line / conflr Goto Github PK

Prepare for release:

• Check that description is informative
• Check licensing of included files
• usethis::use_cran_comments()
• devtools::check()
• devtools::check_win_devel()
• rhub::check_for_cran()
• Polish pkgdown reference index
• Draft blog post

Submit to CRAN:

• usethis::use_version('minor')
• Update cran-comments.md
• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted 🎉
• usethis::use_github_release()
• usethis::use_dev_version()
• Update install instructions in README
• Finish blog post
• Tweet
• Add link to blog post in pkgdown news menu

How can I use this package to access an installation that doesn't support user-password authentication? Is this a mode of operation you would support?

First off, thanks for your work on this package, I'm very excited about it!

After trying to upload a page, I noticed that all the images were resized to "large" in confluence causing them to become unreadable/not match the preview. Is it possible to set them to display at "original" size without manually editing the page?

Thanks!

Plotly is a great and versatile plotting tool for R, especially in R markdown.

I think conflr does not currently support plotly graphs...

However, this would be a fantastic addition to an awesome and much needed tool!

Thanks for this great package I think it'll help me a lot.

Maybe I'm using the wrong workflow but I'm experiencing some awkwardness with space names and page ids.

As a test I opened a confluence page and decide to use it as a parent for the page I wish to create. I believe the full information is in the url, however I have to extract the space name (easy enough as it's in my url), and find the page id, which I do by finding the page nformation here :

Then copy paste the id from the URL, Is there a way to get the pageId and space right from the url ? Or maybe even better, could I provide in the UI the url instead of space and pageID ?

Now the other way around, I wanted to delete the page I created (or maybe I'd want to use it as a new parent), but I don't know its id, so I have to go manually on the page and do the same process.

The most convenient option for me would be to have the choice between inputing space and pageId, or just inputing an "url" parameter. In any case a get_pageId_from_url() function would help.

I'd also appreciate having the pageId logged in the console when running confl_create_post_from_Rmd_addin(), and returned invisibly by confl_create_post_from_Rmd()

Again apologies if I'm just using it wrong, and looking forward for the CRAN release.

Example document:

---
title: "test page"
output:
  conflr::confluence_document:
    space_key: "***"
---

<ac:structured-macro ac:name="expand">
  <ac:parameter ac:name="title">hidden stuff below</ac:parameter>
  <ac:rich-text-body><![CDATA[<p>foo</p>]]></ac:rich-text-body>
</ac:structured-macro>

Note the CDATA section for passing in HTML.

This fails with:

Error in doc_parse_raw(x, encoding = encoding, base_url = base_url, as_html = as_html,  : 
  Opening and ending tag mismatch: confl-ac-rich-text-body line 1 and p [76]
Calls: render ... read_xml.character -> read_xml.raw -> doc_parse_raw

And renders a markdown file as:

<ac:structured-macro ac:name="expand"> <ac:parameter ac:name="title">hidden stuff below</ac:parameter> <ac:rich-text-body><![CDATA[<p>foo

</p>

\]\]\></ac:rich-text-body> </ac:structured-macro>

Note that the closing tag of the CDATA is somehow got escaped.

Not sure what's causing the problem, as conflr:::restore_cdata seems to match it and run correctly (after patching to work with rich-text-body as well besides plain-text-body), but some other mechanism might mess up the closing CDATA tag later on in the process?

Currently, whether to confirm or not is determined when space_key is not provided

This means this might return different page on a different space.

This should add TOC, not a string of "<ac:structured-macro ac:name="toc" ac:schema="1"></ac:structured-macro>"

---
title: test
---

```{r echo=FALSE}
library(htmltools)
p(tag("ac:structured-macro",list("ac:name"="toc", "ac:schema"="1")))
```

### A

### B

### C

I got that

on your pkgdown website https://line.github.io/conflr/

This cannot be done on the page creation, so this might increase the implementational complexity. We need to consider if this is worth adding.

https://developer.atlassian.com/cloud/confluence/rest/#api-api-content-id-label-post

Hi,
I really like the project by the way. Thank you very much for sharing it. You had a data.table issue a while back that you couldn't recreate so I've supplied a sample that fails.

