Multiscrape and multi entity help

Configuration
1

Hi, I’m still very new to HA. I’m trying to put something on my wife’s dashboard, if I keep her happy I get more HA time.

I am using multiscrape to get info from a webpage plain html. and then display it on her dashboard. I feel there must be a more efficient way. The webpage is plain html with a table 11 rows and 5 columns. (I control the webpage so of there is a better way to put that up let me know that also.) I feel my “scrape” is too basic and therefore crazy busy.
right now (testing) I’m only pulling a little data basically a uniquename, name, and select statement for each cell in the table. is there a more efficient route?


sensor:
    - unique_id: r1c1
      name: r1c1
      select: "table > tr > td"
    - unique_id: r1c2
      name: r1c2
      select: "table > tr > td:nth-child(2)"
    - unique_id: r1c3
      name: r1c3
      select: "table > tr > td:nth-child(3)"
    - unique_id: r1c4
      name: r1c4
      select: "table > tr > td:nth-child(4)"
    - unique_id: r2c1
      name: r2c1
      select: "table > tr:nth-child(2) >td"
    - unique_id: r2c2
      name: r2c2
      select: "table > tr:nth-child(2) > td:nth-child(2)"
    - unique_id: r2c3
      name: r2c3
      select: "table > tr:nth-child(2) > td:nth-child(3)"
    - unique_id: r2c4
      name: r2c4
      select: "table > tr:nth-child(2) > td:nth-child(4)"

etc…
Thanks

2

Yes. Output something machine-readable like JSON instead of / as well as the HTML table.

Alternatively, give each cell an HTML id like this:

<table>
<tbody>
<tr>
<td id="r1c1">First</td>
<td id="r1c2">Second</td>
</tr>
</tbody>
</table>

and then your select can be simply "td#r1c1", for example.

1 Like
3

can you point me to some instructions on using the json method?

and calling the cell by name is way simpler also.
Thanks

4

Depends what the data is and how you are generating the web page.

JSON spec is here but that’ll be baffling to a beginner.

It might be something like this, for a bin collection calendar:

{'black_bin': '2024-01-24', 'blue_bin': '2024-02-03'}

You can then pull that in with a rest sensor and do something like:

rest:
  resource: URL
  sensor:
    - name: "Black bin date"
      value_template: "{{ value_json.get('black_bin', 'unavailable') }}"
    - name: "Blue bin date"
      value_template: "{{ value_json.get('blue_bin', 'unavailable') }}"
    - name: "Brown bin date"
      value_template: "{{ value_json.get('brown_bin', 'unavailable') }}"

Doesn’t matter which way around the bin dates are, or how many of them there are: it’ll pick up the appropriate one provided it’s in the data, and this example will return unavailable if not. That’s a trivial example: you can get much more complex data structures yet they’re easy to work with.

In comparison, the scrape sensor is like trying to use character recognition to read off a screenshot — particularly on public sites, where any change of styling and layout can screw up your sensors. It should be a last resort where there is no other method available.

I’m going to be a super-pedant and note that HTML name attribute is not the same as the id attribute I suggested, nor is it valid on a <td> element: