Skip to content

Latest commit

 

History

History
1408 lines (1065 loc) · 45.4 KB

README.md

File metadata and controls

1408 lines (1065 loc) · 45.4 KB

ArangoDB Documentation

This repository contains the source files of the ArangoDB documentation as published on docs.arangodb.com.

The ArangoDB documentation is licensed under Apache-2.0. See LICENSE for details.

Contribute

To suggest a concrete change to the documentation, you may edit files directly on GitHub. It will fork the repository automatically if necessary. Commit the changes to a new branch, give it a succinct name, and open a pull request (PR).

To add more changes to a PR, click on the branch name below the headline of the PR (the latter link of "<User> wants to merge <n> commits into main from branch-name"). Then locate and edit the next file. Commit the changes directly to your PR branch.

For general suggestions, feel free to create an issue.

Automatic Previews

In every pull request, the arangodb-docs-automation bot comments with a deploy preview link. Whenever you push a change to the PR branch, a preview is built. If it succeeds, then you can view the preview hosted on Netlify by following the link.

Note that the automatic previews run plain builds, which means that generated content is not updated. The ArangoDB documentation team takes of regenerating this content if necessary.

Contributor License Agreement

To accept external contributions, we need you to fill out and sign our Contributor License Agreement (CLA). We use an Apache 2 CLA for ArangoDB, which can be found here: https://www.arangodb.com/documents/cla.pdf You can scan and email the CLA PDF file to [email protected] or send it via fax to +49-221-2722999-88.

Build the documentation

The following section describes how you can build the documentation locally. This is not strictly necessary when contributing. You can also rely on automatic previews for basic changes.

The toolchain

At the core, docs are generated with the static site generator Hugo. The hugo build command generates static HTML content from the Markdown, data, and theme files.

The documentation toolchain is containerized and uses Docker Compose to orchestrate builds. The following containers are created:

  • toolchain - contains the code for controlling the toolchain and the generated content, spawning the other containers
  • docs_arangoproxy - contains the arangoproxy web server written in Go that handles generated content using arangosh to run examples against a server and assemble the OpenAPI descriptions
  • docs_site - contains Hugo and the logic to start it
  • docs_server_<version> - an ArangoDB single server
  • docs_server_<version>_cluster - an ArangoDB cluster

Render hooks

For headlines, links, images, and fenced code blocks in Markdown, Hugo can run template files that are called Render Hooks to trigger special processing. This is used to extract and execute examples against actual ArangoDB servers and place the output into the rendered documentation, for example.

  • Heading render hook

    Defined in site/themes/arangodb-docs-theme/layouts/_default/_markup/render-heading.html

    Checks that there is no level 1 # Headline in the content, as this is reserved for the title defined in the front matter. Also injects the widget to copy anchor links to the clipboard.

  • Link render hook

    Defined in site/themes/arangodb-docs-theme/layouts/_default/_markup/render-link.html

    Scans all the hrefs in a file and tries to retrieve the page from that link. If the page is not found, the build fails because of a broken link.

  • Image render hook

    Defined in site/themes/arangodb-docs-theme/layouts/_default/_markup/render-image.html

    Transforms the style attributes defined in the image link as {path.png?{attribute1=value1&attribute2=value2&..}} in a style attribute inside the img HTML tag.

  • Codeblock render hook

    Defined in site/themes/arangodb-docs-theme/layouts/_default/_markup/render-codeblock-*.html

    Triggers a remote call to the arangoproxy web server for assembling the OpenAPI descriptions as well as to run the example generation if the code starts with a front matter block surrounded by ---, like this:

    ```aql
    ---
    name: the_answer
    description: AQL block with front matter
    ---
    RETURN 42
    ```

    The following codeblocks are supported:

    • ```js for arangosh / JavaScript API examples
    • ```aql for AQL query examples
    • ```openapi for REST HTTP API descriptions
    • ```curl for REST HTTP API examples

The hooks trigger a POST call to the dedicated arangoproxy endpoint (/js, /aql, /curl, openapi) with the entire codeblock as request body.

The arangoproxy endpoint parses the request, checks if the examples is cached, otherwise executes the code against the ArangoDB instance with the version as determined from the version folder name and saves the example output in the cache.

The input/output (as defined in the YAML render option) is returned as JSON to Hugo in the render hook, which generates HTML replacing the codeblock in the file with the input/output of the example.

Build workflows

The following build workflows exist:

  • Plain build workflow

    Build docs without re-generating examples (using a committed cache file).

    Includes the assembly of the REST HTTP API descriptions (OpenAPI) with validation at each run. ```curl examples require a different workflow.

    You may need to specify upstream branches.

  • Scheduled workflow

    The following generated content is re-generated periodically using CircleCI:

    • Metrics
    • Error codes and meanings
    • AQL optimizer rules
    • Startup options
    • oasisctl documentation
  • Example generation workflow

    Build docs including re-generating examples for AQL, arangosh (JavaScript API), and cURL (HTTP API).

