diff --git a/llama-index-core/llama_index/core/node_parser/file/markdown.py b/llama-index-core/llama_index/core/node_parser/file/markdown.py index de4450d60605e819c11623a658f9905bce4b9338..6f8ecb984051f9538daea5f03f7a881e49d78853 100644 --- a/llama-index-core/llama_index/core/node_parser/file/markdown.py +++ b/llama-index-core/llama_index/core/node_parser/file/markdown.py @@ -99,11 +99,11 @@ class MarkdownNodeParser(NodeParser): updated_headers = {} for i in range(1, new_header_level): - key = f"Header {i}" + key = f"Header_{i}" if key in headers_metadata: updated_headers[key] = headers_metadata[key] - updated_headers[f"Header {new_header_level}"] = new_header + updated_headers[f"Header_{new_header_level}"] = new_header return updated_headers def _build_node_from_split( diff --git a/llama-index-core/tests/node_parser/test_markdown.py b/llama-index-core/tests/node_parser/test_markdown.py index 46577747e2b229a86557fa2dae31dd8d5144779a..876aae23e887423203bcbe9102bcd7ad93e2c441 100644 --- a/llama-index-core/tests/node_parser/test_markdown.py +++ b/llama-index-core/tests/node_parser/test_markdown.py @@ -19,8 +19,8 @@ Header 2 content ] ) assert len(splits) == 2 - assert splits[0].metadata == {"Header 1": "Main Header"} - assert splits[1].metadata == {"Header 1": "Header 2"} + assert splits[0].metadata == {"Header_1": "Main Header"} + assert splits[1].metadata == {"Header_1": "Header 2"} assert splits[0].text == "Main Header\n\nHeader 1 content" assert splits[1].text == "Header 2\nHeader 2 content" @@ -80,11 +80,11 @@ Content ] ) assert len(splits) == 4 - assert splits[0].metadata == {"Header 1": "Main Header"} - assert splits[1].metadata == {"Header 1": "Main Header", "Header 2": "Sub-header"} + assert splits[0].metadata == {"Header_1": "Main Header"} + assert splits[1].metadata == {"Header_1": "Main Header", "Header_2": "Sub-header"} assert splits[2].metadata == { - "Header 1": "Main Header", - "Header 2": "Sub-header", - "Header 3": "Sub-sub header", + "Header_1": "Main Header", + "Header_2": "Sub-header", + "Header_3": "Sub-sub header", } - assert splits[3].metadata == {"Header 1": "New title"} + assert splits[3].metadata == {"Header_1": "New title"}