django is not included in pip-chill output

[fabiocaccamo]

Environment

• PIP Chill version: 1.0.1
• Python version: 3.9.0
• Operating System: OSX

Description

I'm using pip-chill on a django project to automate requirements.txt generation (without dependencies) and deploy, but pip-chill doesn't recognize django as top-level installed package.

What I Did

Running pip freeze I can see that Django==3.2 is installed correctly, but running pip-chill the django dependency is not included in the output.

[github-actions]

Stale issue message

[fabiocaccamo]

Sure if any maintainer give a feedback the issue become stale.

[astarrh]

I am experiencing the same issue. I'm not sure if Django is considered a dependency in one of my other packages? Even if it is, it's important that I specify the version in my requirements.

These is my pip-chill output:

babel==2.8.0
channels==3.0.4
django-livereload-server==0.3.2
django-mail-factory==0.21
django-newsletter==0.9.1
django-tinymce==2.8.0
django-webpack-loader==0.6.0
pip-chill==1.0.1
psycopg2==2.7.5
psycopg2-binary==2.7.5
pyjwt==1.7.1
pyopenssl==20.0.1
python3-openid==3.1.0
requests-oauthlib==1.0.0
service-identity==21.1.0

[rbanffy]

Sorry for the (very) late answer. I'm not sure if I understood it (and #58) correctly.

The behavior of not listing a manually installed package that is a dependency of another installed package is the intended one. If you need, say, django-newsletter, it's safe to assume you need Django and doesn't need to specify it.

However...

The consequence is that if you want to pin Django's version to a specific one, something that your dependency doesn't do, you'll not still get the django package in the chill results.

Adding a --force-listing option with comma-separated package names, that forces listing that package even if it is a dependency of something else, could be a solution.

Let me know @fabiocaccamo , @astarrh , @GammaGames, @jc9108 what your thoughts are.

I added and enhancement request as #61

[GammaGames]

I know what packages it's happening to so a --force-listing option would solve my issues.

[fabiocaccamo]

@rbanffy in my case django is an installed package (python -m pip install django) but for some reason pip-chill doesn't list it.

[411A]

Same problem. Unfortunately, this library does not work as I expected, don't know why this vital feature does not exist in the pip itself. pip-chill not included some packages that I manually installed, like Pyrogram & Pillow; instead, it shows me the packages that I didn't install manually like: isoduration jsonpointer uri-template etc.

[rbanffy]

I noticed some surprising behavior as well. I'll see what can be done besides what a --force-listing would do.

[rbanffy]

@411A Installing jupyter I noticed the issue you mention. From a clean environment,

$ python3 -m venv env2
$ . env2/bin/activate
$ pip install jupyter pip-chill
Collecting jupyter
...
$ pip-chill
fqdn==1.5.1
isoduration==20.11.0
jsonpointer==2.3
jupyter==1.0.0
pip-chill==1.0.3
uri-template==1.2.0
webcolors==1.13

Only jupyter and pip-chill should be shown, but some of the packages you mentioned still appear. There must be something in their metadata that confuses the chill function.

[github-actions]

Stale issue message

[rbanffy]

Chill, bot. Still thinking about this.

[rangel3l2]

I got the same problem with flask.

[rbanffy]

What is your original dependencies and what does pip-chill output?

[rangel3l2]

What is your original dependencies and what does pip-chill output?

I create a venv and installed “pip install flask”, but quem I run “pip-chill > requirements.txt” flask is not inside my dependencies inside requirements.txt file.

[rbanffy]

I can´t replicate this behavior here (Python 3.11).

When I pip install pip-chill flask from a clean virtual environment, running pip-chill gives me:

flask==3.0.3
pip-chill==1.0.3

When running with --verbose, I get:

flask==3.0.3
pip-chill==1.0.3
# blinker==1.7.0 # Installed as dependency for flask
# click==8.1.7 # Installed as dependency for flask
# itsdangerous==2.2.0 # Installed as dependency for flask
# jinja2==3.1.3 # Installed as dependency for flask
# markupsafe==2.1.5 # Installed as dependency for jinja2, werkzeug
# werkzeug==3.0.2 # Installed as dependency for flask

Are you starting from a clean virtual environment? What do you get when you run pip freeze?