    Specifying upstream branches is required for all versions in which you modified example code and thus require a re-generation. These can be a Docker Hub images or GitHub pull request links.

Plain build

Go to the toolchain/docker/<architecture> folder, with <architecture> being either amd64 for x86-64 CPUs and arm64 for 64-bit ARM CPUs (including Apple silicon like M1).

Run the docker compose services using the docker-compose.pain-build.yml file.

docs-hugo/toolchain/docker/amd64> docker compose -f docker-compose.plain-build.yml up --abort-on-container-exit

The site will be available at http://localhost:1313.

To make the documentation tooling not start a live server in watch mode but rather create a static build and exit, set the environment variable ENV to any value other than local before calling docker compose ...:

export ENV=static  # Bash
set -xg ENV static # Fish
$Env:ENV='static'  # PowerShell

The output files will be written to site/public/.

Scheduled and example generation build

Configuration

The toolchain container needs to be set up via config file in toolchain/docker/config.yaml:

generators:   # Generators to trigger - empty string defaults to all generators
servers:      # Array to define arangodb servers to be used by the toolchain
  - image:    # arangodb docker image to be used, can be arangodb/enterprise-preview:... or a branch name
    version:  # docs branch to put the generated content into
  - ...       # Additional images and versions as needed

Available generators

  • examples
  • metrics
  • error-codes
  • optimizer
  • options
  • oasisctl

The generators entry is a space-separated string.

If metrics or error-codes is in the generators string, the following environment variable has to be exported:

export ARANGODB_SRC_{VERSION}=path/to/arangodb/source

Substitute {VERSION} with a version number like 3_11.

On Windows using PowerShell, use a Unix-like path:

$Env:ARANGODB_SRC_3_11 = "/Drive/path/to/arangodb"

Configuration example

generators: examples oasisctl options optimizer
servers:
  - image: arangodb/enterprise-preview:3.11-nightly
    version: "3.11"
  - image: arangodb/enterprise-preview:devel-nightly
    version: "3.12"

Run the toolchain

Go to the toolchain/docker/<architecture> folder, with <architecture> being either amd64 for x86-64 CPUs and arm64 for 64-bit ARM CPUs (including Apple silicon like M1).

Run the docker compose services without specifying a file:

docs-hugo/toolchain/docker/arm64> docker compose up --abort-on-container-exit

The site will be available at http://localhost:1313

Work with the documentation content

Documentation structure

In the site/content directory, the directories 3.10, 3.11 etc. represent the individual ArangoDB versions and their documentation. There is only one maintained version of the documentation for every minor and major version (3.12, 4.0, etc.) but not for every patch release (e.g. 3.12.1).

Having a folder per version has the advantage that all versions can be built at once, but the drawback of Git cherry-picking not being available and therefore requiring to manually apply changes to different versions as necessary.

  • site/ - Folder with all files that Hugo needs to generate the site
    • config/ - The base Hugo configuration in _default/ as well as additional configurations for different build workflows
    • content/ - The Markdown source files in version folders as well as a shared folder for images
    • data/ - Contains JSON and YAML files for the documentation versions, the OpenAPI tag list, the example cache, etc.
    • public/ - Default output directory for the generated site (not committed)
    • resources/ - Holds the various cached resources that are generated by Hugo when using hugo serve
    • themes/ - Folder for Hugo themes, containing the customized ArangoDB docs theme
  • toolchain/ - Folder for the docs toolchain tools and scripts
    • arangoproxy/ - Source code of the arangoproxy web server
    • docker/ - The Docker container and compose files, with two sets of configurations for the amd64 and arm64 architectures that are needed for the scheduled and example generation build workflows
    • scripts/ - The toolchain scripts

Markup overview

The documentation is written in the light-weight markup language Markdown, using the GitHub flavor, and further extended by Hugo shortcodes for advanced templating needs.

For an overview over the basic markup, see the CommonMark help.

The following extensions are available:

Admonitions

You can use admonitions for hints and callouts that render in a colored box with an icon, highlighting useful or important information.

{{< danger >}}
Critical information to prevent data loss or similarly impactful events.
{{< /danger >}}
{{< warning >}}
Be careful and mind restrictions to avoid issues.
{{< /warning >}}
{{< security >}}
Mind this information to keep the system and your data safe.
{{< /security >}}
{{< info >}}
Helpful information to have.
{{< /info >}}
{{< tip >}}
Follow best practices and utilize features to set yourself up for success.
{{< /tip >}}

Admonitions can also be used in description fields inside of ```openapi code blocks but the syntax then needs to be like this:

```openapi
paths:
  /_api/endpoint:
    post:
      description: |
        {{</* warning */>}}
        Admonition inside of REST HTTP API description.
        {{</* /warning */>}}
        ...
```

For the OpenAPI JSON output, this is rendered as > **WARNING:** ... (a blockquote with the admonition type in bold and uppercase). Only a single paragraph is properly supported by the toolchain. Additional paragraphs will not be part of the blockquote.

