Remove leon wiki graph command

This commit is contained in:
Joscha 2024-08-25 20:32:06 +02:00
parent c612bd35ad
commit 0eb745e928
5 changed files with 0 additions and 129 deletions

28
brood/Cargo.lock generated
View file

@ -56,7 +56,6 @@ name = "brood"
version = "0.0.0"
dependencies = [
"clap",
"csv",
"rustc-hash",
"serde",
"serde_json",
@ -108,27 +107,6 @@ version = "1.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0b6a852b24ab71dffc585bcb46eaf7959d175cb865a7152e35b348d1b2960422"
[[package]]
name = "csv"
version = "1.3.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ac574ff4d437a7b5ad237ef331c17ccca63c46479e5b5453eb8e10bb99a759fe"
dependencies = [
"csv-core",
"itoa",
"ryu",
"serde",
]
[[package]]
name = "csv-core"
version = "0.1.11"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5efa2b3d7902f4b634a20cae3c9c4e6209dc4779feb6863329607560143efa70"
dependencies = [
"memchr",
]
[[package]]
name = "heck"
version = "0.5.0"
@ -147,12 +125,6 @@ version = "1.0.11"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "49f1f14873335454500d59611f1cf4a4b0f786f9ac11f4312a78e4cf2566695b"
[[package]]
name = "memchr"
version = "2.7.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "78ca9ab1a0babb1e7d5695e3530886289c18cf2f87ec19a575a0abdce112e3a3"
[[package]]
name = "proc-macro2"
version = "1.0.86"

View file

@ -5,7 +5,6 @@ edition = "2021"
[dependencies]
clap = { version = "4.5.7", features = ["derive", "deprecated"] }
csv = "1.3.0"
rustc-hash = "2.0.0"
serde = { version = "1.0.203", features = ["derive"] }
serde_json = "1.0.118"

View file

@ -1,5 +1,4 @@
pub mod ingest;
pub mod leon_wiki_graph;
pub mod list_pages;
pub mod longest_shortest_path;
pub mod path;

View file

@ -1,94 +0,0 @@
use std::collections::HashSet;
use std::fs::File;
use std::io::{self, BufReader, BufWriter};
use std::path::Path;
use serde::{Deserialize, Serialize};
use crate::data::AdjacencyList;
#[derive(Deserialize)]
struct Article {
title: String,
url: String,
language: String,
}
#[derive(Serialize)]
struct NodeRow {
id: u32,
label: String,
}
#[derive(Serialize)]
struct EdgeRow {
source: u32,
target: u32,
}
pub fn run(datafile: &Path, articlesfile: &Path, language: &str) -> io::Result<()> {
let mut databuf = BufReader::new(File::open(datafile)?);
let data = AdjacencyList::read(&mut databuf)?;
let articlesbuf = BufReader::new(File::open(articlesfile)?);
let articles: Vec<Article> =
serde_json::from_reader(articlesbuf).expect("failed to parse articles file");
let titles = articles
.into_iter()
.filter(|a| a.language == language)
.map(|a| a.title)
.collect::<HashSet<_>>();
let page_ids = data
.pages
.split_last()
.unwrap()
.1
.iter()
.enumerate()
.filter(|(_, p)| titles.contains(&p.data.title))
.map(|(i, _)| i as u32)
.collect::<Vec<_>>();
let mut node_rows = vec![];
for i in &page_ids {
let page = data.page(*i);
let row = NodeRow {
id: *i,
label: page.data.title.clone(),
};
node_rows.push(row);
}
let mut edge_rows = vec![];
for i in &page_ids {
let links = data
.link_range(*i)
.map(|li| data.link(li).to)
.filter(|to| page_ids.contains(to))
.collect::<HashSet<_>>();
for to in links {
let row = EdgeRow {
source: *i,
target: to,
};
edge_rows.push(row);
}
}
let node_writer = BufWriter::new(File::create("nodes.csv")?);
let mut node_writer = csv::Writer::from_writer(node_writer);
for node in node_rows {
node_writer.serialize(node).unwrap();
}
let edge_writer = BufWriter::new(File::create("edges.csv")?);
let mut edge_writer = csv::Writer::from_writer(edge_writer);
for edge in edge_rows {
edge_writer.serialize(edge).unwrap();
}
Ok(())
}

View file

@ -25,8 +25,6 @@ enum Command {
LongestShortestPath { from: String },
/// Print all page titles.
ListPages,
/// Construct wikipedia article graph for Leon.
LeonWikiGraph { articles: PathBuf, language: String },
}
#[derive(Debug, Parser)]
@ -52,8 +50,5 @@ fn main() -> io::Result<()> {
commands::longest_shortest_path::run(&args.datafile, &from)
}
Command::ListPages => commands::list_pages::run(&args.datafile),
Command::LeonWikiGraph { articles, language } => {
commands::leon_wiki_graph::run(&args.datafile, &articles, &language)
}
}
}