I've attached a txt file - just change the extension to .rmd. The data.table operations work under knitr but fail under conflr saying that dt is not a data.table.

TestDataTable.txt

When I tried to use a YAML param to create a dynamic title, it showed up fine when I knit it, but in the confluence page, it just showed the R code. (added quotes to syntax for formatting)

---
output: html_document
params:
  peril: 'Peril Name'
  version: 0.03
title: "`r paste0(params$peril,' - v', params$version,'a')`"
---

I'm getting an error when trying to post a markdown file that includes manual page breaks (<br>) in it. See the reproducible example below.

library(conflr)
md_text <- c("`Test Markdown`",
             "==============",
             "",
             "This is a test file.",
             "<br>",
             "It has a break in it.")
html_text <- commonmark::markdown_html(md_text)


page <- confl_update_page(
  id = "244877109",
  title = "Test",
  body = html_text)
#> Error in doc_parse_raw(x, encoding = encoding, base_url = base_url, as_html = as_html, : Opening and ending tag mismatch: br line 3 and p [76]

^{Created on 2019-03-08 by the reprex package (v0.2.1)}

It looks like this error occurs in the private function translate_to_confl_macro, specifically in line 2 below:

html_doc <- xml2::read_xml(html_text, options = c("RECOVER", "NOERROR", "NOBLANKS"))

It seems like I can resolve this error by replacing xml2::read_xml with xml2::read_html, which doesn't seem to check to make sure every tag has a closing pair (which is not needed for manual breaks using <br>).

Is it possible to replace xml2::read_xml with xml2::read_html in the package source?

See #41

Is it already somehow possible to directly invoke Confluence's macros (https://confluence.atlassian.com/doc/macros-139387.html) from within the source Rmd?

This is easy with expand macro.

<ac:structured-macro ac:name="expand">
  <ac:parameter ac:name="title">title1</ac:parameter>
  <ac:rich-text-body>
    <p>aaa</p>
  </ac:rich-text-body>
</ac:structured-macro>

Hi,

First of - 'conflr' is a real time saver for me!
Your work is appreciated!

I am facing similar issue are described here:
https://jira.atlassian.com/browse/CONFSERVER-38804

i.e. watchers are being notifed not only when I post new page but also when I add content or modify it. This seems to be resulting in sending notification each time I invoke knitr::include_graphics().
According to the link provided this ca be partially solved by using "minorEdit" : true .
Do you plan on implementing it or having any other workaround for the problem ?

Best regards,
P.

Since named capture group is not supported on the older systems, it seems better to tweak this regex pattern for compatibility.

c.f. yutannihilation/tiedr#10

Hello !

Everything works fine until I try to "publish" my page on confluence. The process seems to be stuck and I have only one warning which is not very helpful :

Warning: Error in : Condition message must be a string
86: <Anonymous>

I integrate "confluence_document: " on my rmarkdown, I log in to confluence directly on the command line, I have access to the preview before publishing. It's just when I click on "Publish" that the issue appears. I use the following parameters :

    parent_id: 436666369
    toc: TRUE
    toc_depth: 4
    code_folding: hide
    supported_syntax_highlighting:
      r: r
      foo: bar
    update: true 
    use_original_size: true 

Does anyone have any idea ?

Hello, I am using the nvimr plugin and not Rstudio.
I'd like to be able to create/update a confluence page from the R console.
I tried to do

R> conflr::confl_create_post_from_Rmd(here::here("fi_aware.Rmd")) 
Error: RStudio not running
In addition: Warning message:
In stop(e, call. = FALSE) : additional arguments ignored in stop()

Is there a function I can use that would be like

conflr::confl_create_post_from_Rmd(Rmd_file, spaceId, username, ...)

that would allow me to work without any rstudio ?

Currently, it seems the language of an rmarkdown code chunk is lost in translation between the Rmd to md conversion, and then can't be reported to html conversion, and finally confluence post processing of html to create xml storage format.

