Update links to the Katacoda tutorials to point to the official versions
at https://katacoda.com/tektoncd and update the dashboard scenario naming
to be consistent with other (i.e. drop the redundant `tekton-` from the
name + URL).

Update the Katacoda tutorial playground to the latest available releases:
- Pipelines 0.20
- Dashboard 0.13
- CLI 0.15

Update the Dashboard katacoda tutorial to use the latest releases:
- Dashboard 0.13.0
- Pipelines 0.20.0

Update screenshots to reflect the new design in recent dashboard versions.
While doing this I converted the images to JPEG for significant filesize
savings over the existing PNG images (>50% on average)

The website header partial sets the meta index, follow if the
HUGO_ENV is set to production, and no index no follow otherwise.

Change the HUGO_ENV to be production for the production environment.

Related to #191
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

The baseURL is not set for dev and prod which causes the following
issues:
- for dev, the URL shown on the console is not the correct one
- for prod, the generated sitemap.xml uses relative URLs, which
  makes it not parsable by the search engine

This is already setup correctly for deploy-previews.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

fixing the overview section markdown metadata so it is visually consistent with the other sections of the doc

There is a chance that utf-8 encoded text may be decoded
successfully but wrongly as ascii. UTF-8 instead should always be
able to decode ascii successfully, so let's try UTF-8 first.

Fixes: #211
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Add the new version and add some docs that were previously excluded:
- deprecations
- runs
- tekton bundles contract

Add an option to update caches.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

In case of non-md files in nested folders, the filename used in the
rewrite included the target path leading to double-path in the link.
Fixed that and added a use case for the non-md file in a subfolder.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Change the sync configuration from files to folders, so make the
configuration easier to maintain as well to allow syncing folders
of generated docs, where the list of files may vary and for which
it is not possible to pre-add the YAML headers used by docsy.

This change implies the following modification to the sync process:
- we do not download files one-by-one to disk and rewrite them to
  a different location on disk. Instead we git clone all relevant
  repositories, and pull files at the correct reference into the
  site content folder direcly
- we now support pulling content from different folders in the
  tekton repos. As part of the configuration a target folder in
  the website can be specified. The default is '/'.
  This will allow pulling in content from other folders that may
  not even be under `docs`.

As part of the re-work, a couple of issues have been fixed:
- links to the index page were re-written to github because the
  `_index.md` could not be found locally in the source. This case
  is now handled by the code
- links to files in nested folders were re-written to github
  because the files were downloaded to the parent folders and
  thus not found locally. Files configured from a nested folder
  are now downloaded to the configured target folder and the
  check has been fixed.
  In one case this issue had the side effect of fixing the broken
  link in content/en/vault/Triggers-v0.7.0/exposing-eventlisteners.md:
  - [`create-ingress`](./create-ingress.yaml)
  the YAML file is not in the `docs` folder so the link should be:
  - [`create-ingress`](./getting-started/create-ingress.yaml)
  which would have to be fixed on triggers side, with a new minor
  release.

Overall the execution time ("real") has been improved (real), at
least on my laptop. The "user" time is bit more so on systems
with fewer CPU cores it may take slightly longer.

- before:
real	0m41.870s
user	1m52.244s
sys	0m29.440s

- after, no cache:
real	0m31.981s
user	2m25.869s
sys	0m46.176s

- after, with cache:
real	0m21.382s
user	2m24.826s
sys	0m42.637s

Added more unit test coverage (75% now).
To verify this change I also diffed the content generated before
this change with the content generated after the change.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Instead of ignoring the whole content folder, only ignore folders
that we sync. Add a few more ignore files for python and IDEs.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

We may need to remove versions: when releasing a minor version
we might want to remove previous releases on the same major version.
Adding a version rm command for that.

As part of the process, move away from absl-py and use click instead
that helps handling different commands.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Pipeline release v0.17.3 which supercedes v0.17.2.
Updating the docs accordingly.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Use a simple scheduled github action to deploy
the website nightly. The deploy step includes sync,
so this will pull in an up to date version of the docs
on the master branch.

When an older version is selected, we still want to be able
to select the latest version. Alter the version switcher to
allow for that.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Set the latest version of each component as the current one, add
everything else, including master, to the vault.
Remove the _index.md files that are brought in via sync.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Hugo produces lowercase anchors into rendered markdown pages.
If a fragment points to a markdown page, make it lower case.
We do not touch fragments on external URLs or on github files
that are not markdown.

Fixes #159
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

I think it's safe to say that Napoleon has gone above and beyond
anything we could have asked for help on this repo.

Thanks for everything, and here's to more!
Signed-off-by: NJJ Asghar <jjasghar@gmail.com>

For relative URLs that are not matched to a file on disk, we
rewrite them to a github URL. The base URL is joined to the
path with urljoin. If the the path includes a `..` then
urljoin goes up one level from the last `/` in the base URL.
Because of that, we need the base url to end with a `/`.
Added tests to avoid this from breaking in future.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

The previous logic parses URLs with a regex first and then using
the markdown library. It then uses urlopen().read() to validate
links.

We use now the markdown library only to extract the list of links,
and then urlparse to deconstruct, analyse, adapt and reconstract
the link. We do not attempt to fetch links anymore, which means
that external links are not guaranteed to be working.

