Need help with generating GUID/UUID values for context variables

[chechu2k5]

• Cookiecutter version: 1.6
• Template project url: none
• Python version: 3.7 (virtual env created using win python 3.7 x64)
• Operating System: Windows 10, 64 Bit

Description:

First off many thanks for creating this project !
Here is some context of what I am trying to do and where I need some guidance

• I am trying to use CC to generate new a Visual Studio 2017 solution and project files with a particular folder/file organization that I like
• I was able to most of it working but for the below:
• Parts of the above project, solution files involves generating several unique GUIDs
• my first approach was creating a pre_gen_project.py inside the hooks folder and update/create new variables that could be added to the ones loaded from cookiecutter.json or entered by the user
• I was however blocked as I could not figure out how to access the context being used by CC and the jinja2 engine
• I proceeded to go over the many issues on github and found some related ones like the following: Context injection from pre hook / access from post hook #60, run hooks through template allowing for dynamic contexting in pre/post hooks #102, Add ability to inject context in pre-generate hooks #180, Inject context into hooks #288 but no clear answer on how to achieve what I'd like
• I also followed some others issues that suggested creating custom jinja2 extension/filter (Allow Jinja extensions to be defined/bundled in the template repository #944) but I couldnt figure out how or where to put them in the template folder so the cookiecutter.exe can identify them and pick them up
• Lastly, I also tried going over the CC source code and tried to create a new executable from my script (similar to cli.py) that passes the guids via the extra_context to cookiecutter.main(...) but ran into some other problems that I am still trying to figure out

Appreciate any pointers on how I can inject GUID values for the context variables

[glenfant]

Computing new Jinja context variables in the pre hook would be a great feature.

[gmr]

Having a pre-hook for computing values that are injected into the context for use in templates would be very useful.

[agrant3d]

Was also looking to do this exact thing. Fixing the vcxproj in a post hook script is not too hard, but I have other cases where I need to generate unique ids in unstructured files (cpp).

I was surprised when I started using cookiecutter that the scripts are rendered and ran from temp folder, rather than simply importing the script and giving access to the context dict. Seems like that would make it much simpler and more powerful!

[agrant3d]

As a solution, what I do is: in pre_gen_project.py I generate my ids, and write them to a tmp file inside the template. Then I {%- include tmp/IDFILE.txt -%} where needed. In post_gen_project.py I cleanup the tmp folder.

[rymurr]

@agrant3d could you show a bit more of your code? I used pre_gen_project.py to write a file at tmp/id1.txt and I am doing {%- include 'tmp/id1.txt' =%} from file a/b/c.xml and its giving me a TemplateNotFound exception

[agrant3d]

Sure, looks something like this:

import os
import uuid

CURDIR = os.path.realpath(os.path.curdir)
TEMPLATEDIR = os.path.normpath(os.path.join(CURDIR, '..', '{{cookiecutter._template}}'))
TMPDIR = os.path.join(TEMPLATEDIR, '{{'{{'}}cookiecutter.project_slug{{'}}'}}', 'tmp')

GUIDS = ['PROJECT_GUID', 'SOLUTION_GUID']
for guid in GUIDS:
    file_path = os.path.join(TMPDIR, guid + '.txt')
    with open(file_path, 'w') as f:
        f.write("{" + str(uuid.uuid4()).upper() + "}")

The thing you're probably missing is that you need to put the files inside the {{cookiecutter.project_slug}} directory.

[agrant3d]

Note - this method only works when you are working with a local copy of the cookiecutter template. If it's a cached version - you cannot simply write to the directory.

[rymurr]

worked a dream, thanks @agrant3d

[metaskills]

I am trying to write some Lambda CookieCutter projects because AWS SAM uses this project. Though I do like a good challenge, I've not had a fun time at it. Hoping someone can help.

I need to take the project_name and derive folder/file names, as well as Ruby class names to match in my project. I really wish #944 was merged and I'd be able to use some Python inflector code and do what I need in the templates. Of course #60 would have been helpful too. But I've been forced to try the method above using pre/post hooks to write files to include. I cant get it working. I keep ending up with a jinja2.exceptions.TemplateNotFound: tmp/class_name.txt error.

Note - this method only works when you are working with a local copy of the cookiecutter template. If it's a cached version - you cannot simply write to the directory.

So my cookiecutter._tempalte value is gh:customink/lambda-starter-ruby. My guess is that I'd have to write to ~/.cookiecutters/lambda-starter-ruby where the cloned repo is for this to work?

[metaskills]

Confirmed on putting these tmp files in the ~/.cookiecutters directory. My pre_gen_project.py file is basically the inflector code I needed followed by this.

project_name  = '{{cookiecutter.project_name}}'
cc_template   = '{{cookiecutter._template}}'
template_name = cc_template.split("/")[-1]

file_name  = singularize(tableize(project_name))
class_name = singularize(camelize(tableize(project_name)))
dash_name  = singularize(dasherize(project_name))
all_names  = {
  'file_name': file_name,
  'class_name': class_name,
  'dash_name': dash_name
}

import os
import pathlib
from os.path import expanduser
home_dir = expanduser("~")
template_dir = os.path.join(home_dir, '.cookiecutters', template_name)
temp_dir = os.path.join(template_dir, '{{'{{'}}cookiecutter.project_name{{'}}'}}', '_cctmp')
pathlib.Path(temp_dir).mkdir(parents=True, exist_ok=True)

for name in all_names.keys():
  file_path = os.path.join(temp_dir, name + '.txt')
  with open(file_path, 'w') as f:
    f.write(all_names[name])

[RobRoseKnows]

Looks like there's a PR out to fix this by @jonaswre in #1493