Lander

- Pin to metasrc 0.1.3
- Via metasrc, Lander has improved LaTeX source processing, including handling of referenced source files (`\input` and `\include`) and macros (`\def` and `\newcommand`).
- Improved treatment of draft status.
The heuristic is that a document is considered a draft if the branch is not `master` and `lsstdraft` is not present in a lsstdoc document's options.

- Fix logic for determining it Lander is running in a Travis PR environment.
- Log the Lander version at startup.

- Fixed Travis deployment issue. Used `skip_cleanup: true` to `.travis.yml` to prevent CSS and JS assets from bring cleaned up before creating a release.

- Detect if running from a Travis PR build (using the `TRAVIS_PULL_REQUEST` environment variable) and if so, abort the page build and upload.
This is to prevent duplicate uploads from both branch and PR-based Travis jobs.
- Pin inuitcss to 6.0.0-beta4 because of the removal of rem functions in beta5.

- Update to `metasrc>=0.1.1,<0.2`.
- Use `remove_comments` and `remove_trailing_whitespace` feature from metasrc.
This improves the accuracy of metadata extraction from tex source.
For example, comment characters won't appear in extract abstract content.

Initial version.

- Native PDF display via [PDFObject](https://pdfobject.com).
- Multi-level metadata model for populating the landing page.
Lander uses https://github.com/lsst-sqre/metasrc to extract content from the `tex` source (more work on this is needed).
Otherwise, it gets data from environment variables (including Travis CI variables).
Finally, metadata can be specified explicitly with command line arguments.
- Uses https://github.com/lsst-sqre/squared for CSS, icons, and logos.
A Gulp and webpack workflow build these sources into deployable dependencies.
Gulp and webpack are only needed by developers and CI; the deployable artifacts are included in PyPI releases.
- Release workflow is fully automated in Travis.
Create a PEP 440 version tag, push it, and the release appears in PyPI.
- Includes an LSST the Docs upload client (via https://github.com/lsst-sqre/ltd-conveyor) built-in so that https://github.com/lsst-sqre/ltd-mason isn't required.
- Usage and development docs are currently in the README.