denkan / cheerio-json-mapper Goto Github PK

Love this library, its exactly what cheerio is missing. Hopefully they'll merge it in!

I'm using the JSON mapper to pull marketing data about our competition from Clutch.

The following errors though when Im trying to get the attribute of the current element. I assume I'm doing it wrong.

Error: Empty sub-selector

{"companies": [
  {
    "$": "ul.directory-list li.provider-row",
    "clutchId": "$ | attr:data-clutch-pid",
    "title": ".company_info a",
    "tagline": ".tagline",
    "rating": ".rating",
    "reviewCount": ".reviews-link",
    "minProjectSize": "div [data-content='<i>Min. project size</i>'] span",
    "avgHourlyRate": "div [data-content='<i>Avg. hourly rate</i>'] span",
    "employees": "div [data-content='<i>Employees</i>'] span",
    "location": "div [data-content='<i>Location</i>'] span"
}
]}

The error occurs when I include the line

"clutchId": "$.attr",

This is the HTML Im parsing at this point

<ul class="directory-list"> <li data-position="1" id="provider-12345" data-type="Sponsor" class="provider provider-row sponsor" data-is-list="true" data-clutch-pid="12345" data-title="XXXX"

Is it possible to to do something like

const html= `
  <table>
    <tr>
      <th>Name</th>
      <th>Age</th>
    </tr>
    <tr>
      <td>John</td>
      <td>30</td>
    </tr>
    <tr>
      <td>Jane</td>
      <td>28</td>
    </tr>
  </table>
`;

const schema = [
  {
    "th: "td"
  }
]

which would map to:

[
  {
    "Name": "John",
    "Age": "30"
  },
  {
    "Name": "James",
    "Age": "28"
  },
]

I was expecting the pipe "parseAs:date" to return a Date object and not a string.

https://github.com/denkan/cheerio-json-mapper/blob/ec23188ce90353c3d046671e073de3537a622759/src/cheerio-json-
mapper.ts#L53

What is the idea behind that? Just to see if the string is a valid date?

Hi, I think this librarry is awsome so i create a website with next js to visually and make scraping easy, You just need the schema and lists of website you want to scrape, I hope this website help someone, Thank you .

Website Screenshot

Links

website: https://perfrico.vercel.app/
github: https://github.com/iabdrahim/perfrico

Thanks for writing this app, since cheerio can't seem to release a new version since 6 months!

I wonder why you choose to return a Promise<unknown> instead of a promise with the structure of the template (types)?
I guess its not easy to derive the type from the piped values, but how about adding a template argument to the function where we can specifiy the resulting objects shape?

Or have you looked at [zod](https://www.npmjs.com/package/zod) which is completely typesafe and also supports the concept of pipes (althought differently named).