bernardodiasc / filestojson Goto Github PK

include and exclude arrays are only accepting a very specific format of data, this should be more flexible.

Resources:

• https://github.com/isaacs/node-glob
• https://www.npmjs.com/package/glob-to-regexp

Actual

include: ['.md', '.png'],
exclude: ['README.md'],

Expected

Example:

include: ['*.md', '*.png', '*.jpg', '*.gif', 'scriptsamples/**/*.js'],
exclude: ['README.md', 'foldername'],

Specs

• https://facebook.github.io/react/
• Client-side rendered

Specs

• https://pugjs.org
• Server-side rendered

Actual

Currently the script will generate a single JSON file.

Expected

The way the output files will be created need to be configurable. One simple use case is to have a single JSON for each content type, but this could be even more flexible.

Resources:

• https://www.npmjs.com/package/gh-pages-deploy

Actual

Image files are stored together with content files, as they are content as well. The problem is that the content path is not public, therefore images aren't be accessible.

Expected

Images should be copied to a public place respecting original path structures from contents. This step need to be together with the data generation.

Looks like broken markdown can cause trouble with gh-pages and probably on data generation as well. Need to include a step somewhere somehow to avoid invalid stuff to be published.

Resources:

• https://github.com/markdownlint/markdownlint
• https://github.com/wooorm/remark-lint

Specs

• https://vuejs.org/
• Client-side rendered

Specs

• Vanilla JS
• Client-side rendered

This one is already implemented here https://github.com/bernardodiasc/filestojson/tree/master/examples/plainjs

Actual

Expected

• Include filtering functionality on tags
• Posters should use images from gallery (required #9)

Include unit tests for all functions.

Resources:

• https://github.com/yargs/yargs
• https://github.com/chalk/chalk
• https://github.com/sindresorhus/meow
• https://github.com/tj/commander.js
• https://www.npmjs.com/package/progress
• http://blog.liangzan.net/blog/2012/07/30/how-to-write-a-command-line-application-in-node-dot-js/

For each Content Type there is a translation file that return the custom structured data based on an index of files of that kind.

Actual

content type from examples in this repo

// posts.js
function translation (content, contentType) {
  let output = {}
  const allFiles = content.filter(each => each.dir.includes(`/${contentType}`))
  const index = allFiles.filter(each => each.base === 'index.md')[0]
  index.attr.forEach(each => {
    allFiles.forEach(file => {
      if (file.name === each) {
        output[each] = file
      }
    })
  })
  return output
}

export default translation

// gallery.js
function translation (content, contentType) {
  let output = {}
  const allFiles = content.filter(each => each.dir.includes(`/${contentType}`))
  const index = allFiles.filter(each => each.base === 'index.md')[0]
  index.attr.forEach(each => {
    Object.keys(each).forEach(key => {
      output[key] = Object.assign({ title: each[key] }, allFiles.filter(file => file.name === key)[0])
    })
  })
  return output
}

export default translation

Expected

psedo-code bellow demonstrate what is the translation for each content type, the rest should be abstracted away to keep theses translations as simple as possible.

// posts.js
  index.attr.forEach(each => {
    allFiles.forEach(file => {
      if (file.name === each) {
        output[each] = file
      }
    })
  })

// gallery.js
  index.attr.forEach(each => {
    Object.keys(each).forEach(key => {
      output[key] = Object.assign({ title: each[key] }, allFiles.filter(file => file.name === key)[0])
    })
  })

Actual

When running filestojson ./tests/fixtures/config.js, the config path is just a string and is also a relative path.

Expected

The config path argument need to return the actual config module.

Specs

• http://node-swig.github.io/swig-templates/
• Server-side rendered

User story

As a developer, while editing existing or creating a new content type, I want to quickly preview that specific content type data structure on console in order to have simpler development experience.

Expected

1. create new content type, as ./contents/posts
2. create index and a sample post, respectively ./contents/posts/index.md and ./contents/posts/post-1.md
3. include post-1 key/value in Posts Index:

---
- post-1: Post 1
---

4. include Post 1 content:

---
- title: Post 1
---

This is a sample post.

## Featuring Markdown

> actually Front-matter

- n
- i
- c
- e

5. [complicated configuration part here, depends on: #4 and #6]
6. run command filestojson --preview ./content/posts (depends on #2) and get on console:

# Posts content type

"posts": {
  "post-1": {
    "file": "/posts/post-1.md",
    "dir": "/posts/post-1",
    "name": "post-1",
    "base": "post-1.md",
    "ext": ".md",
    "attr": {
      "title":"Post 1",
    },
    "body": "This is a sample post.\n\n## Featuring Markdown\n\n> actually Front-matter\n\n- n\n- i\n- c\n- e\n"
  }
}

Use Jest.

Resources:

• https://x-team.com/blog/a-guide-to-creating-a-nodejs-command/
• https://developer.atlassian.com/blog/2015/11/scripting-with-node/

Actual

Currently only supporting .md and .png.

Expected

Support more textual types, more image types and maybe allow to include code types too (not to run the source code, but as a source of content, with proper treatment).

Currently the contentTypes validation are limited to an array of object. It should define better the object shape.

Actual

contentTypes: Joi.array().min(1).items(Joi.object()).required(),

Expected

Example:

contentTypes: Joi.array().min(1).items(
  Joi.object({
    arg: Joi.string().required(),
    value: Joi.func().required(),
  })
).required(),

The problem with the snippet above is that is not valid for the need, using it just as pseudo-code.