Admonitions inside of other shortcodes need to use the special syntax, too:

{{< expand title="Outer shortcode" >}}

{{</* tip */>}}
Inner shortcode
{{</* /tip */>}}

{{< /expand >}}

Tags

Tags let you display badges, usually below a headline.

This is mainly used for pointing out if a feature is only available in the Enterprise Edition of ArangoDB, the ArangoGraph Insights Platform, or both. See Edition remarks for details.

Tabs

Display content with a tabbed interface, like for code examples using different languages or if there are multiple ways of achieving a goal like configuring ArangoDB.

{{< tabs "startup-options" >}}

{{< tab "Command-line" >}}
Start `arangod` with the startup option `--log.level startup=trace`.
{{< /tab >}}

{{< tab "Configuration file" >}}
Include the following startup option setting in your `arangod.conf`:

```conf
[log]
level = startup=trace

{{< /tab >}}

{{< /tabs >}}


The parameter for the `tabs` shortcode is a group identifier. If there are
multiple tab groups in one page, changing the active tab of one of them also
changes the active tabs of all other groups with the same identifier while
groups with different identifiers are unaffected. The browser remembers the last
active tab of each group.

The parameter for the `tab` shortcode is the label to display for the tab in the
tab panel. Tab groups using the same identifier should use the same tab labels.

#### Figures

If you want to add an image with a caption, use this shortcode instead of the
native Markdown syntax `![alt](/images/file.png)`:

```markdown
{{< image src="../images/file.png" alt="Description of image content, used as caption" >}}

Available attributes:

  • src: location of the image file
  • class: CSS class to apply
  • style: CSS inline styles to apply
  • size: image width, can be numeric or one of small, medium, large
  • alt: image description for accessibility

Icons

Display an image with special styling.

{{< icon src="../images/file.png" alt="Description of image content, used by screen readers" >}}

Available attributes:

  • src: location of the image file
  • class: CSS class to apply
  • style: CSS inline styles to apply
  • size: image width, can be numeric or one of small, medium, large
  • alt: image description for accessibility

Cards

To prominently link to other content, you may use cards:

{{< cards >}}

{{% card title="Graphs" link="graphs/" icon="/images/file.png" %}}
Learn everything about graphs.
{{% /card %}}

{{% card title="Data science" link="data-science/" icon="/images/file.png" %}}
Read about ArangoDB's features for analytics.
{{% /card %}}

{{< /cards >}}

Comments

If you want to place a remark in the source files that should not end up in the generated output, you can use a comment as follows:

{{% comment %}}
Content or reminder that should not be rendered.
{{% /comment %}}

Special shortcodes

The following shortcodes also exist but are rarely used:

  • {{< expand title="A short description" >}}
    Content that is collapsed by default but can be expanded.
    {{< /expand >}}
  • {{< youtube id="dQw4w9WgXcQ" >}} can be used to embed a single YouTube video, and {{< youtube-playlist id="PL0tn-TSss6NV45d1HnLA57VJFH6h1SeH7" >}} for a YouTube playlist.

  • {{% optimizer-rules %}} is used once to render the list of AQL optimizer rules from a JSON source file.

  • {{% program-options name="arangod" %}} renders the startup options of a component like the ArangoDB server (arangod) or shell (arangosh).

  • {{% error-codes %}} renders the ArangoDB server error codes and their meaning.

  • {{% metrics %}} renders the list of ArangoDB server metrics.

Content Guidelines

  • Use American English spelling, e.g. behavior instead of behaviour.

  • Use inclusive language, e.g. work-hours instead of man-hours.

  • Get to point quickly on every page. Add lead paragraphs that summarizes what the page is about.

  • Target end-users and focus on the outcome. It should be about solutions, not features.

  • Do not use jargon or technical abbreviations in headlines or the navigation. Define the meaning if you use it in the content.

  • Do not add too many admonitions or everything ends up being a callout.

Syntax Guidelines

  • Wrap text at 80 characters where possible. This helps tremendously in version control. Pre-wrap lines if necessary.

  • Put Markdown links on a single line [link label](target.md#hash), even if it violates the guideline of 80 characters per line.

  • Avoid breaking lines of code blocks and where Markdown does not allow line breaks, e.g. in Markdown table rows (you can use <br> if really necessary).

  • Avoid using here as link label. Describe what the link points to instead.

  • Avoid overly long link labels, such as entire sentences.

  • Use relative links for cross-references to other documentation pages, e.g. ../drivers/js/_index.md instead of /3.12/drivers/js/_index.md or https://docs.arangodb.com/3.12/drivers/js/.

  • Avoid bold and italic markup in headlines. If you have to use it, then prefer **bold** and *italic* over __bold__ and _italic_ because the underscores are preserved in anchor links but asterisks are removed!