However, it is possible to add the option to code block macro using

<ac:structured-macro ac:name="code">
  <ac:parameter ac:name="language">r</ac:parameter>
  <ac:plain-text-body><![CDATA[this is my code]]></ac:plain-text-body>
</ac:structured-macro>

https://confluence.atlassian.com/conf59/code-block-macro-792499083.html#CodeBlockMacro-Storageformatexample

I think it could be possible to add such support. I can look into it, but first what do you think ?
Did you already try something ?

Thanks.

Probably we can do this by setting some post_processor.

---
title: "foo"
output:
  confluence_document:
    space_key: "my_space"
    parent_id: "123456789"
    ...
---

Hi. I tried to post to our company Confluence page yesterday and am experiencing the error above. I'm not sure if it's an issue per se or just unique to my set up, any guidance would be much appreciated.

Duplicate post on rstudio community: https://community.rstudio.com/t/conflr-errors-when-trying-to-publish-a-rmd-page-to-confluence/61832

I have created a Rmd doc and using output: github_document I was also able to generate a readme for a Github repo.

Some folk in our company do not have access to Github and they wouldlike me to publish the Rmd documentation of a project to Confluence. I found the https://github.com/line/conflr library for doing this.

I've followed the set up guidelines and added global variables to my .Renviron, including

CONFLUENCE_URL="https://confluence.somecompany.com"
CONFLUENCE_USERNAME="doug"
CONFLUENCE_PASSWORD="[my pass]"

I then used the conflr add in to attempt to post to a page. I entered the Space of the page which is called 'Analytics' and then the parent page ID (Which I found after clicking on 'page info' in Confluence on the parent page I'd like to post to). After I click publish I get this output in the console:

No encoding supplied: defaulting to UTF-8.
Listening on http://127.0.0.1:3415
No encoding supplied: defaulting to UTF-8.
No encoding supplied: defaulting to UTF-8.
No encoding supplied: defaulting to UTF-8.
Warning: Error in parse_url: length(url) == 1 is not TRUE
65:

I tried Google searching for this error but did not find anything in relation to it within the context of conflr. How should I interpret this error and how can I overcome it?

I tried some other functions from the package just to see if it recognized by credentials in .Renviron (I restarted R after updating .Renviron with my credentials).

e.g. list pages in a space

confl_list_pages(spaceKey = 'Analytics')
No encoding supplied: defaulting to UTF-8.
{html_document}
<html>
[1] <head>\n<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">\n<title>Sign in ・ Cloudflare Access</title ...
[2] <body>\n    <div class="Surface">\n      <div class="AuthBox">\n        <div class="AuthBox-body">\n          <a class ...

I'm not sure if this means there's a successful read of my credentials or not?

How can I post my Rmd file to Confuence?

I can not use conflr to upload LaTex.

here is my latex content:

$$ y = \frac {1 }{2}$$

result in wiki:

it seems that the latex macro is wrong

Hello!

It looks like confl_create_post_from_Rmd() no longer returns the URL of the published page? Is there a way to get that URL still? I was sending it along with an alert that a page was published.

Thanks!

The default value of the Shiny UI component for the Confluence space is the user's personal space. The conflr package attempts to use the username as the space key, but personal spaces often a numeric space key (see here).

It would be good to suppress the warning message like:

Warning message:
In confl_verb("GET", glue::glue("/space/{spaceKey}"), query = purrr::compact(query)) :
  Error in stop(httr::http_condition(res, type = "error"), httr::content(res)): Not Found (HTTP 404).
404list(authorized = FALSE, valid = TRUE, errors = list(), successful = FALSE)No space found with key : ~david.hall

This is exciting!

But per Atlassian, usernames are no longer a thing in Atlassian Cloud. Is it still possible to authenticate with conflr? If so, how?

We must pass escaped version of paths to confl_upload_interactively(). Maybe this is too early:

currently, I can not update an existing Wiki by conflr directly.

