MarkdownExtractionService.cs

Service that extracts structured sections from Markdown files using strategy-based extraction.

What it does

Reads a Markdown file from disk, splits it into lines, and applies extraction strategies defined in a SourceConfig to produce named sections. Returns the same ExtractionResult as the PDF extraction service, making source types interchangeable.

Interface

public interface IMarkdownExtractionService
{
    ExtractionResult ExtractSectionsFromConfig(string filePath, SourceConfig sourceConfig);
}

Extraction Strategies

Strategy	Implementation
firstHeading	Finds the first line matching #{level} prefix. Returns heading text without the # prefix.
firstParagraph	Finds the first non-empty, non-heading line after a heading. Returns the text.
betweenPatterns	Scans lines, starts capturing at startPattern regex, stops at any stopPatterns regex. Respects includeStartLine.
regex	Matches a regex pattern against each line. Returns the first match (supports capture groups).

Key difference from PDF extraction

Markdown is already structured text — no PdfPig, no column detection, no font analysis needed. The service is simple line-based text parsing (~170 lines).

Strategy Parameters

firstHeading

• level (int) — Heading level to match (1 = #, 2 = ##, etc.). Default: 1.

betweenPatterns

• startPattern (string) — Regex to match the start line
• stopPatterns (string[]) — Regexes that stop capturing when matched
• includeStartLine (bool) — Whether to include the matched start line. Default: true.

regex

• pattern (string) — Regex pattern to match
• flags (string) — Regex flags (e.g. "i" for case-insensitive)
• captureGroup (int) — Which capture group to return. Default: 0 (full match).

Registration

Registered as scoped via UpDocComposer:

builder.Services.AddScoped<IMarkdownExtractionService, MarkdownExtractionService>();

Dependencies

• ILogger<MarkdownExtractionService> — for logging extraction progress and errors

Namespace

namespace UpDoc.Services;