Bigfunctions

BREAKING CHANGES

**The documentations generation mecanism CHANGED for the page listing bigfunctions.**

BEFORE

- We used to define bigfunctions categories in each bigfunction yaml file
- The categories to show in the page listing bigfunctions were loaded from mkdocs.yml config.


NOW
- There is no category field in bigfunction yaml file anymore
- The bigfunctions yaml files can be put in subfolders.
- These subfolders are like the old categories, they enable to group functions
- By default, all subfolders are included in the page listing bigfunctions
- You can configure this in the front matter of bigfunctions/README.md.

Enabled the possiblity to put functions in subfolders

- Add `bigfunction_dataset_location` variable available in python function

Small fixes:

- `bigfun test` for python function correctly uses `init_code`
- doc: fix the replacement of `{BIGFUNCTIONS_DATASET}` in documentation
- If you share your dataset to `allUsers` instead of `allAuthenticatedUsers`, it will work too share the function to `allAuthenticatedUsers`

**BUGFIX**
fixed package_data issues in setup.py

This release brings the following features compared to v0.7:

**New Features** ~

- dockerfile base image is now configurable with dockerfile.image variable in yaml files.
- `bigfun get` works with any url and not only bigfunction name from bigfunctions repo.
- New type of function _aggregate function_ available
- In remote function `user_project` variable is available and set to the project used by the user to trigger this function.
- Cloud Run default settings changed to 512 MB for memory and `all` as ingress (to prevent hard to understand failures in function execution)

**Bug Fix**
- Fixed bad dependencies