I'm trying to publish webpage using org-mode. Two questions:
Is there a way to "sync" the org-mode files in the base-directory and the html files in the publishing-directory? Specifically, if I delete an org file in the base-directory, can I get org-publish-html to delete the corresponding file in the html directory also?
If I have pages within subdirectories, how can I specify a single .css file in the root directory to be used for the style sheet? For instance, my directory structure is as follows:
public_html/
css/
mystyle.css
index.html
subdir/
index.html
With the following specifications in org-publish-project-alist (this is just a subset) --
:publishing-directory "public_html"
:style "<link rel=\"stylesheet\" href=\"css/mystyle.css\" type=\"text/css\"/>"
mystyle.css is used by public_html/index.html but not by public_html/subdir/index.html. Is there a simple remedy to this (I want the style sheet to be used by both/all files in subdirectories)?
Thanks much ~
There is no straightforward way of doing this. Org-mode doesn't know (or care) about the location to which it is publishing - it just sends things there and makes sure the correct directory structure exists. There is a hook in the publishing process that gets called after the files have been pushed to their published location. This is controlled by setting the :completion-function property in your org-publish-project-alist. You could use this hook to write a function that compares the *.org files in your base-dir and subdirectories to the accompanying *.html published files, and remove those *.html files that don't have an accompanying *.org file.
I suspect this will be most easily accomplished by making your Lisp completion-function call a shell script that removes the necessary files. If you are doing something fancy with the :include, :exclude, or :base-extension properties, you'll likely want your completion-function to grab the pertinent information from the plist and then pass them to your shell script. This org-mode page has an example completion-function that shows how to get property values for the org-publish-project-alist. You would then need to pass them to your shell script.
There are several ways to do this. Perhaps the simplest is to just override the default style sheet in each file with a line such as:
#+STYLE: <link rel="stylesheet" type="text/css" href="../stylesheet.css" />
for your first level of subdirectory files, and keep adding ../ as you get deeper in the directory structure.
Another possibility is generate generic template files for each level within the directory tree. This org-mode page gives a nice example of how to set this up.
Lastly, another option is to use the :preparation-function property of org-publish-project-alist to define a function that will automatically change the style file for each file. Again, this is probably best done by having the Lisp preparation-function call a shell script to parse the files. I could imagine doing this with the Unix sed program to find a regular expression denoted something like href="#MYLOC#/stylesheet.css" /> and substitute the stuff between #'s with the appropriate level within the directory tree. This seems like overkill, given the other options.
Related
I'm using the Gulp to build my SCSS, Pug and ES6 assets for my static website. I know it's possible to hash file names and output the files in a different directory.
For my specific example:
my Pug markdown are found in the ~/src/pages directory and getting built to the ~/public/ directory.
My SCSS stylesheets are found in the ~/src/stylesheets directory. These are getting built to the and getting ~/public/style directory
My problem is, when I'm referring to my stylesheets files from Pug, I have to refer to the already-built folder like this:
link(rel='stylesheet', href='./style/example.css')
For my IDE, this doesn't make sense, because the style directory doesn't exist in the ~/src/pages directory.
What I would find the most useful is that I can refer to my stylesheets like the example below:
link(rel='stylesheet', href='../stylesheets/example.scss')
Is there any way this is possible or am I completely going in the wrong direction? If not, where am I looking for?
Solution to make the file name like hash
gulp, for automating our task
gulp-rev, for renaming our files with random hashes.
gulp-rev-collector, for switching non-hashed references by hashed-references inside our files.
rev-del, for deleting non-hashed files in our /dist folder.
Sample code :
gulpfile.js
gulp.task("revision:rename", ["serve"], () =>
gulp.src(["dist/**/*.html",
"dist/**/*.css",
"dist/**/*.js",
"dist/**/*.{jpg,png,jpeg,gif,svg}"])
.pipe(rev())
.pipe(revdel())
.pipe(gulp.dest("dist"))
.pipe(rev.manifest({ path: "manifest.json" }))
.pipe(gulp.dest("dist"))
);
manifest.json
{
style.css: style-ds9udjvci.css,
main.js: main-dijds9xc9.min.js
}
For creating our revision update in the file like
Rewrite every reference for every key of manifest.json to it’s respective value inside every html/json/css/js file (i.e: <link href="style.css"> would become <link href="style-ds9udjvci.css">)
gulp.task("revision:updateReferences", ["serve", "revision:rename"], () =>
gulp.src(["dist/manifest.json","dist/**/*.{html,json,css,js}"])
.pipe(collect())
.pipe(gulp.dest("dist"))
);
You can use something like gulp-watch for real-time compiling of your .scss files, then your /style/example.css file will exist and it will be recompiled automatically when you modify example.scss
You may need to move some directories around to get everything to link, but you can use watch to build your Pug files too, so your site will always be up to date.
Basically, you make a change on any file in your project and view the update live.
Gulp cannot automatically change the file paths used inside the htmls. Therefore you will have to use the generated file path for accessing the style files.
Although if you want to have the file path as the folder structure of your scss, then you will have to replace the contents of the pug file after gulp has finished converting it to HTML.
