partner with # a non-zero distance which is not above the threshhold are # considered changes - transformations of one token into the other. # The following shows the resulting list of alignment data.frames. dp$text_align(distance = "lv", maxDist = 15) dp$alignment # To measure the change between those two texts we can e.g. aggregate # the distances by change type: sum_up_changes <- function(x){ x %>% dplyr::group_by(type) %>% dplyr::summarise(sum_of_change = sum(distance)) } lapply( dp$alignment, sum_up_changes) ## Coding Texts # Now let us put some data into our diffrproject. # The most basic method to do so is simply called text_code. # Text_code takes up # to five arguments (the first three are mandatory), where one # specifies the text to be coded (`text`, either by index # number or by name), how the variable to store the information # is called (`x`), and the index number or a vector of those # indicating which characters of the text should be coded. # The last two parameters are optional and specify which value # the variable should hold (`val`) and at which hierarchy # level the coding is placed (`hl`, higher or equal hierarchy # levels will overwrite existing codings of lower hierarchy # level for the same text, character span, and variable). dp$text_code(text = 1, x = "start", i=1:5, val = TRUE, hl = 0) dp$text_code(text = "version2", x = "start", i=1:5, val = TRUE, hl = 0) # The text_code method is quite verbose and in most cases more suited # to be accessed by a machine or algorithm than by a human. # Therefore, there are three other methods to code text: # text_code_regex, text_code_alignment_token, # text_code_alignment_token_regex. # The text_code_regex method allows to search for text patterns and # code a whole pattern instead of assigning codes character by # character - the `i` argument of text_code gets replaced by a # `pattern` argument. The in addition further arguments can be # passed to the pattern search functions via `...` - see e.g. # `?grep` for possible further arguments and # https://stat.ethz.ch/R-manual/R-devel/library/base/html/regex.html for a # description of regular expressions in R. # In this example we are searching for the word *"it"* in text 1 and code # each instance. dp$text_code_regex(text = 1, x = "it", pattern = "\\bit\\b", ignore.case=TRUE) # Another variant of coding text is by using alignment tokens. # Having alignment data availible, this allows for selecting: # link, alignment and text while the other arguments from above stay the # same. # having a look at alignment number 4 dp$alignment[[1]][4,] # coding text connected by alignment number 4 dp$text_code_alignment_token( link = 1, alignment_i = 4, text1 = TRUE, text2 = TRUE, x = "token_coding", val = 4, hl = 0 ) ## Getting Text Codings # The most basic way to get text data is to use the text_data method. # This method will go through all or only selected texts, gather all # the data stored there and put it into a neat data.frame where name # identifies the text from which the data comes per name, char informs # us about the character that was coded, and i refers to the characters # position within the text. All other variables hold the data we added # during the examples above. dp$text_data(text = 1) %>% head() ## Aggregating Text Codings # The usage of text_data has its merits but often one is more # interested in text data aggregated to a specific level. # The following three aggregation functions offer a solution # to this problem: tokenize_text_data_lines, tokenize_text_data_words, # and tokenize_text_data_regex. These three methods make use # of the similiary named methods provided by the rtext package. # One important thing to keep in mind is that using these methods # implies aggregating several data values on character level # into one data value at token level. Therefore there has # to be some aggregation function to be involved. The default # is to use the value that occurs most often on character # level, if more than one distinct values occur more than # once the first is choosen. # The aggregation function can be changed to whatever function the # user seems appropriate by passing it to `aggregate_function` # - as long as it # reduces a vector of values into a vector with only one value. # The `join` argument allows to decide how text and data are # joined # into the resulting data.frame - left: all token, right: all data, full: # token with or without data and data with or without token. dp$tokenize_text_data_lines( text = 1, join = "right", aggregate_function = function(x){ paste(x[1:3], collapse = ",") } ) ## Text Coding Inheritence # Having aligned two texts via token pairs another functionality of # diffrprojects becomes availible: text coding inheritance via no-change # tokens. This means that text codings can get copied to those tokens they # are aligned with, given that they are considered the same - i.e. the # distance equals zero and the change type therefore is no-change. # To show this feature we use the text_inherit method and we will # start with a fresh example. A new project with two texts. The first text # gets some codings, then they are aligned, and in a last step codings are # transfered from one text to the other via the text_data_inherit method. dp <- diffrproject$new()$ text_add(text_version_1)$ text_add(text_version_2)$ text_code_regex( text = 1, x = "test1", pattern = "This part.