I wish to use conflr in Rscript and directly update the content with function.

it looks like this:

conflr::confl_login(username = "xxx",passwd = "xxx",url = "xxx.com")
conflr::confl_update_post_from_Rmd("test.Rmd",space = "xxx",page_id = 123456)

https://rlang.r-lib.org/reference/list2.html

I am using knitr::include_graphics() with .png files to include graphics in the confluence page. The images are uploaded but don't show up in the page. I am able to manually edit the page to put the images in without having to re-upload them.

• Emacs/Vim
• non-interactive sessions

maxLevel seems the corresponding parameter.

<ac:structured-macro ac:name="toc">
  <ac:parameter ac:name="maxLevel">3</ac:parameter>
</ac:structured-macro>

conflr is great package to talk to Wiki in my daily job. Wish someone can propose a lovely hex sticker like bookdown logo.

Use modal.

When I call confl_create_post_from_Rmd(interactive=F) from a terminal R session, I receive the following error.

Error in browseURL(result_url) : 
  'browser' must be a non-empty character string

The function works when I call it from the command line in R studio, opening a browser with the created page. But this is a behavior I do not need or want in "Batch Use". I want to create large numbers of pages without R attempting to open a browser to them.
I troubleshooted in two ways.
First, I tried setting the environmental variable R_BROWSER=FALSE in my .Rprofile (to suppress showing URLs altogether).
Second, I commented out #readLines(paste0(output_file, "_result_url"), warn = FALSE) in the function and had it return a number. This produced another error message.
I think the issue is the render function in rmarkdown.

Can I use this function or your package to create / publish pages without R trying to open the page in a browser?

Getting an error when trying to publish. Not sure if related to #45 ?

Listening on http://127.0.0.1:7685
Warning: Error in confl_verb: Error in stop(httr::http_condition(res, type = "error"), httr::content(res)): Bad Request (HTTP 400).
400list(authorized = FALSE, valid = TRUE, errors = list(), successful = FALSE)com.atlassian.confluence.api.service.exceptions.BadRequestException: Error parsing xhtml: Unexpected close tag </ac:structured-macro>; expected </p>.
 at [row,col {unknown-source}]: [22,21]
  78: stop
  77: confl_verb
  76: confl_update_page
  75: observeEventHandler
   4: shiny::runApp
   3: shiny::runGadget
   2: confl_addin_upload
   1: conflr::confl_create_post_from_Rmd

Given that html gives access to richer content than md (interactive graphs etc...)
Is it a goal to be able to publish html to confluence ?
Thanks

I looks like when I use data.table in the Rmd that I want to push on confluence I get issues with data.table. Typically the attributes of the data.table might be messed up.
I add similar issues before when writing my own packages and I had to add Depends: data.table in DESCRIPTION

After knit the rmarkdown, I don's see the options type, space key and parent page in the preview window as mentioned in the below link.
https://engineering.linecorp.com/en/blog/conflr-published/

This might be good to have, but this requires non-default macro. What's tougher is that the type of macro is different between the Server version and the Cloud version...

• Deck of Cards macro: https://docs.servicerocket.com/composition/macro-reference/deck-of-cards
• Tabs macro: https://docs.servicerocket.com/composition/macro-reference/tabs

<ac:structured-macro ac:name="deck">
  <ac:parameter ac:name="id">card1</ac:parameter>
  <ac:rich-text-body>
    <ac:structured-macro ac:name="card">
      <ac:parameter ac:name="label">aaa</ac:parameter>
      <ac:rich-text-body>
        <p>aaa</p>
      </ac:rich-text-body>
    </ac:structured-macro>

    <ac:structured-macro ac:name="card">
      <ac:parameter ac:name="label">bbb</ac:parameter>
      <ac:rich-text-body>
        <p>bbb</p>
      </ac:rich-text-body>
    </ac:structured-macro>
  </ac:rich-text-body>
</ac:structured-macro>

As we keep adding many features, I think it's a matter of time until they get diverged and behave inconsistently. Let's figure out the steps needed and merge these two files.

Add this tag

<ac:structured-macro ac:name="toc" />

c.f. #7