specfold_core/corpus/
scan.rs

1//! What specfold extracts from one Markdown file. This is the unit the cache stores.
2
3use crate::markdown::{body, frontmatter};
4use crate::model::doc::Frontmatter;
5use regex::Regex;
6use serde::{Deserialize, Serialize};
7use sha2::{Digest, Sha256};
8
9#[derive(Debug, Clone, Serialize, Deserialize)]
10pub struct FileScan {
11    pub frontmatter: Frontmatter,
12    pub body: body::Body,
13    pub body_hash: String,
14}
15
16pub fn hex(bytes: &[u8]) -> String {
17    bytes.iter().map(|b| format!("{b:02x}")).collect()
18}
19
20/// Splits frontmatter from body and scans the body for references and anchors.
21pub fn scan(text: &str, cite_re: &Regex) -> FileScan {
22    let split = frontmatter::split(text);
23    let fm = match frontmatter::parse(split.raw) {
24        frontmatter::Parsed::None => Frontmatter::default(),
25        frontmatter::Parsed::Map(m) => Frontmatter { map: Some(m), error: None },
26        frontmatter::Parsed::Invalid(e) => Frontmatter { map: None, error: Some(e) },
27    };
28    FileScan {
29        frontmatter: fm,
30        body: body::scan(split.body, split.body_line, cite_re),
31        body_hash: hex(&Sha256::digest(split.body.as_bytes())),
32    }
33}
specfold_core/corpus/scan.rs

specfold_core/corpus/
scan.rs