*?change", val = "inherited" )$ text_code_regex( text = 1, x = "test2", pattern = "This part.*?change", val = "inherited" ) dp$tokenize_text_data_lines(1) dp$ text_link()$ text_align()$ text_data_inherit( link = 1, direction = "forward" ) dp$tokenize_text_data_lines(2) ## Saving and Loading Projects # Diffrprojects also allow for storing and loading project to and # from disk. # note, uncomment code lines to run # save to file # dp$save(file = "dp_save.RData") # remove object # rm(dp) # create new object and load saved data into new object # dp <- diffrproject$new() # dp$load("dp_save.RData") # dp$tokenize_text_data_lines(2) > > > # Another way is to call the ls() method. This will present us with a > # data frame listing all fields where data is stored and all the methods > # (aka object specific functions) of our diffrprojects instance. > # Those methods and fields located in *private* are not for the user > # to mess around with while non-private (*self* aka public) data fields > # can be read by the user and public methods can be triggered by the > # user to manipulate the data or retrieve data in a specific format. > > > dp$ls() name where class 1 execute_load private function 2 hash private function 3 hashed private function 5 prepare_save private function 4 hashes private list 9 alignment_data self alignment_data_list, list 6 alignment self alignment_list, list 21 link self alignment_list, list 7 alignment_add self function 8 alignment_code self function 10 alignment_data_full self function 11 alignment_data_set self function 12 alignment_delete self function 13 clone self function 14 debug self function 15 export_csv self function 16 export_sqlite self function 17 get self function 18 import_csv self function 19 import_sqlite self function 20 initialize self function 22 load self function 23 ls self function 24 message self function 27 save self function 29 text_add self function 30 text_align self function 31 text_code self function 32 text_code_alignment_token self function 33 text_code_regex self function 34 text_data self function 35 text_data_inherit self function 36 text_delete self function 37 text_link self function 38 text_meta_data self function 39 tokenize_text_data_lines self function 40 tokenize_text_data_regex self function 41 tokenize_text_data_words self function 42 warning self function 25 meta self list 26 options self list 28 text self list > > > # The base R class() function furthermore reveals from which classes the > # diffrproject class inherits: > > class(dp) [1] "diffrproject" "dp_inherit" "dp_align" [4] "dp_export" "rtext_loadsave" "dp_base" [7] "R6_rtext_extended" "R6" > > > > ## Adding Texts to Projects > > # Our diffrproject (`dp`) has one method called `text_add()` that allows to > # add texts to the project. Basically the method can be used in three > # different flavors: adding character vectors, adding texts stored on disk, > # or by adding rtext objects (see rtext package: > # https://CRAN.R-project.org/package=rtext; rtext objects are the way > # individual texts are represented within diffrprojects). > # For each of these used cases there is one option: > # `text`, `text_file`, `rtext`; respectively. > > # Below are shown examples using each of these methods: > > > # **adding text files** > > test_file1 <- stringb:::test_file("rc_1_ch1.txt") > test_file2 <- stringb:::test_file("rc_2_ch1.txt") > dp$text_add(text_file = c(test_file1, test_file2) ) rtext : initializing rtext : initializing > > > # **adding rtext objects** > > test_file <- stringb:::test_file("rc_1_ch1.txt") > rt <- rtext$new( text_file = test_file) rtext : initializing > dp$text_add(rtext = rt) > > > # **adding character vectors** > > test_file1 <- stringb:::test_file("rc_1_ch1.txt") > test_file2 <- stringb:::test_file("rc_2_ch1.txt") > cv <- "" > cv[1] <- text_read(test_file1, NULL) > cv[2] <- text_read(test_file2, NULL) > dp$text_add(text = cv) rtext : initializing rtext : initializing > > # In the last case make sure to put each text in one separate line. > # Functions like readLines() or text_read() read in texts such that > # each line corresponds to one element in a character vector. With e.g. > # text_read()'s tokenize parameter to NULL the text will be read in as one > # long string. > > > > > ## Piping Methods > > # Now is a good time