Jekyll allows data to be specified in YAML, CSV, and JSON format in the =_data= dir.
If the data is an array, it is straightforward to build an index page, containing all records, using a Liquid loop. In some occasions, however, you also want to generate one page per record. Consider, e.g., a list of team members, for which you want to generate an individual page for each member.
This generator allows one to specify data files for which we want to generate one page per record.
Among the advantages:
general purpose: it works with any array of data: people, projects, events, ... you name it
it manages multiple data sources in the same website
Installation
Option 1. Add =gem "jekyll-datapage-generator"= to your project Gemfile and then load it as a plugin in the configuration:
plugins:
(See https://jekyllrb.com/docs/plugins/installation/ for more details.)
Option 2. Download =jekyll-datapage-generator.rb= and put it in the =_plugins= directory of your website. /Remember to delete older versions of the source file from the =_plugins= directory or you might run into errors./
Read the [[example]] section or the included example to get started quickly.
You have some data in a data file for which you want to generate individual pages.
The specification in =config.yml= looks like:
page_gen-dirs: [true|false]
page_gen:
data: [name of a data set in _data] template: [name of template in _layouts: will be used to render pages] dir: [directory where filenames will be generated] index_files: [true|false] name: [field used to generate the filename] name_expr: [a Ruby expression to generate the filename (alternative to name)] title: [field used to generate the page title] title_expr: [a Ruby expression to generate the filename (alternative to title)] extension: [extension used to generate the filename.] filter: [property to filter data records by] filter_condition: [a Ruby expression to filter data] page_data_prefix: [prefix used to name variables] debug: [boolean, if true output more informative messages]
[another set of data, if you want ...]
where:
~page_gen-dirs~ specifies whether we want to generate named folders (~true~) or not (~false~) for all generated data.
~page_gen~ is an array specifying a data for which individual pages have to be generated.
Each entry in ~page_gen~ can (in some cases must) contain the following fields:
~data~ is the name of the data file to read (YAML, Json, or CSV). Use the full path, if your data is structured in a hierarchy. For instance: ~hierarchy.people~ will loop over the variable ~people~ in the ~_data/hierarchy.yml~ file.
~template~ is the name of a template used to layout data in the page. /Optional, if not specified, the plugin uses the layout with the same name of the value of ~data~./
~dir~ is the directory where pages are generated. /Optional: if not specified, the generator puts pages in a directory with the same name of the value of the ~data~ field./
~name~ is the name of a field in the data used to generate a filename. You need to ensure values in the chosen field are unique, or some files will get overwritten.
~name_expr~ is an optional Ruby expression used to generate a filename. The expression can reference fields of the data being read using the ~record~ hash (e.g., ~record['firstname'] + "" + record['last_name']~; see the documentation of ~filter_condition~ for more details.) /Optional: used in alternative to ~name~. If set it overrides ~name~/
~index_files~ specifies whether we want to generate named folders (true) or not (false) for the current set of data. /Optional: if specified, it overrides the value of the global declaration ~page_gen-dirs~./
~title~ is the name of a field in data which is used for the page title. /Optional: if not specified, the generator uses the filename as the page title./
~title_expr~ is an optional Ruby expression used to generate the page title. The expression can reference fields of the data being read using the ~record~ hash (e.g., ~record['firstname'] + "" + record['last_name']~). /Optional, but if set, it overrides ~title~./
~extension~ is the extension of the generated files. /Optional: if not specified, the generator uses ~html~ extension./ If set to false, the file will be generated without an extension
~page_data_prefix~ is the prefix used to output the page data. Data read from each record is made available in the page so that it can be accessed using liquid tags. In some cases, however, there might be clashes with existing tags. ~page_data_prefix~ can be used to prefix all data read from records and avoid the problem mentioned above. /Optional: if not specified, no prefix is used./
~filter~ is a property of each data record that must return a true-ish value for the record to be included in the list of files to be generated. See [[Filtering Data]], below, for more details. /Optional: if not specified, all records from the dataset are included (see also ~filter_condition~)./
~filter_condition~ is a string containing a Ruby expression which evaluates to a true-ish value. The condition can reference fields of the data being read using the ~record~ hash (e.g., ~record['author'] ~~ 'George Orwell'~). See [[Filtering Data]], below, for more details. /Optional: if not specified, all records from the dataset are included (see also ~filter~)./
~debug~ is a Boolean value specifying whether the plugin will output information about the configuration and data read. /Optional: if not specified, no debug information is outputted./
Note. The same data structure can be referenced different times, maybe with different target directories. This is useful to group pages in different directories, using ~filter_condition~.