  • Inline `code` in headlines is acceptable for code values, and required for startup options because -- otherwise gets turned into an n-dash.

  • - is preferred for bullet points in unordered lists over *

  • Don't use # or === for level 1 headlines. Every page should only have a single <h1> headline, and this is automatically generated from the title front matter parameter.

  • Use ## for level 2 headlines, not --- underlines.

  • There should be a blank line above and below fenced code blocks and headlines (except if it is at the top of the document, right after the end of the front matter ---).

  • Use all lowercase languages in fenced code blocks without whitespace before or after the language, and use the following languages in favor of the ones in parentheses for consistency:

    • ```py (instead of ```python)
    • ```yaml (instead of ```yml)
    • ```sh (instead of ```shell)
    • ```js (instead of ```javascript)
    • ``` (instead of ```plain or ```text)
  • Use the exact spelling of Enterprise Edition and its features, as well as for all other terms coined by ArangoDB:

    • EnterpriseGraphs
    • SmartGraphs, Disjoint SmartGraphs
    • SmartGraphs using SatelliteCollection, not Hybrid SmartGraphs
    • SmartJoins
    • OneShard
    • Community Edition
    • Enterprise Edition
    • DB-Server, not dbserver, db-server, DBserver (unless it is a code value)
    • Coordinator (uppercase C)
    • Agent, Agency (uppercase A)
    • ArangoGraph Insights Platform and ArangoGraph for short, but not Oasis, ArangoDB Oasis, or ArangoDB Cloud
    • Deployment mode (single server, cluster, etc.), not deployment type
  • Never capitalize the names of executables or code values, e.g. write arangosh instead of Arangosh.

  • Do not write TODOs right into the content and avoid using <!-- HTML comments -->. Use {{< comment >}}...{{< /comment >}} instead.

Add links

For external links, use standard Markdown. Clicking these links automatically opens them in a new tab:

[ArangoGraph Insights Platform](https://dashboard.arangodb.cloud)

For internal links, use relative paths to the Markdown files. Always link to files, not folders (e.g. ../graphs/_index.md instead of ../graphs/). This way, links can be followed in tools like Visual Studio Code and on GitHub.

[Graphs](../graphs/_index.md)

For anchor links, append #fragment-identifier to the path if the content is in a different file, or use the fragment ID only to link to a headline in the same file:

See [Named Graphs](#named-graphs)

Version Remarks

The main page about a new feature should indicate the version the feature was added in, as shown below:

---
title: New feature
...
---
<small>Introduced in: v3.12.0</small>

...

Similarly, the remark should be added if only a section is added to an existing page, as shown below:

## Existing feature

...

### New feature section

<small>Introduced in: v3.12.0</small>

...

The value v3.12.0 implies that all later versions also have this feature (3.12.1, 3.12.2, etc., as well as 4.0.0 and later). If this is not the case, then also mention the other relevant versions. For example, if a feature is added to 3.11.5 and 3.12.2, then write the following in the 3.12 documentation:

<small>Introduced in: v3.11.5, v3.12.2</small>

All later documentation versions should use a copy of the content, as thus the 4.0 documentation would contain the same.

In the 3.11 documentation, only mention versions up to this documentation version (excluding 3.12 and later in this example), pretending no later version exists to be consistent with the rest of the 3.11 documentation and to avoid additional maintenance burdens:

<small>Introduced in: v3.11.5</small>

New options in the JavaScript and HTTP APIs are covered by the release notes, but if new options are added mid-release (not in the x.x.0 release but a later bugfix version), then this should be pointed out as follows:

- `existingOption` (number, _optional_): ...
- `newOption` (string, _optional_): ... (introduced in v3.11.5, v3.12.2).

You may also add a remark if an existing feature or option is significantly extended by a new (sub-)option in a x.x.0 release.

While version remarks are mostly Introduced in: ..., you can also mark deprecated features in the same manner with Deprecated in: ....

Edition Remarks

Pages and sections about Enterprise Edition features should indicate that the Enterprise Edition is required using a hint box. Use the following include in the general case:

{{< tag "ArangoDB Enterprise Edition" "ArangoGraph" >}}

This shortcode should be placed immediately after a headline, before any version remarks (<small>Introduced in: ...</small>).

To tag options in lists, place the shortcode as follows:

- **optionName** (data type):

  {{< tag "ArangoDB Enterprise Edition" "ArangoGraph" >}}

  Version remarks and description of the option

Most Enterprise Edition features are also available in ArangoGraph, but some features are not or in a different form (e.g. Hot Backup). If a feature is not available in ArangoGraph, use the following include instead:

{{< tag "ArangoDB Enterprise Enterprise" >}}

In the release notes, add the following at the end of a section about a new Enterprise Edition feature:

This feature is only available in the Enterprise Edition.

