Fixes incorrect ID generation for identical chunks in RecursiveDocumentSplitter (#9517)

* fix(preprocessor): ensure RecursiveDocumentSplitter generates unique chunk IDs

* fix: update meta handling in RecursiveDocumentSplitter to ensure correct overlap information

---------

Co-authored-by: Michele Pangrazzi <xmikex83@gmail.com>

haystack/components/preprocessors/recursive_splitter.py

@@ -423,10 +423,12 @@ class RecursiveDocumentSplitter:
         new_docs: List[Document] = []
 
         for split_nr, chunk in enumerate(chunks):
-            new_doc = Document(content=chunk, meta=deepcopy(doc.meta))
-            new_doc.meta["split_id"] = split_nr
-            new_doc.meta["split_idx_start"] = current_position
-            new_doc.meta["_split_overlap"] = [] if self.split_overlap > 0 else None
+            meta = deepcopy(doc.meta)
+            meta["parent_id"] = doc.id
+            meta["split_id"] = split_nr
+            meta["split_idx_start"] = current_position
+            meta["_split_overlap"] = [] if self.split_overlap > 0 else None
+            new_doc = Document(content=chunk, meta=meta)
 
             # add overlap information to the previous and current doc
             if split_nr > 0 and self.split_overlap > 0:

releasenotes/notes/fix-recursive-splitter-unique-ids-5ae9901b81131302.yaml

@@ -0,0 +1,4 @@
+---
+fixes:
+  - |
+    **RecursiveDocumentSplitter** now generates a unique `Document.id` for every chunk. The meta fields (`split_id`, `parent_id`, etc.) are populated _before_ `Document` creation, so the hash used for `id` generation is always unique.

test/components/preprocessors/test_recursive_splitter.py

@@ -990,3 +990,21 @@ def test_run_complex_text_with_multiple_separators():
     assert len(chunks[3].content) == 152
     assert chunks[3].content.startswith("C")
     assert chunks[3].content.endswith("D" * 50)
+
+
+def test_recursive_splitter_generates_unique_ids_and_correct_meta():
+    text = "Haystack is awesome. " * 5
+    source_doc = Document(content=text)
+
+    splitter = RecursiveDocumentSplitter(split_length=3)
+    splitter.warm_up()
+
+    chunks = splitter.run([source_doc])["documents"]
+
+    # IDs must be unique
+    assert len({c.id for c in chunks}) == len(chunks)
+
+    # parent_id and split_id checks
+    for idx, chunk in enumerate(chunks):
+        assert chunk.meta["parent_id"] == source_doc.id
+        assert chunk.meta["split_id"] == idx