to mention a feature of diffrprojects that comes in > # handy: All functions that do not explicitly extract data > # (those usually have some 'get' as part of their name) do return the > # object itself so that one can pipe together a series of method calls. > > # Consider the following example where we initiate a new diffrprojects > # instance and add two texts in just one pipe: > > dp <- + diffrproject$ + new()$ + text_add(text_version_1, name = "version1")$ + text_add(text_version_2, name = "version2") rtext : initializing rtext : initializing > > length(dp$text) [1] 2 > > > > > > > ## Getting Infos About Texts > > # If we want to get some general overview about the texts gathered in our > # project, we can use the text_meta_data() method to do so. > # The method has no parameters and returns a data.frame with several > # variables informing us about its source, length, encoding used for > # storage, and its name. > > > dp$text_meta_data() text_file character encoding sourcetype name 1 479 UTF-8 text version1 2 539 UTF-8 text version2 > > > > > ## Showing Text > > # If you want to have a look at your texts you may do so by using the > # text's own text_show methods. Per default this method only shows the > # first 500 characters, but it can be set to higher numbers as well. > > > dp$text$version1$text_show(length=1000) This part of the document has stayed the same from version to version. It shouldn't be shown if it doesn't change. Otherwise, that would not be helping to compress the size of the changes. This paragraph contains text that is outdated. It will be deleted in the near future. It is important to spell check this dokument. On the other hand, a misspelled word isn't the end of the world. Nothing in the rest of this paragraph needs to be changed. Things can be added after it. > dp$text$version2$text_show(length=1000) This is an important notice! It should therefore be located at the beginning of this document! This part of the document has stayed the same from version to version. It shouldn't be shown if it doesn't change. Otherwise, that would not be helping to compress anything. It is important to spell check this document. On the other hand, a misspelled word isn't the end of the world. Nothing in the rest of this paragraph needs to be changed. Things can be added after it. This paragraph contains important new additions to this document. > > > > > ## Getting And Setting Infos About the Project > > # Similar to the text_meta_data() method we can access the projects > # meta data via data fields meta and options. But contrary to the > # text_meta_data() method that gathers data from all the texts within the > # project and does not allow for manipulation of the data, the data > # fields allow reading and writing. > > # First let us have a look and thereafter turn off the message > # notification service: > > # **getting data fields** > > dp$options $verbose [1] TRUE $warning [1] TRUE $ask [1] TRUE > > > # **setting data fields** > > dp$options$verbose <- FALSE > > > # (note, ask is deprecated and only remains for compatibility > # reasons but has no function anymore) > > # Now it's time to have a look at the projects meta data. > # It tells us when the project was created, which path to use for > # SQLite exports, which path to use for saving data as in RData > # format and what is the projects id. The id is a hash of a time stamp > # as well as session information which should ensure uniqueness across > # space and time. > > # All these values can be manipulated by the user to her liking. > > > dp$meta $ts_created [1] "2020-04-22 20:06:22 UTC" $db_path [1] "./diffrproject.db" $file_path [1] "" $project_id [1] "97760e0ec337d3480d98cc61ab797fc0" > > dp$meta$file_path = "./diffrproject.RData" > > > > > > > > > ## Deleting Texts > > # Of cause we can not only add texts but delete them from the project as > # well. For this purpose there is the text_delete() method. > > # Let's just add two texts and delete one by providing its index number and > # the second by providing its name to the text_delete() method. > > > dp$text_add(text = "nonesense", "n1") > dp$text_add(text = "nonesense", "n2") > > dp$text_delete(3) > dp$text_delete("n2") > > length(dp$text) [1] 2 > names(dp$text) [1] "version1" "version2" > > > > > > > > > ## Defining Relationships Between Texts: Linking > > # The purpose of diffrprojects is to enable data collection on the > # difference of texts. Having filled a project with various texts, > # there are endless possibilities to form pairs of text for comparison > # and change measurement - where endless actually is equal to: $n^2-n$. > > # Linking can be done via the text_link method which accepts either > # index numbers or text names for its from and to arguments > # (a third argument delete will delete a specified link if set to TRUE). > > > dp$text_link(from = 1, to = 2) > dp$text_link(from = 1, to = 2, delete = TRUE) > > > # If no arguments are specified, text_link will link the first text to > # the second, the third to the fourth, the fourth to the fifths and so on. > > > dp$text_link() > > > > # To get an idea of what links are currently specified, we can > # directly access the link data field or/and ask R to transform the > # list found there into a data.frame. > > > dp$link $`version1~version2` $`version1~version2`$from [1] "version1" $`version1~version2`$to [1] "version2" attr(,"class") [1] "alignment_list" "list" > > dp$link %>% as.data.frame() from to link 1 version1 version2 version1~version2 > > > > > > > > > > ## Aligning Texts and Measuring Change > > # At the heart of each diffrproject lies the text_align method. > # This method compares two texts and tries to align parts > # of one text with parts of the other text. The first two > # arguments (`t1` and `t2`) are for specifying which pair > # of texts to compare - if left as-is, all text pairs that > # are specified within the link data field will be aligned. > > # Text parts are arbitrary character spans defined by the > # `tokenizer` argument. This argument expects a function splitting > # text into a token data.frame. If the tokenizer argument > # is left as-is, it will default to text_tokenize_lines function > # from the stringb package. > > # Text tokens can be pre-processed before alignment. The `clean` > # argument allows to hand over a function tranforming a charactr > # vector of text tokens into their clean counterparts. > > # The `ignore` arguments expects a function that is able to > # transform a character vector of tokens into a logical vector > # of same length, indicating which tokens to ignore throughout > # the alignment process and which to consider. > > # The next argument - `distance` - specifies which distance > # metrics to use to calculate distances between strings. > > # Since the text_align method basically is a wrapper around > # diff_align you can get more information via `?diff_align` > # and since again diff_align is a wrapper around stringdist > # from the stringdist package `?stringdist::stringdist` and > # also ``?stringdist::`stringdist-metrics` `` will provide > # further insights about possible metrics and how to use the > # rest of the arguments to text_align (these are passed through > # to stringdist). > > # Let's have an example using the Levenshtein distance to > # calculate distances between tokens (lines per default). > # Furthermore we allow the distance between two aligned tokens > # to be as large as 15. Tokens which do not find a partner > # below that distance are considered to have been deleted > # or respectively inserted. Tokens which find a partner with > # a non-zero distance which is not above the threshhold are > # considered changes - transformations of one token into the other. > > # The following shows the resulting list of alignment data.frames. > > > dp$text_align(distance = "lv", maxDist = 15) > > dp$alignment $`version1~version2` alignment_i token_i_1 token_i_2 distance type from_1 to_1 from_2 to_2 1 1 1 6 0 no-change 1 16 97 112 2 2 2 7 0 no-change 18 40 114 136 3 3 3 8 0 no-change 42 61 138 157 4 4 4 9 0 no-change 63 84 159 180 5 5 5 10 0 no-change 86 107 182 203 6 6 6 11 0 no-change 109 132 205 228 7 7 7 12 0 no-change 134 156 230 252 8 8 8 13 14 change 158 181 254 271 9 9 9 5 8 change 183 190 86 94 11 10 10 23 0 no-change 193 215 475 497 12 11 11 25 13 change 217 238 523 539 13 12 12 NA 25 deletion 240 264 NA NA 14 13 13 5 11 change 266 277 86 94 16 14 14 14 0 no-change 280 303 274 297 17 15 15 15 1 change 305 327 299 321 18 16 16 16 0 no-change 329 345 323 339 19 17 17 17 0 no-change 347 367 341 361 20 18 18 18 0 no-change 369 389 363 383 21 19 19 19 0 no-change 391 412 385 406 22 20 20 20 0 no-change 414 436 408 430 23 21 21 21 0 no-change 438 459 432 453 24 22 22 22 0 no-change 461 478 455 472 15 23 NA 1 20 insertion NA NA 1 20 25 24 NA 2 17 insertion NA NA 22 38 31 25 NA 3 23 insertion NA NA 40 62 41 26 NA 4 21 insertion NA NA 64 84 27 27 NA 24 23 insertion NA NA 499 521 attr(,"class") [1] "alignment_list" "list" > > > > # To measure the change between those two texts we can e.g. aggregate > # the distances by change type: > > > sum_up_changes <- function(x){ + x %>% + dplyr::group_by(type) %>% + dplyr::summarise(sum_of_change = sum(distance)) + } > > lapply( dp$alignment, sum_up_changes) $`version1~version2` # A tibble: 4 x 2 type sum_of_change 1 change 47 2 deletion 25 3 insertion 104 4 no-change 0 > > > > > > > ## Coding Texts > > # Now let us put some data into our diffrproject. > > # The most basic method to do so is simply called text_code. > # Text_code takes up > # to five arguments (the first three are mandatory), where one > # specifies the text to be coded (`text`, either by index > # number or by name), how the variable to store the information > # is called (`x`), and the index number or a vector of those > # indicating which characters of the text should be coded. > # The last two parameters are optional and specify which value > # the variable should hold (`val`) and at which hierarchy > # level the coding is placed (`hl`, higher or equal hierarchy > # levels will overwrite existing codings of lower hierarchy > # level for the same text, character span, and variable). > > > dp$text_code(text = 1, x = "start", i=1:5, val = TRUE, hl = 0) > dp$text_code(text = "version2", x = "start", i=1:5, val = TRUE, hl = 0) > > > # The text_code method is quite verbose and in most cases more suited > # to be accessed by a machine or algorithm than by a human. > # Therefore, there are three other methods to code text: > # text_code_regex, text_code_alignment_token, > # text_code_alignment_token_regex. > > # The text_code_regex method allows to search for text patterns and > # code a whole pattern instead of assigning codes character by > # character - the `i` argument of text_code gets replaced by a > # `pattern` argument. The in addition further arguments can be > # passed to the pattern search functions via `...` - see e.g. > # `?grep` for possible further arguments and > # https://stat.ethz.ch/R-manual/R-devel/library/base/html/regex.html for a > # description of regular expressions in R. > > # In this example we are searching for the word *"it"* in text 1 and code > # each instance. > > > dp$text_code_regex(text = 1, x = "it", pattern = "\\bit\\b", ignore.case=TRUE) > > > # Another variant of coding text is by using alignment tokens. > # Having alignment data availible, this allows for selecting: > # link, alignment and text while the other arguments from above stay the > # same. > > > > # having a look at alignment number 4 > > dp$alignment[[1]][4,] alignment_i token_i_1 token_i_2 distance type from_1 to_1 from_2 to_2 4 4 4 9 0 no-change 63 84 159 180 > > > # coding text connected by alignment number 4 > > dp$text_code_alignment_token( + link = 1, + alignment_i = 4, + text1 = TRUE, + text2 = TRUE, + x = "token_coding", + val = 4, + hl = 0 + ) > > > > > > > > > ## Getting Text Codings > > # The most basic way to get text data is to use the text_data method. > # This method will go through all or only selected texts, gather all > # the data stored there and put it into a neat data.frame where name > # identifies the text from which the data comes per name, char informs > # us about the character that was coded, and i refers to the characters > # position within the text. All other variables hold the data we added > # during the examples above. > > > dp$text_data(text = 1) %>% head() i char start it token_coding name 1 1 T TRUE NA NA version1 2 2 h TRUE NA NA version1 3 3 i TRUE NA NA version1 4 4 s TRUE NA NA version1 5 5 TRUE NA NA version1 6 63 v NA NA 4 version1 > > > > > > > ## Aggregating Text Codings > > # The usage of text_data has its merits but often one is more > # interested in text data aggregated to a specific level. > # The following three aggregation functions offer a solution > # to this problem: tokenize_text_data_lines, tokenize_text_data_words, > # and tokenize_text_data_regex. These three methods make use > # of the similiary named methods provided by the rtext package. > > # One important thing to keep in mind is that using these methods > # implies aggregating several data values on character level > # into one data value at token level. Therefore there has > # to be some aggregation function to be involved. The default > # is to use the value that occurs most often on character > # level, if more than one distinct values occur more than > # once the first is choosen. > > # The aggregation function can be changed to whatever function the > # user seems appropriate by passing it to `aggregate_function` > # - as long as it > # reduces a vector of values into a vector with only one value. > > # The `join` argument allows to decide how text and data are > # joined > # into the resulting data.frame - left: all token, right: all data, full: > # token with or without data and data with or without token. > > > dp$tokenize_text_data_lines( + text = 1, + join = "right", + aggregate_function = + function(x){ + paste(x[1:3], collapse = ",") + } + ) Error in aggregate.data.frame(., by = list(token_i = token_i), FUN = aggregate_function, : arguments must have same length Calls: ... withVisible -> -> -> aggregate.data.frame Execution halted ** running examples for arch 'x64' ... [7s] OK * checking for unstated dependencies in 'tests' ... OK * checking tests ... ** running tests for arch 'i386' ... [71s] OK Running 'test-all.R' [70s] ** running tests for arch 'x64' ... [64s] OK Running 'test-all.R' [63s] * checking PDF version of manual ... OK * DONE Status: 1 ERROR