HTTP API options, that is options described in an ```openapi code block, should have a remark as follows if they are only available in the Enterprise Edition:

- `enterpriseOption` (boolean, _optional_): ...
  (Enterprise Edition only).

If there are both a version remark and an Enterprise Edition remark, use:

- `enterpriseOption` (boolean, _optional_): ...
  (introduced in v3.11.5 and v3.12.2, Enterprise Edition only).

Add lead paragraphs

A lead paragraph is the opening paragraph of a written work that summarizes its main ideas. Only few pages have one so far, but new content should be written with such a brief description. It is supposed to clarify the scope of the article so that the reader can quickly assess whether the following information is of relevance, but also acts as an introduction.

You can set the lead paragraph via the description parameter in the front matter of a page:

---
title: Feature X
description: >-
  You can do this and that with X, and it is ideal to solve problem Y
---
...

The lead paragraph text should end without a period, contain no links, and usually avoid other markup as well. However, bold, italic, and inline code are acceptable.

Add a page or section

Start off by finding a file name. It should be:

  • All lower-case
  • Use hyphen-minus - instead of spaces
  • Be very short but descriptive
  • Follow the patterns of existing files

Note that the file name is independent of what will show in the navigation or what will be used as headline for that page. The file name will be used as part of the final URL, however. For example, 3.12/aql/examples.md will become http://docs.arangodb.com/3.12/aql/examples/.

Create a new file with the file name and a .md file extension. Open the file in a text editor (Visual Studio Code is recommended). Add the following front matter:

---
title: The level 1 headline
description: >-
  A meaningful description of the page
menuTitle: Short navigation menu title
weight: 42 # controls navigation position within the current section (low to high)
---

Add the actual content formatted in Markdown syntax below the front matter.

Rename a page or section

Netlify supports server-side redirects configured with a text file (documentation). This is helpful when renaming folders with many subfolders and files because there is support for splatting and placeholders (but not regular expressions). See Redirect options for details. The configuration file is site/static/_redirects.

Otherwise, the following steps are necessary for moving content:

  1. Rename file or folder
  2. Set up aliases via the front matter as needed
  3. Adjust weight of pages in the front matter if necessary
  4. Update cross-references in all of the content to use the new file path

The URL of a page is derived from the file name and the parent folders, with special handling for sections (folders with a _index.md file). For example, 3.12/aql/operators.md becomes the URL path /3.12/aql/operators/, and 3.12/aql/functions/_index.md becomes /3.12/aql/functions/.

If you rename a file, from section/old-name.md to section/new-name.md for instance, make sure to add a redirect for the old URL by adding the following to the front matter of section/new-name.md:

aliases:
  - old-name

Don't forget to update any references to the old file in the content to the new path.

If you move a file from one folder to another, from old/file.md to new/file.md for instance, use a relative path as shown below:

aliases:
  - ../old/file

If you rename a folder, from old/ to new/ for instance, add the following front matter to new/_index.md:

aliases:
  - old

For aliases in _index.md files, think of the folder they are in as a file. In the above example, the folder is new/. Treating it like the file that defines the page means that the alias old is relative to its parent folder (here: the root folder of the content, site/content/). Therefore, the alias needs to be old, not ../old.

Note that you need to set up aliases for all files in new/ so that every URL which includes the old folder name redirects to the corresponding new URL. For example, for a file new/other.md (previously old/other.md), add the following:

aliases:
  - ../old/other

Aliases create HTML files with client-side redirects before any content is rendered to HTML, which means that aliases can get overwritten by content files. If this is not a mistake, the affected aliases should be removed.

Disable or limit the table of contents

The table of contents (ToC) or "On this page" on the right-hand side at the top of a page lists the headlines if there are at least two headlines on the page (excluding the title).

The ToC can be restricted to a maximum headline level to omit the deeper nested headlines for less clutter:

---
...
pageToc:
  maxHeadlineLevel: 3
---

A setting of 3 means that <h1>, <h2>, and <h3> headlines will be listed in the ToC, whereas <h4>, <h5>, and <h6> will be ignored.

Deprecate a version

When an ArangoDB version reaches End-of-Life, its documentation needs to be marked as such. For the respective version, set the deprecated attribute to true in the site/data/versions.yaml file:

 - name: "3.10"
   version: "3.10.9"
   alias: "3.10"
-  deprecated: false
+  deprecated: true

It makes a warning show at the top of every page for that version.

