Document ID doesn't updated upon metadata update

[wochinge]

Describe the bug
If you assign the meta field post initialization to a Document, the id of the document doesn't get updated.
This is e.g. done in the PyPDFConverter.

Documents having the same ID although they have different metadata leads to issues with document stores and duplicate policy OVERWRITE as all documents end up as the same document then and even overwrite each other.

Error message
Error that was thrown (if available)

Expected behavior
The ID should update itself if the metadata is changed. Same applies to the other properties.

Additional context
Ideally we find a solution that the ID is automatically updated but also can be overridden manually?

To Reproduce

def test_set_meta_afterwards():
    doc = Document()
    old_id = doc.id
    doc.meta = {"test": 10}
    assert doc.meta == {"test": 10}
    assert doc.id != old_id

FAQ Check

• Have you had a look at our new FAQ page?

System:

• OS:
• GPU/CPU:
• Haystack version (commit or version number):
• DocumentStore:
• Reader:
• Retriever:

[julian-risch]

With #8698 and #8708 merged, the immediate issue was addressed. Before closing this issue, we should check whether it makes sense to emit a warning when a document's attribute is updated saying that the id is no re-created. At least for content and metadata I think it makes sense. Not sure about embedding.

[sjrl]

Closing this one to track @julian-risch request in a new issue #8856