cardosofelipe/syndarix
1
0
Fork 0
forked from cardosofelipe/pragma-stack
Code Issues 17 Pull Requests Wiki Activity

refactor(knowledge-base mcp server): adjust formatting for consistency and readability

Browse Source 
Improved code formatting, line breaks, and indentation across chunking logic and multiple test modules to enhance code clarity and maintain consistent style. No functional changes made.
This commit is contained in:
Felipe Cardoso
2026-01-06 17:20:31 +01:00
parent 3c30923e5b
commit 6df8787c6e
15 changed files with 306 additions and 155 deletions
Download Patch File Download Diff File Expand all files Collapse all files

56
mcp-servers/knowledge-base/chunking/markdown.py
View File

@@ -69,9 +69,7 @@ class MarkdownChunker(BaseChunker):
if not sections:
# No headings, chunk as plain text
return self._chunk_text_block(
content, source_path, file_type, metadata, []
)
return self._chunk_text_block(content, source_path, file_type, metadata, [])
chunks: list[Chunk] = []
heading_stack: list[tuple[int, str]] = [] # (level, text)
@@ -292,7 +290,10 @@ class MarkdownChunker(BaseChunker):
)
# Overlap: include last paragraph if it fits
if current_content and self.count_tokens(current_content[-1]) <= self.chunk_overlap:
if (
current_content
and self.count_tokens(current_content[-1]) <= self.chunk_overlap
):
current_content = [current_content[-1]]
current_tokens = self.count_tokens(current_content[-1])
else:
@@ -341,12 +342,14 @@ class MarkdownChunker(BaseChunker):
# Start of code block - save previous paragraph
if current_para and any(p.strip() for p in current_para):
para_content = "\n".join(current_para)
paragraphs.append({
"content": para_content,
"tokens": self.count_tokens(para_content),
"start_line": para_start,
"end_line": i - 1,
})
paragraphs.append(
{
"content": para_content,
"tokens": self.count_tokens(para_content),
"start_line": para_start,
"end_line": i - 1,
}
)
current_para = [line]
para_start = i
in_code_block = True
@@ -360,12 +363,14 @@ class MarkdownChunker(BaseChunker):
if not line.strip():
if current_para and any(p.strip() for p in current_para):
para_content = "\n".join(current_para)
paragraphs.append({
"content": para_content,
"tokens": self.count_tokens(para_content),
"start_line": para_start,
"end_line": i - 1,
})
paragraphs.append(
{
"content": para_content,
"tokens": self.count_tokens(para_content),
"start_line": para_start,
"end_line": i - 1,
}
)
current_para = []
para_start = i + 1
else:
@@ -376,12 +381,14 @@ class MarkdownChunker(BaseChunker):
# Final paragraph
if current_para and any(p.strip() for p in current_para):
para_content = "\n".join(current_para)
paragraphs.append({
"content": para_content,
"tokens": self.count_tokens(para_content),
"start_line": para_start,
"end_line": len(lines) - 1,
})
paragraphs.append(
{
"content": para_content,
"tokens": self.count_tokens(para_content),
"start_line": para_start,
"end_line": len(lines) - 1,
}
)
return paragraphs
@@ -448,7 +455,10 @@ class MarkdownChunker(BaseChunker):
)
# Overlap with last sentence
if current_content and self.count_tokens(current_content[-1]) <= self.chunk_overlap:
if (
current_content
and self.count_tokens(current_content[-1]) <= self.chunk_overlap
):
current_content = [current_content[-1]]
current_tokens = self.count_tokens(current_content[-1])
else: