block_pour_docx but only for part of file to be poured

[SchmidtPaul]

I am writing a docx-document using RMarkdown and would like to include parts of another file otherdoc.docx. I guess it is like using block_pour_docx, but only for parts of the document.

Let's say otherdoc.docx looks like this:

In my RMarkdown document I would like to work like so:

---
output: officedown::rdocx_document
---

```{r setup, include=FALSE}
library(officedown)
library(officer)
```

# Introduction

Some text and r chunks here

```{r}
source("somecode.R"))
result
```

# Part 1

More text and r chunks here

```{r}
# Pour ONLY THE PART 1 content of the docx
block_pour_docx("otherdoc.docx")
```

# Part 2

More text and r chunks here

```{r}
# Pour ONLY THE PART 2 content of the docx
block_pour_docx("otherdoc.docx")
```

What are my options to achieve my goal here?

P.S.: Thanks for the fantastic officeverse

[davidgohel]

Hello

I don't know. There is no function like this in officer nor officedown.

Did you try with officer::cursor_* in a loop with officer::body_remove()?

[SchmidtPaul]

Ok, thanks - I see.
I've come up with a first solution that works given that the parts I want are named via and separated by heading 1-headers. This would be the R code that extracts a part (here it is Part 2).

However, please find my question at the end.

slicedocx.R

library(officedown)
library(officer)
library(tidyverse)

# gather necessary info ---------------------------------------------------
doc <- read_docx("otherdoc.docx")
doc_summ <- docx_summary(doc)

index_max <- max(doc_summ$doc_index)

doc_slices <- doc_summ %>%
  filter(style_name == "heading 1") %>%
  transmute(text = text,
            index_start = doc_index,
            index_end = lead(doc_index)-1)


# get a part --------------------------------------------------------------
want <- "Part 2"

docpart <- doc

i <- doc_slices %>% 
  filter(text == want) %>% 
  pivot_longer(-text) %>% 
  pull(value)


# delete everything after part
if (!is.na(i[2])) {
  for (j in (i[2] + 1):index_max) {
    docpart <- docpart %>% cursor_end() %>% body_remove()
  }
}

# delete everything before part
if (i[1]>1) {
  for (j in 1:(i[1]-1)) {
    docpart <- docpart %>% cursor_begin() %>% body_remove()
  }
}

For comparison:

docx_summary(doc) # before
#    doc_index content_type     style_name                         text level num_id row_id is_header cell_id col_span row_span
# 1           1    paragraph      heading 1                       Part 1    NA     NA     NA        NA      NA       NA       NA
# 2           2    paragraph           <NA> This is some text in part 1.    NA     NA     NA        NA      NA       NA       NA
# 3           3    paragraph List Paragraph                     Part 1 a     1      1     NA        NA      NA       NA       NA
# 4           4    paragraph List Paragraph                     Part 1 b     1      1     NA        NA      NA       NA       NA
# 5           5    paragraph           <NA>     And more text in part 1.    NA     NA     NA        NA      NA       NA       NA
# 6           6    paragraph      heading 1                       Part 2    NA     NA     NA        NA      NA       NA       NA
# 7           7    paragraph           <NA> This is some text in part 2.    NA     NA     NA        NA      NA       NA       NA
# 1.1         8   table cell     Table Grid                    Tabhead A    NA     NA      1     FALSE       1        1        1
# 1.4         8   table cell     Table Grid                            1    NA     NA      2     FALSE       1        1        1
# 2.2         8   table cell     Table Grid                    Tabhead B    NA     NA      1     FALSE       2        1        1
# 2.5         8   table cell     Table Grid                            2    NA     NA      2     FALSE       2        1        1
# 3.3         8   table cell     Table Grid                    Tabhead C    NA     NA      1     FALSE       3        1        1
# 3.6         8   table cell     Table Grid                            3    NA     NA      2     FALSE       3        1        1
# 11          9    paragraph           <NA>                                 NA     NA     NA        NA      NA       NA       NA
# 12         10    paragraph      heading 1                       Part 3    NA     NA     NA        NA      NA       NA       NA
# 13         11    paragraph           <NA> This is some text in part 3.    NA     NA     NA        NA      NA       NA       NA

doc_slices
#      text index_start index_end
# 1  Part 1           1         5
# 6  Part 2           6         9
# 12 Part 3          10        NA

docx_summary(docpart) # after
#     doc_index content_type style_name                         text level num_id row_id is_header cell_id col_span row_span
# 1           1    paragraph  heading 1                       Part 2    NA     NA     NA        NA      NA       NA       NA
# 2           2    paragraph       <NA> This is some text in part 2.    NA     NA     NA        NA      NA       NA       NA
# 1.1         3   table cell Table Grid                    Tabhead A    NA     NA      1     FALSE       1        1        1
# 1.4         3   table cell Table Grid                            1    NA     NA      2     FALSE       1        1        1
# 2.2         3   table cell Table Grid                    Tabhead B    NA     NA      1     FALSE       2        1        1
# 2.5         3   table cell Table Grid                            2    NA     NA      2     FALSE       2        1        1
# 3.3         3   table cell Table Grid                    Tabhead C    NA     NA      1     FALSE       3        1        1
# 3.6         3   table cell Table Grid                            3    NA     NA      2     FALSE       3        1        1
# 11          4    paragraph       <NA>                                 NA     NA     NA        NA      NA       NA       NA

@davidgohel How would I now pour the docpart into my RMarkdown file? If I understand it correctly I could print a new .docx file first in order to use block_pour_docx(), but there is probably an easier way? This obviously does not work:

mymarkdown.Rmd

---
output: officedown::rdocx_document
---

# Introduction

Some text and r chunks here

```{r}
source("slicedocx.R")
print(docpart)
```

More text and r chunks here

[SchmidtPaul]

Hey @davidgohel, I took another shot at this and would be thankful for a comment.

Basically, I created a function get_docx_part() that

1. reads a docx file,
2. removes everything except the part I am interested in. The part I am interested in is simply defined as everything after a certain header and before the next header (both of style heading 1).
3. prints a temporary docx file temp.docx

I can then use block_pour_docx("temp.docx") inside my RMarkdown file.

Function

library(officedown)
library(officer)
library(tidyverse)

get_docx_part <- function(infile, heading1title, outfile){
  
  # import and summary ------------------------------------------------------
  doc <- read_docx(infile)
  doc_summ <- docx_summary(doc)
  index_max <- max(doc_summ$doc_index)
  
  # info on all parts (i.e. sections separated by "heading 1" headers)
  doc_parts_info <- doc_summ %>%
    filter(style_name == "heading 1") %>%
    transmute(text = text,
              index_start = doc_index,
              index_end = lead(doc_index)-1)
  

  # delete unwanted parts ---------------------------------------------------
  # prepare
  docpart <- doc
  
  i <- doc_parts_info %>% 
    filter(text == heading1title) %>% 
    pivot_longer(-text) %>% 
    pull(value)
  
  # delete everything after part
  if (!is.na(i[2])) {
    for (j in (i[2] + 1):index_max) {
      docpart <- docpart %>% cursor_end() %>% body_remove()
    }
  }
  
  # delete everything before part
  if (i[1]>1) {
    for (j in 1:(i[1]-1)) {
      docpart <- docpart %>% cursor_begin() %>% body_remove()
    }
  }
  

  # print -------------------------------------------------------------------
  # create temporary docx file
  print(docpart, target = outfile)

}

Example Rmd

---
output: officedown::rdocx_document
---

# Introduction

Some text and r chunks here

```{r, echo=FALSE, message=FALSE}
source("get_docx_part.R") # get function

get_docx_part(infile = "BaseDocument.docx",
              heading1title = "Second header title",
              outfile = "temp.docx")

block_pour_docx("temp.docx")
```

# Conclusion

More text and r chunks here

Check results

Questions

As you can see, it so far works as intended. However,

• what do you generally think about the approach?
• I actually wanted to include the block_pour_docx("temp.docx") in the get_docx_part() function, but it did not work when inside the function?!
• I wanted to delete temp.docx immediately after pouring, but it led to an error?!

[davidgohel]

Hello

1. I did not test but it seems ok to me
2. yes, it needs to be executed in a knitr context, otherwise knitr::knit_print() is not called
3. wait the final document is created and then you can delete the file you want to pour

[SchmidtPaul]

Hey @davidgohel, I have been implementing this method successfully now, but realized it does not extend well for repeating the step multiple times in the same knitting process. This is because everytime a "temp.docx" is created to be poured, which overwrites the old one, but apparently the pouring happens at the very end. So I ultimately end up with all poured section being identical to the last one that was poured.

I guess this goes along with you telling me to "wait the final document is created and then you can delete the file you want to pour".

Any suggestions how I could tackle this issue?

[davidgohel]

I would use tempfile(fileext = ".docx") instead of hard coded value ("temp.docx")