assert(heading_level("# Title") == 1);

assert(heading_level("## Subtitle") == 2);

assert(heading_level("### Level 3") == 3);

assert(heading_level("###### Level 6") == 6);

assert(heading_level("Not a heading") == 0);

assert(heading_level("") == 0);

assert(heading_level("#NoSpace") == 0); // No space after #

assert(heading_level("##NoSpace") == 0);

assert(heading_text("# Title") == "Title");

assert(heading_text("## Sub Title") == "Sub Title");

assert(heading_text("### Multi Word Heading") == "Multi Word Heading");

std::string text = "This is a paragraph of text without any headings.\n\n"

"This is another paragraph that should be merged.";

auto chunks = chunk_markdown(text, 300);

// Both paragraphs are short, should merge into one chunk

assert(chunks.size() == 1);

assert(chunks[0].section.empty());

std::string text = "# My Document\n\n"

"This is the content under the heading. "

"It has enough text to be meaningful.";

auto chunks = chunk_markdown(text, 10);

assert(!chunks.empty());

// First chunk should include the heading

assert(chunks[0].text.find("# My Document") != std::string::npos);

assert(chunks[0].section == "My Document");

std::string text = "# Chapter One\n\n"

"Content for chapter one goes here with plenty of words to exceed the minimum chunk size easily.\n\n"

"# Chapter Two\n\n"

"Content for chapter two goes here with plenty of words to exceed the minimum chunk size easily.\n\n"

"# Chapter Three\n\n"

"Content for chapter three.";

auto chunks = chunk_markdown(text, 50);

assert(chunks.size() >= 2); // Should split on headings

std::string text = "# Title\n\n"

"## Section A\n\n"

"Content A with enough text to be a real chunk by itself when needed.\n\n"

"## Section B\n\n"

"Content B with enough text to be a real chunk by itself when needed.\n\n"

"### Subsection B1\n\n"

"Content B1 details.";

auto chunks = chunk_markdown(text, 30);

// Check section breadcrumbs

bool found_nested = false;

for (auto& c : chunks) {

if (c.section.find("\xC2\xBB") != std::string::npos) {

found_nested = true; // Found a "»" separator = nested section

}

}

assert(found_nested);

// When a same-or-higher level heading appears, the stack should pop

std::string text = "# A\n\n## A1\n\nContent A1.\n\n# B\n\nContent B.";

auto chunks = chunk_markdown(text, 10);

// The "B" chunk should NOT have "A" in its section

bool b_clean = false;

for (auto& c : chunks) {

if (c.text.find("Content B") != std::string::npos) {

assert(c.section.find("A1") == std::string::npos);

b_clean = true;

}

}

assert(b_clean);

std::string text = "# Title\n\nShort.\n\nAlso short.\n\nStill short.";

// With high min_chunk_size, everything should merge

auto chunks = chunk_markdown(text, 1000);

assert(chunks.size() == 1);

std::string text = "# Title\n\n";

// Add enough content to exceed min_chunk_size

for (int i = 0; i < 20; ++i) {

text += "Paragraph number " + std::to_string(i) + " with some filler text to bulk it up.\n\n";

}

auto chunks = chunk_markdown(text, 50);

assert(chunks.size() > 1);

std::string text = "# Doc\n\n"

"Some text before image.\n\n"

"\n\n"

"Some text after image.";

auto chunks = chunk_markdown(text, 10);

// base64 data should be stripped

for (auto& c : chunks) {

assert(c.text.find("base64") == std::string::npos);

}

std::string text = "# Title\n\n"

"Text with multiple spaces.\n\n\n\n\n"

"After many blank lines.";

auto chunks = chunk_markdown(text, 10);

// Multiple spaces collapsed, excessive newlines collapsed

for (auto& c : chunks) {

assert(c.text.find(" ") == std::string::npos ||

c.text.find(" ") == std::string::npos); // some tolerance

}

std::string text = "Hi.";

auto chunks = chunk_markdown(text, 300);

assert(chunks.size() == 1);

assert(chunks[0].text == "Hi.");

using ragger::clean_document_text;

assert(clean_document_text("## Heading") == "Heading");

assert(clean_document_text(" ### Sub ") == "Sub");

assert(clean_document_text("#") == "");

assert(clean_document_text(" # ") == "");

assert(clean_document_text("No markup here") == "No markup here");

assert(clean_document_text(" leading spaces") == "leading spaces");

assert(clean_document_text("trailing spaces ") == "trailing spaces");

assert(clean_document_text("line1\n## line2\nline3 \n") ==

"line1\nline2\nline3");

assert(clean_document_text("###No space heading") == "No space heading");

assert(clean_document_text("mid # hash stays") == "mid # hash stays");

assert(clean_document_text("line1\r\n## line2\r\n") == "line1\nline2");

test_heading_levels();

test_heading_text_extraction();

test_empty_input();

test_no_headings();

test_single_heading_with_content();

test_multiple_headings();

test_nested_headings();

test_heading_level_reset();

test_minimum_chunk_size_merging();

test_minimum_chunk_size_splitting();

test_base64_stripping();

test_whitespace_collapsing();

test_tiny_file();

test_clean_document_text();

std::println("test_import: all passed");

return 0;