Add a new version

  1. In the site/data/versions.yaml file, add a new entry for the version you want to add. Example:

    +- name: "4.0"
    +  version: "4.0.0"
    +  alias: "4.0"
    +  deprecated: false
    +
     - name: "3.12"
       version: "3.12.0"
       alias: "devel"
       deprecated: false
  2. Near the top of the .circleci/config.yml file under parameters, find the last entry with the format arangodb-X_XX where X_XX is a version number like 3_12, so arangodb-3_12 for example. Duplicate the block and adjust the version number. Example:

       arangodb-3_12:
         type: string
         default: "undefined"
    +
    +  arangodb-4_0:
    +    type: string
    +    default: "undefined"

    Near the end of the file under jobs.generate-config.steps[0].run.command, find the line of code that calls python3 generate_config.py. In one of the following lines, a parameter for every version is passed as argument. Add one for the new version. Example:

                 python3 generate_config.py \
                   --workflow << pipeline.parameters.workflow >> \
    -              --arangodb-branches << pipeline.parameters.arangodb-3_11 >> << pipeline.parameters.arangodb-3_12 >> \
    +              --arangodb-branches << pipeline.parameters.arangodb-3_11 >> << pipeline.parameters.arangodb-3_12 >> << pipeline.parameters.arangodb-4_0 >> \
  3. In the toolchain/docker/amd64/docker-compose.yml file, add an entry under services.toolchain.volumes for the new version. Simply increment the value after :-/tmp/. Example:

           - ${ARANGODB_SRC_3_12:-/tmp/2}:/tmp/3.12
    +      - ${ARANGODB_SRC_4_0:-/tmp/3}:/tmp/4.0

    Under services.toolchain.environment, you need to add two different entries for the new version. Example:

           ARANGODB_SRC_3_11: ${ARANGODB_SRC_3_11}
           ARANGODB_SRC_3_12: ${ARANGODB_SRC_3_12}
    +      ARANGODB_SRC_4_0: ${ARANGODB_SRC_4_0}
           ARANGODB_BRANCH_3_11: ${ARANGODB_BRANCH_3_11}
           ARANGODB_BRANCH_3_12: ${ARANGODB_BRANCH_3_12}
    +      ARANGODB_BRANCH_4_0: ${ARANGODB_BRANCH_4_0}

    The same changes are required in the toolchain/docker/arm64/docker-compose.yml file.

  4. In the toolchain/docker/config.yaml file, add an entry for the new version. Example:

       - image: ${ARANGODB_BRANCH_3_12_IMAGE}
         version: ${ARANGODB_BRANCH_3_12_VERSION}
    +
    +  - image: ${ARANGODB_BRANCH_4_0_IMAGE}
    +    version: ${ARANGODB_BRANCH_4_0_VERSION}
  5. In the toolchain/scripts/toolchain.sh file, find the code that accesses environment variables with the format $ARANGODB_BRANCH_X_XX where X_XX is a version number like 3_12, so $ARANGODB_BRANCH_3_12 for instance. Duplicate the block of an existing version and adjust all version numbers. Example:

     if [ "$ARANGODB_BRANCH_3_12" != "" ] ; then
           export ARANGODB_BRANCH_3_12_IMAGE="$ARANGODB_BRANCH_3_12"
           export ARANGODB_BRANCH_3_12_VERSION="3.12"
     fi
     
    +if [ "$ARANGODB_BRANCH_4_0" != "" ] ; then
    +      export ARANGODB_BRANCH_4_0_IMAGE="$ARANGODB_BRANCH_4_0"
    +      export ARANGODB_BRANCH_4_0_VERSION="4.0"
    +fi
  6. In the site/data folder, create a new folder with the short version number as the name, e.g. 4.0. In the new site/data/4.0 folder, create a cache.json file with the following content:

    {}

    Add this untracked file to Git!

  7. Duplicate the folder of the most recent version in site/content, e.g. the 3.12 folder, and rename the copy to the new version, e.g. 4.0.

    The menuTitle in the front matter of the version homepage, e.g. site/content/4.0/_index.md, needs to adjusted to the new version, like menuTitle: '4.0'.

    In the folder for release notes, e.g. site/content/4.0/release-notes/, duplicate the folder of the most recent version, e.g. version-3.12, and rename it, e.g. to version-4.0. In this folder, rename the files to replace the old with the new version number, e.g. api-changes-in-3-12.md to api-changes-in-4-0.md and so on.

    In the _index.md file in the folder, e.g. site/content/4.0/release-notes/version-4.0/_index.md, you need to replace the version numbers in the front matter and links. You also need to adjust the weight in the front matter. Decrement the value by one to make the new version appear before the existing versions, but make sure that it is greater than 0, which may require adjusting the weights in all of the section files.

    In the other files of the release notes, remove the version-specific content and only leave the front matter, introductions, and headlines that are commonly used across different versions in the release notes. Adjust the version numbers in the front matter and content.

    Search the entire version folder, e.g. site/content/4.0/, for links that are meant to point to the release notes of the own version, but which are still pointing to the version the content has been copied from. For example, if you duplicated the 3.12 folder, search the 4.0 folder for version-3.12/. You should find links to version-3.12/known-issues-in-3-12.md that need to be updated to version-4.0/known-issues-in-4-0.md.