Absolute URLs are not changed (they may be external)
Fragments are relative to the page and do not need changes
Path only links should point to a file synced to the website
but sometimes the file may be missing (if it's not in the sync
configuration), so we follow this approach:
- prefix with base_path and check for the file locally
- if not found, prefix with base_url instead

Note that urlparse treats URLs without scheme like path only
URLs, so 'github.com' will be rewritten to base_url/github.com
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Most of the time required for a sync is spent in two tasks:
- wget of all the original files
- validating the URLs with urlopen().read()

Running those two tasks in parallel greatly reduces the
processing time. This will be even more relevant when syncing
files from multiple releases.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Add a new python script that can be used to add a new version to
the configuration of a component.
This script can be used by various projects as part of their
release process to keep the version on the website up-to-date.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Cleanup typos, simplify logic in a few places.
Make the list of files config a dict instead of a list of dicts.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

As a developer, it is difficult to make changes to a script and ensure that nothing is broken without tests. Test are added in this patch with 80% code coverage. The script has also been refactored to increase readability. This refactor is not a pure refactor. The sync did not work properly. There is a list of bugs that reported the script generating broken links, #160, #126, #158, #133, #134. This refactor also includes a patch that fixes these bugs. The patch redirects a user to github if there is a relative link to a file that does not exist. This is important because some links will reference a snippet of code, test, or example. These artifacts are not docs, they are not rendered properly on the site. It is better to redirect the user to the intended artifact if the file is not synced. This prevents random 404s, and preserves the original behavior of the sync script. There are other approaches to fix this bug. This includes, manually editing links, downloading the files that the broken link are referencing in the sync config file, or removing the broken links. This is the best approach because broken links are automatically fixed and do not require manual intervention. This approach causes a side effect of not requiring the config file to be up-to date because the script will redirect the user to github. We want to encourage frequent sync. So eventually, this script should become tekton-ified (tekton task) and placed into the tekton catalog repo. Sync needs to become a tekon task in order for every repo to sync docs on push event.

The version switcher dropdown box mentiones releases, however currently
it refers to "master" which is not a release at all. I think we could
use the wording "version" which is more generic.

Archived versions instead is a link to releases on github. Changing the
wording accordingly.
Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

Signed-off-by: NAndrea Frittoli <andrea.frittoli@gmail.com>

The modal in the interactive tutorials is difficult to navigate. I can not scroll (left to right) and (up and down) because of the embedded iframe.

Publishes a comprehensive overview of Tekton.

Instead, assume a cluster already exists, or briefly document using
`kind` to set up a local cluster for testing.

Update all Katacoda scenarios to run on Kubernetes 1.18, with
a 2 node cluster.

Fix some issues in the tasks and runs scenarios that seem to have
been undone in rebases before they were merged.

Update the Dashboard scenario to run the latest Pipeline and
Dashboard versions at time of writing.

Remove the use of the ingress in the Dashboard scenario since
it appears unstable on Katacoda with kubernetes 1.18. Using a
port-forward is sufficient for this scenario.

Add a Katacoda scenario that can act as a playground for users
to experiment with Tekton in a sandbox environment.

This environment provides a 2-node Kubernetes cluster with the
following projects installed:
- Pipelines v0.14.3
- Dashboard v0.8.2
- CLI v0.11.0

Typical startup time is <1minute, but may take 2 or 3 minutes
in some cases. The installation should not require any manual
steps by the user.

without --dry-run=client option, the create & replace command will fail.

Katacoda supports linking to images in the assets folder. Update our
existing links to use this approach instead of using the GitHub 'raw'
URLs as this gives us greater portability for the tutorials and ensures
the correct version of the images is loaded even if we stop loading content
from the `master` branch in future.

Tekton is a cloud native project, not necessarily a kubernetes native
project, and in fact (beyond basing our design around k8s style
resources and using k8s containers) we try to minimize our coupling to
k8s (at least in our API).

If you look at our README you can see that we are k8s-style and cloud
native: https://github.com/tektoncd/pipeline#-tekton-pipelines

You can see this reflected in our mission as well:
"Be the industry-standard, cloud-native CI/CD platform components and
ecosystem."
https://github.com/tektoncd/community/blob/master/roadmap.md#mission-and-vision

This commit updates the wording to be a bit more accurate.

- Moving the code from the decentralized to the centralized
- Create .vimrc in the background to cut down noise in the terminal
- Add a short delay before calling launch.sh to avoid intermittent failures with launch.sh not being ready
- Fix pipeline definition to ensure deploy runs after build
- Fix `kubectl expose` command since there's no deployment involved, just a pod
- Fix some lingering bad PipelineResource versions
Signed-off-by: NJJ Asghar <jjasghar@gmail.com>
Co-authored-by: Nmichaelawyu <michaelawyu@google.com>
Co-authored-by: NAlan Greene <alangree@ie.ibm.com>

Some of the Triggers content was moved in https://github.com/tektoncd/triggers/pull/642,
specifically: create-ingress.yaml and create-webhook.yaml

These have been moved to the getting-started folder which broke
the website sync script, blocking other changes.

Update the sync config to point to the new paths to unblock
other changes.

down in the docs