Remove autopublish for sub-documents

Nobody has claimed this yet.

• #454 by @ricaragao — closed without merging

**Is your feature request related to a problem? Please describe.**

I'm using Tabular with a collection with many fields and a lot of them in subdocuments
Example:

```js
{
_id: 'aaaaa',
subDoc1: {
field1: 'a',
field2: 'b',
},
subDoc2: {
field3: 'c',
field4: 'd',
}
}
```

So when I want to create a tabular with the field1 and field3, I just put definition of data as 'subDoc1.field1', and 'subDoc2.field3'. This works fine, no issues
The problem is that Tabular publish just use the top level field name to publish the document. Then, all fields are being published to the client.

**Describe the solution you'd like**
I took a look on the code and see in the file `tableInit.js` a line that remove the sub-fields name, returning just the main level of the fields. The line is 36 with the code:

```js
fields[cleanFieldName(data)] = 1;
```

**Describe alternatives you've considered**
None

**Additional context**

I did a test to my scenario removing the cleanFieldName call, and all the functions seems as fine.
new code:

```js
fields[data] = 1;
```

Just to understanding, with this change I could reduce data size on client in 80% in my case, and I don't expose any unnecessary data to client.

Start in tableInit.js at line 36 and inspect how cleanFieldName transforms nested field definitions during publication. Compare the current behavior with using the original field path, then verify that selecting subDoc1.field1 and subDoc2.field3 publishes only those nested fields without breaking existing field selection.

Written by the indexing model from the issue text.