TBX termEntry id handling

### Describe the problem

Some TBX files use `id` attribute of the `termEntry` tag as their internal identifier which is not suitable for context usage in Weblate. Importing such files in more languages into Weblate results in not matching source strings causing each language to have a unique set of source strings because of context.

This in turn leads to issues like #15420 and #15427.

The XML looks like:

```xml

```

An example of this can be https://learn.microsoft.com/en-us/globalization/reference/microsoft-terminology

### Solution brainstorm

For Weblate-generated glossaries, the `id` attribute is used when there are more terms with the same source (unique terms do not have the `id` attribute).

We might introduce an option to ignore the `id` attribute, but users will likely to figure out need for this *after* importing the strings.

### Describe alternatives you have considered

_No response_

### Screenshots

_No response_

### Additional context

_No response_