    In the release notes root file, e.g. site/content/4.0/release-notes/_index.md, add the links for the new version following the existing pattern. Do this after updating the links to the known issues so that you don't accidentally change the 3.12 link in the release notes root file.

    In the Highlights by Version page, e.g. site/content/4.0/introduction/features/highlights-by-version.md, add a section for the new version including a link to the release notes.

    Add the new, untracked files to Git!

  8. In the PULL_REQUEST_TEMPLATE.md file, add a new line for the new version. Example:

     - 3.12: 
    +- 4.0: 

    Stage all changes and commit them. Open a pull request (PR) on GitHub. You only need to specify a Docker image or PR link for - 4.0: if you plan to use the /generate or /generate-commit command to re-generate the examples. If you follow the next step, the example generation is run manually along with some other generators, so using the commands shouldn't be necessary.

    Expect the plain build to fail for the time being because of missing data files.

  9. You can use CircleCI to initially generate the data files for the new version, like the startup option dumps. You can also populate the example cache at the same time.

    Before you continue, make sure there are no conflicts in the PR with the main branch. The CircleCI workflow will otherwise create a merge commit favoring the main branch (in order to update the cache file but also affecting other files), and this can cause e.g. the versions.yaml file to get reverted in case of a conflict. The toolchain would then be unaware of the newly added version.

    In CircleCI at https://app.circleci.com/pipelines/github/arangodb/docs-hugo, select the branch of your PR and click Trigger Pipeline. Enter the parameters similar to this example:

    Type Name Value
    string workflow generate
    string arangodb-4_0 Docker Hub image (e.g. arangodb/enterprise-preview:devel-nightly) or GitHub main repo PR link (e.g. https://github.com/arangodb/arangodb/pull/123456)
    string generators examples metrics error-codes optimizer options
    string deploy-url deploy-preview-{PR-number} with the number of the docs PR
    boolean commit-generated true

    Approve the workflow in CircleCI. After a successful run, you should see a commit in the docs PR with the updated data files, as well as an updated example cache file. The plain build should no longer fail and provide you with a Netlify preview.

Remove a version

The process is similar to adding a version but you need to undo the changes.

Change the stable version

For example, if the current stable version is 3.11 and the devel version is 3.12, the site/data/versions.yaml file may look like this:

- name: "4.0"
  version: "4.0.0"
  alias: "4.0"
  deprecated: false

- name: "3.12"
  version: "3.12.0"
  alias: "devel"
  deprecated: false

- name: "3.11"
  version: "3.11.4"
  alias: "stable"
  deprecated: false

To make 3.12 the new stable version, you need to set alias to "stable" for the 3.12 entry. As there can only be a single stable version (in the documentation sense), the alias value of the former stable version needs to be changed. In this example, it is the 3.11 entry where you need to change alias to the version name, which is "3.11" in this case. Finally, you need to re-assign the "devel" alias to the version that comes after the new stable version. In this example, you need to adjust the alias of the 4.0 entry. The final configuration would then look lik this:

- name: "4.0"
  version: "4.0.0"
  alias: "devel"   # was "4.0"
  deprecated: false

- name: "3.12"
  version: "3.12.0"
  alias: "stable"  # was "devel"
  deprecated: false

- name: "3.11"
  version: "3.11.4"
  alias: "3.11"    # was "stable"
  deprecated: false

Add a new arangosh example

A complete example:

```js
---
name: ensureUniquePersistentSingle
description: Create a unique persistent index on a single document attribute
---
~db._create("ids");
db.ids.ensureIndex({ type: "persistent", fields: [ "myId" ], unique: true });
db.ids.save({ "myId": 123 });
db.ids.save({ 
  "myId": 123
}); // xpError(ERROR_ARANGO_UNIQUE_CONSTRAINT_VIOLATED)
~db._drop("ids");
```

Groups of examples should have the same name prefix.

If an example needs to be run against an ArangoDB cluster instead of a single server (default), then add the following front matter option:

type: cluster

To not render the transcript comprised of input and output but only the input or output, set the render front matter option:

render: input # or 'output', default 'input/output'

After the front matter, you can write the JavaScript code for arangosh:

db._create("collection");
db.collection.save({ _key: "foo", value: 123 });
db._query(`FOR doc IN collection RETURN doc.value`).toArray();
db._drop("collection");

Statements can span multiple lines:

db.collection.save([
  { multiple: true },
  { lines: true }
]);

The statements as well as the results will be visible in the example transcript. To hide certain statements from the output, e.g. for setup/teardown that is not relevant for the example, you can use a leading tilde ~ to suppress individual lines:

~db._create("collection");
db.collection.save({ _key: "foo" });
~db._drop("collection");

Examples need to remove the collections and Views they create. Not dropping them will raise an error unless they are specifically exempt:

~db._create("collection");
~db._createView("view", "arangosearch", {...});
db.collection.save({...});
~addIgnoreView("view");
~addIgnoreCollection("collection");

This is helpful for creating collections and Views once, using them in multiple examples, and finally dropping them instead of having to create and drop them in each example.