You can convert the html to String and use the .replace method to replace whatever content you want to change and finally parse the string to a HTML document.
Hope this helps!!
I first met html-minifier today after running a small site I've created using Hugo through Google PageSpeed.
First thing I noticed is that although it does have recursion capabilities it stops working on unsupported files like images (my speakers started beeping and I freaked a little)
I've found this stack showing an apparently undocumented command-line option --file-ext
That worked perfectly but in the output directory, I noticed that the folders with the unmatching contents were gone.
From the directory root, I saw it was Hugo's folders for CSS, JS, images and Github Pages' CNAME file. Not only I can't tell for sure there's not even one piece of static file in any of the folders Hugo generated (you may know that Hugo is sometimes unpredictable) but also I would like to keep language specific XML Sitemaps I've created for some specific folders.
Long story short, is there a way to copy-over unmatching files "as is", keeping input directory ready for a commit/push?
After analyzing the whole directory structure I could be sure that within all the directory structure Hugo creates there are nothing more than HTML and XML files so then the Ockham's Razor took place.
Since both my Hugo's source code and output contents are in totally different directories, it was a simple matter of pointing the output directory to the same path of the input directory.
All HTML files are minified, overwriting those Hugo generated.
I have a folder with bla.html file and some assets (*.png and *.json) it it.
I want doxygen to copy it into HTML documentation root.
How to make it do such thing in doxygen configuration file (not using external script)?
Upon request:
How about (from the documentation / Doxyfile):
HTML_EXTRA_FILES The HTML_EXTRA_FILES tag can be used to specify one
or more extra images or other source files which should be copied to
the HTML output directory. Note that these files will be copied to the
base HTML output directory. Use the $relpath^ marker in the
HTML_HEADER and/or HTML_FOOTER files to load these files. In the
HTML_STYLESHEET file, use the file name only. Also note that the files
will be copied as-is; there are no commands or markers available. This
tag requires that the tag GENERATE_HTML is set to YES.
The full doxygen documentation can be found at : http://doxygen.nl/manual/index.html and the referred part at: http://doxygen.nl/manual/config.html#cfg_html_extra_files
All my projects contains alot of files with the same name 'file1.php file1.less file1.css file1.min.js file1.js file1.json'
Is there a plugin or a way to group all those files with same basename under eg. file1.php, just like the filewatchers does it.
The grouping relations can only be established when running watchers, so using watchers is the right way to go. It can indeed be a batch script that does nothing - this is the 'Output paths to refresh' option that matters: it should provide a pattern for files that has to be nested. For example, let you have same-named files with .php, .css and .js extensions in some folder, and like to nest the latter 2 in .php file (doesn't make much sense, but it's just an example).
create a .bat/sh file that does nothing (echos some string, etc.)
create a new custom file watcher, set 'php' as file type, your .bat - as a program
specify $FileName$ and $FileDir$ as 'Arguments' and 'Working directory' respectively
set $FileNameWithoutExtension$.css:$FileNameWithoutExtension$.js as 'Output paths to refresh'
Now when you modify your phpfile watcher will be run and nest .js and .css in .php
Currently it is not possible to perform such nesting/grouping manually (excluding File Watchers).. and I personally know no plugin that can do this.
Watch these tickets (star/vote/comment) to get notified on progress:
https://youtrack.jetbrains.com/issue/IDEA-113347
https://youtrack.jetbrains.com/issue/WEB-7635
I am creating a Zip file from a folder (and subfolders). it works fine and creates a new .zip file also but I am having an issue while using glob.glob. It is reading all files from the desired folder (source folder) and writing to the new zip file but the problem is that it is, however, adding subdirectories, but not adding files form the subdirectories.
I am giving user an option to select the filename and path as well as filetype also (Zip or Tar). I don;t get any problem while creating .tar.gz file, but when use creates .zip file, this problem comes across.
Here is my code:
for name in (Source_Dir):
for name in glob.glob("/path/to/source/dir/*" ):
myZip.write(name, os.path.basename(name), zipfile.ZIP_DEFLATED)
myZip.close()
Also, if I use code below:
for dirpath, dirnames, filenames in os.walk(Source_Dir):
myZip.write(os.path.join(dirpath, filename) os.path.basename(filename))
myZip.close()
Now the 2nd code taks all files even if it inside the folder/ subfolders, creates a new .zip file and write to it without any directory strucure. It even does not take dir structure for main folder and simply write all files from main dir or subdir to that .zip file.
Can anyone please help me or suggest me. I would prefer glob.glob rather than the 2nd option to use.
Thanks in advance.
Regards,
Akash
Glob by design does not expand into subdirectories. It follows UNIX style path rules and expansions see the documentation for fnmatch for more information. If you want to get at the subdirectories you need to add it to the path. This example will get everything at one level down.
for name in (Source_Dir):
for name in glob.glob("/path/to/source/dir/*/*" ):
myZip.write(name, os.path.basename(name), zipfile.ZIP_DEFLATED)
myZip.close()
Doug Hellman has an excellent discussion here. If you are not using the pattern features of glob (like *.txt for all text files or *[0-9].txt for all text files that have a number before the extension) then I think your os.walk solution is better