A liquid tag is also made available to generate a link to a given page. For instance:
{{ page_name | datapage_url: dir }}generates a link to ~page_name~ in ~dir~.
There are three different ways which you can use to show only the relevant records of a data structure in your website:
** Do not link uninteresting pages
Generate pages for all records (relevant and not), but link only the interesting pages.
The uninteresting pages will still get generated but will not be easily accessible. A visitor has to guess the URL to access them. This is more of a workaround, rather than a solution.
This is shown in the ~books.md~ file, in the section "Books I have read".
The filter is applied to the links to tha generated pages. Pages will still be generated for all books, but only those for which ~book.read~ is true will be easily accessible (since only these have an explicit link in our website).
** Use the ~filter~ condition
Use the ~filter~ property.
In this case, all records in your data structure should have a boolean field, let us say, ~publish~. Pages will be generated only for those records in which the ~publish~ field is true(-ish).
Consider the following declaration in ~_config.yml~:
In this case, a page will be generated only for the books in which the field ~read~ is ~true~.
** Use the ~filter_condition~ condition
Use the ~filter_condition~ property.
The field should contain a string which evaluates to a boolean expression. The
string may reference fields of the data structure using the
~record[
In this case pages will be generated only for the records satisfying the evaluation of the ~filter_condition~.
Example 1. Consider the following declaration in ~_config.yml~:
that allows me to generate a list of the books I have not read. The ~filter~ keyword, in this case, is no good, since I need to test for falsity (~read~ has to be false).
The filter condition allows to select only those records in which ~record['read']~ is false.
Remark If you want to filter on nested fields, use multiple ~[]~. For instance:
filter_condition: "record['did-i']['read'] ~~ false"
works with the following data structure:
Example 2. Consider the following declaration in ~_config.yml~:
In this case, I am testing the ~author~ field and generating pages only for the books by George Orwell.
As a final consideration, ~filter_condition~ allows one to deploy pages in different directories according to specific properties.
Consider the following example:
which splits the ~book~ data structure in two different folders, according to the value of the ~read~ flag.
Of course, such an approach makes sense only for variables with a limited number of values, since one needs to explicitly specify in ~_config.yml~ conditions and target directories.
You can generate filenames with an expression, by replacing ~name~ with ~name_expr~. For example, if you have data in a .yml file that looks like this:
- first_name: adolfo
last_name: villafiorita
bio: long bio goes here
- first_name: pietro
last_name: molini
bio: another long bio
- first_name: aaron
last_name: ciaghi
bio: another very long bioYour ~_config.yml~ could contain the following:
page_gen:
Alternatively, you could have a ~members.json~ (or a ~members.csv~ file) stored in the ~_data~ directory with the following content and the example would work the same:
[ { "name": "adolfo villafiorita", "bio": "long bio goes here" }, { "name": "pietro molini", "bio": "another long bio" }, { "name": "aaron ciaghi", "bio": "another very long bio" } ]
{{page.bio}}
page_gen:
Then, when building the site, this generator will create a directory ~people~ containing, for each record in ~members.yml~, a file with the record data formatted according to the ~profile.html~ layout. The record used to generate the filename of each page is ~name~, sanitized.
$ cd example $ jekyll build $ cat _site/people/adolfo-villafiorita.html
long bio goes here
Check the example directory for a live demo. (Notice that the ruby file in ~_plugins~ is a symbolic link; you might have to remove the link and manually copy the ruby file in the ~_plugins~ directory, if symbolic links do not work in your system.)
See the [[file:CHANGELOG.org][CHANGELOG]] file.
Run successfully at least once with the following Jekyll versions: 4.2.1., 4.0.1, 3.8.5, 3.6.2, 3.1.6. Try with the included example and open an issue if you find any compatibility issue.
[[http://ict4g.net/adolfo][Adolfo Villafiorita]] with several excellent [[https://github.com/avillafiorita/jekyll-datapage_gen/graphs/contributors][contributions from various authors]].
Some known bugs and an unknown number of unknown bugs.
(See the open issues for the known bugs.)
Distributed under the terms of the [[http://opensource.org/licenses/MIT][MIT License]].