The last example of the series should undo the ignore to catch unintended leftovers:

~removeIgnoreCollection("collection");
~removeIgnoreView("view");
~db._dropView("view");
~db._drop("collection");

Note that a series of examples needs to be contained within a single file.

If a statement is expected to fail (e.g. to demonstrate the error case), then this has to be indicated with a special JavaScript comment:

db._document("collection/does_not_exist"); // xpError(ERROR_ARANGO_DOCUMENT_NOT_FOUND)

This will make the example generation continue despite the error. See Error codes and meanings for a list of all error codes and their names. If a unexpected error is raised, then the example generation will abort with an error.

Every backslash in a query needs to be escaped with another backslash, i.e. JSON escape sequences require two backslashes, and literal backslashes four:

db._query(`RETURN REGEX_SPLIT("foo\\t bar\\r baz\\n foob", "\\\\s+|(?:b\\\\b)")`).toArray();

This does not apply to backslashes in bind variables:

db._query(`RETURN REGEX_SPLIT(@t, @r)`, {t: "foo\t bar\r baz\n foob", r: "\\s+|(?:b\\b)"}).toArray();

Add a new AQL example

Complete example:

```aql
---
name: joinTuples
description:
bindVars: {
  friend: "friend"
}
dataset: joinSampleDataset
explain: true
---
FOR u IN users
  FILTER u.active == true
  LIMIT 0, 4
  FOR f IN relations
    FILTER f.type == @friend && f.friendOf == u.userId
    RETURN {
      "user" : u.name,
      "friendId" : f.thisUser
    }
```

An example can optionally specify a dataset in the front matter that will be loaded before the query is run:

dataset: name_of_dataset

See datasets.json for the available datasets.

To get the query explain output including the execution plan instead of the actual query result, you can optionally specify the explain option in the front matter:

explain: true

Then the actual AQL query follows, e.g.

FOR i IN 1..3
  RETURN i

The query can optionally use bind parameters that can be set via the bindVars option in the front matter:

---
...
bindVars:
  '@coll': products
  attr:
    - title
    - de
# or using JSON notation:
#bindVars: { "@coll": "products", "attr": ["title", "de"] }
---
FOR doc IN @@coll
  RETURN doc.@attr

Add a new OpenAPI endpoint description

Used to describe an HTTP REST API endpoint using the OpenAPI Specification standard in version 3.1.0.

The content inside the codeblock is a standard OpenAPI endpoint description in YAML format for a single ArangoDB endpoint. The headline above the code block is also used as the endpoint summary automatically:

### Get the service README

```openapi
paths:
  /_api/foxx/readme:
    get:
      operationId: getFoxxReadme
      description: |
        Fetches the service's README or README.md file's contents if any.
      parameters:
        - name: mount
          in: query
          required: true
          description: |
            Mount path of the installed service.
          schema:
            type: string
      responses:
        '200':
          description: |
            Returned if the request was successful.
        '204':
          description: |
            Returned if no README file was found.
      tags:
        - Foxx
```

Only define a single tag in tags as this is used to categorize the endpoints. See the openapi_tags.yaml file for the available categories, or add new ones if necessary.

The REST HTTP API endpoint descriptions are rendered in the documentation, but arangoproxy also converts the YAML to JSON and assembles a single api-docs.json file. This file is needed by the web interface for Swagger UI.

Add a new cURL example

Complete example:

```curl
---
name: HttpAqlToBase64
description: Encode example credentials using the HTTP API
---
var url = "/_api/cursor";
var body = { query: `RETURN TO_BASE64("user:pass")` };
var response = logCurlRequest('POST', url, body);
assert(response.code === 201);
logJsonResponse(response);
```

Unlike arangosh examples (```js), requests and responses need to be output explicitly by calling one of the following functions:

  • logCurlRequest(method, url, body) → response: make and output an HTTP request
  • logCurlRequestRaw(method, url, body) → response: make and output an HTTP request without code highlighting
  • logCurlRequestPlain(method, url, body) → response: make and output an HTTP request, with the payload decoded (new lines instead of \r\n etc.). Useful for formatting complex requests.
  • logJsonResponse(response): output a JSON server reply (fails on invalid JSON)
  • logJsonLResponse(response): output a JSONL server reply (fails on invalid JSON)
  • logRawResponse(response): output plaintext response (do not use for JSON replies; can be used for endpoint calls that return an empty body to output the headers)
  • logPlainResponse(response): output decoded response (new lines instead of \r\n etc.). Useful for formatting complex responses, like from batch requests.
  • logHtmlResponse(response): output HTML
  • logErrorResponse(response): dump reply to error log for testing (makes example generation fail)

To test whether requests and replies are as expected, you can add assert(expression) calls. Expressions that evaluate to false will make the example generation fail. You can inspect the CircleCI logs for details.