Ploomber

* Fixes look up of conda root when running `ploomber install` when conda binary is inside the `Library` directory (Windows)
* No longer looking up pip inside conda when running `ploomber install` and `setup.py` does not exist
* Adds `--use-lock/-l` option to `ploomber install` to install using lock files

* Simplifies serializer and unserializer creation with new decorators
* Adds guide on serializer/unserializer decorators to the documentation

* Clearer error message when failing to import function
* Better error message when `tasks` in YAML spec is not a list
* Fixes an issue that caused dag plot to fail when using `.svg`
* Fixes duplicated log entries when viewing a file in Jupyter

* Fixes cell injection when using the `--notebook-dir` during Jupyter initialization
* Reduces verbosity in Jupyter logs ([314](https://github.com/ploomber/ploomber/issues/314))
* Adds `tasks[*].params.resources_` to track changes in external files
* Minor bug fixes

* Lazy load for `serializer`, `unserialize`, DAG clients, Task clients, Product clients, and task hooks, which allows the Jupyter plugin to work even if the Jupyter process does not have the dependencies required to import such dotted paths
* CLI `--help` message shows if `ENTRY_POINT` environment variable is defined
* `ploomber scaffold` now takes a `-e/--entry-point` optional argument
* Fixes error that caused the `{{here}}` placeholder not to work if an `env.yaml` exists
* Adds `--empty` option to `ploomber scaffold` to create a `pipeline.yaml` with no tasks

- Allowing `pipeline.yaml` at project root if setup.py but `src/*/pipeline.yaml` is missing
- Fixes bug in `EnvDict.find` that caused the `{{here}}` placeholder to point to the `env.yaml` file instead of its parent
- `DAGSpec._find_relative` returns relative path to spec
- Fixes error that missed `env.yaml` loading when initializing DAGSpecPartial