tinysearch-action is a Github action to build a tinysearch engine from an index file. This action is meant to be combined with static site generators (Hugo, mdBook, ...).

This example takes a json index located in public/index.json and outputs all wasm and js files to the public/wasm directory.

- name: Build tinysearch
  uses: leonhfr/tinysearch-action@v1
  with:
    index: public/index.json
    output_dir: public/wasm
    output_types: |
      wasm
      js

- uses: leonhfr/tinysearch-action@v1
  with:
    # Path to the json index.
    # Optional.
    index: public/index.json # default
    # Path to the output directory.
    # If the directory does not exist, it will be created.
    # Required.
    output_dir: public/wasm
    # Extensions that will be copied from the tinysearch output to the output directory.
    # This should be a multiline input using the pipe operator |, with one extension per line.
    # Optional.
    output_types: | # default
      wasm
      js

The tinysearch command outputs those files:

$ ls wasm_output
demo.html                      # search demo
package.json                   # metadata
storage                        # no clue what this is but we don't use it
tinysearch_engine.d.ts         # library typescript types
tinysearch_engine.js           # library to initiate and call the tinysearch wasm
tinysearch_engine_bg.wasm      # tinysearch engine asm
tinysearch_engine_bg.wasm.d.ts # wasm typescript types

Most of the time, we're only interested in the actual wasm engine (tinysearch_engine_bg.wasm) and the Javascript glue library (tinysearch_engine.js). By default, only those files will be copied to the output directory.

The demo.html file is useful to have an implementation example for reference.

In addition to copying the specified file extensions to the output directory, the action outputs the list of copied files to be used in following steps. This is useful for example if some files have to be moved to other directories.

For example, the last step will echo file names:

- uses: actions/checkout@v3
- name: Build tinysearch
  id: tinysearch
  uses: leonhfr/tinysearch-action@v1
  with:
    output_dir: public/wasm
- name: List emitted files
  run: |
    echo ${{ steps.tinysearch.outputs.files }}

Only the scenario I'm using this action is listed at the moment, but any contributions with how you're using this action are welcome.

(Tutorial source in the tinysearch repository.)

With Hugo, the index can be built by adding a custom JSON layout. Create layouts/_default/list.json.json:

[
  {{ range $index , $e := .Site.RegularPages }}{{ if $index }}, {{end}}{{ dict "title" .Title "url" .Permalink "body" .Plain | jsonify }}{{end}}
]

Then, the config.toml has to be updated to also output JSON:

# ...

[outputs]
    home = ["html","rss","json"] # Add json to the list

# ...

The index file will be in public/index.json (hence the action default).

Finally, use the action after building the hugo website and before publishing:

- uses: actions/checkout@v3
- name: Setup Hugo
  uses: peaceiris/actions-hugo@v2
  with:
    hugo-version: '0.104.2'
- name: Build website
  run: hugo
  env:
    HUGO_ENV: production
- name: Build tinysearch
  uses: leonhfr/tinysearch-action@v1
  with:
    output_dir: public/wasm
    output_types: |
      wasm
- name: Deploy
  uses: peaceiris/actions-gh-pages@v3
  with:
    github_token: ${{ secrets.GITHUB_TOKEN }}
    publish_branch: gh-pages
    publish_dir: ./public

The example index in fixtures/index.json has been copied from the tinysearch repository.

To test this repository locally using act and the test.yaml workflow, run:

# removing previous container to force rebuild test any changes
$ docker rmi act-dockeraction ; act -v

The scripts and documentation in this project are released under the MIT License.