Garmelon/wikilyze
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Make data representation more flexible

Browse source
This commit is contained in:
Joscha 2022-10-22 15:52:07 +02:00
parent 49b27715f0
commit 67f405a21e
3 changed files with 87 additions and 72 deletions
Download patch file Download diff file Expand all files Collapse all files

46
brood/src/commands/ingest.rs
View file

@ -6,7 +6,7 @@ use std::path::Path;
use rustc_hash::FxHashMap;
use serde::Deserialize;
use crate::data::{AdjacencyList, Link, Page};
use crate::data::{AdjacencyList, Link, LinkInfo, Page, PageInfo};
use crate::util;
#[derive(Deserialize)]
@ -61,7 +61,7 @@ impl Titles {
}
}
fn first_stage() -> io::Result<(AdjacencyList<(), ()>, Titles)> {
fn first_stage() -> io::Result<(AdjacencyList<PageInfo, LinkInfo>, Titles)> {
let mut titles = Titles::default();
let mut result = AdjacencyList::default();
@ -71,20 +71,19 @@ fn first_stage() -> io::Result<(AdjacencyList<(), ()>, Titles)> {
result.pages.push(Page {
link_idx: result.links.len() as u32,
id: json_page.id,
length: json_page.length,
redirect: json_page.redirect.is_some(),
title: json_page.title,
data: (),
data: PageInfo {
id: json_page.id,
length: json_page.length,
redirect: json_page.redirect.is_some(),
title: json_page.title,
},
});
for (to, start, end) in json_page.links {
let to = titles.insert(util::normalize_link(&to));
result.links.push(Link {
to,
start,
end,
data: (),
data: LinkInfo { start, end },
});
}
@ -100,23 +99,29 @@ fn first_stage() -> io::Result<(AdjacencyList<(), ()>, Titles)> {
result.pages.push(Page {
link_idx: result.links.len() as u32,
id: 0,
length: 0,
redirect: false,
title: "Sentinel page at the end of all pages, Q2AKO3OYzyitmCJURghJ".to_string(),
data: (),
data: PageInfo {
id: 0,
length: 0,
redirect: false,
title: "Sentinel page at the end of all pages, Q2AKO3OYzyitmCJURghJ".to_string(),
},
});
Ok((result, titles))
}
/// Create map from normalized title to index in pages.
fn initialize_pages_map(pages: &[Page<()>]) -> FxHashMap<String, u32> {
fn initialize_pages_map(pages: &[Page<PageInfo>]) -> FxHashMap<String, u32> {
let mut result = FxHashMap::default();
for (i, p) in pages.iter().enumerate() {
match result.entry(util::normalize_link(&p.title)) {
match result.entry(util::normalize_link(&p.data.title)) {
Entry::Occupied(entry) => {
eprintln!("{:?} already exists at index {}", p.title, entry.get());
eprintln!(
"{:?} already exists at index {} as {:?}",
p.data.title,
entry.get(),
util::normalize_link(&p.data.title)
);
}
Entry::Vacant(entry) => {
entry.insert(i as u32);
@ -126,7 +131,10 @@ fn initialize_pages_map(pages: &[Page<()>]) -> FxHashMap<String, u32> {
result
}
fn second_stage(first_stage: &AdjacencyList<(), ()>, titles: &Titles) -> AdjacencyList<(), ()> {
fn second_stage(
first_stage: &AdjacencyList<PageInfo, LinkInfo>,
titles: &Titles,
) -> AdjacencyList<PageInfo, LinkInfo> {
let pages_map = initialize_pages_map(&first_stage.pages);
let mut result = AdjacencyList::default();

10
brood/src/commands/path.rs
View file

@ -8,7 +8,7 @@ use crate::util;
pub fn path(datafile: &Path, from: &str, to: &str) -> io::Result<()> {
eprintln!(">> Import");
let mut databuf = BufReader::new(File::open(datafile)?);
let mut data = AdjacencyList::read(&mut databuf)?.change_page_data(f32::INFINITY);
let data = AdjacencyList::read(&mut databuf)?;
eprintln!(">> Locate from and to");
let from = util::normalize_link(from);
@ -17,15 +17,15 @@ pub fn path(datafile: &Path, from: &str, to: &str) -> io::Result<()> {
.pages
.iter()
.enumerate()
.filter(|(_, p)| !p.redirect)
.find(|(_, p)| util::normalize_link(&p.title) == from)
.filter(|(_, p)| !p.data.redirect)
.find(|(_, p)| util::normalize_link(&p.data.title) == from)
.unwrap_or_else(|| panic!("no article called {from}"));
let (to_i, to_p) = data
.pages
.iter()
.enumerate()
.filter(|(_, p)| !p.redirect)
.find(|(_, p)| util::normalize_link(&p.title) == to)
.filter(|(_, p)| !p.data.redirect)
.find(|(_, p)| util::normalize_link(&p.data.title) == to)
.unwrap_or_else(|| panic!("no article called {to}"));
dbg!(from_i, from_p, to_i, to_p);