raymon
/
wikipedia-infobox-analyzer
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Move table formatting function to main

Browse Source
master
Raymon Zutekouw 3 weeks ago
parent bd14c26285
commit 4b8562089e
Signed by: raymon
GPG Key ID: 0E62222846283925
2 changed files with 57 additions and 55 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 59
      src/lib.rs
  2. 53
      src/main.rs

59
src/lib.rs
Unescape View File

@ -1,6 +1,5 @@
#[doc = include_str!("../README.md")]
pub mod wikipedia_infobox_analyzer {
use ascii_table::{Align, AsciiTable};
use regex::Regex;
use itertools::Itertools;
@ -20,43 +19,20 @@ pub mod wikipedia_infobox_analyzer {
.collect()
}
/// Creates an ascii infobox out of the properties
pub fn format_infobox_from_used_properties(name: String, qid: String, properties_template: Vec<u64>, properties_item: Vec<u64>, ) -> String {
let mut ascii_table = AsciiTable::default();
ascii_table.set_max_width(80);
ascii_table
.column(0)
.set_header("Infobox requires")
.set_align(Align::Left);
ascii_table
.column(1)
.set_header(format!("Wikidata {name} ({qid})"))
.set_align(Align::Left);
// Display the property columns as a table
let data: Vec<Vec<String>> = join_columns(&properties_template, &properties_item)
.iter()
.map(|&m| {
let (v1, v2) = m;
vec![ v1.map_or("".to_owned(), |v| format!("P{v}").to_owned())
, v2.map_or("".to_owned(), |v| format!("P{v}").to_owned())]
}
).collect();
ascii_table.format(data)
}
/// Properties will be displayed in the right column if they are in the list for that column
///
/// All items in the left columns will be displayed
/// The remaining items in the right columns not in the left columns will trail in the right column
/// ```md
/// ┌───────┬───────┐
/// │ 1 │ 1 │
/// │ 3 │ │
/// │ │ 2 │
/// │ │ 4 │
/// └───────┴───────┘
/// ```
/// So for example, [1, 3] and [1, 2, 4] become [1, 2, 3, 4], as shown above
fn join_columns<T: Copy + Eq>(vec1: &[T], vec2: &[T]) -> Vec<(Option<T>, Option<T>)> {
pub fn join_columns<T: Copy + Eq>(vec1: &[T], vec2: &[T]) -> Vec<(Option<T>, Option<T>)> {
vec1.iter()
.flat_map(|n1| {
if let Some(&n2) = vec2.iter().find(|&x| x == n1) {
@ -124,7 +100,7 @@ pub mod wikipedia_infobox_analyzer {
.collect::<Vec<u64>>()
}
/// Fetch the wikidata identifier associated with the wikipedia article
/// Fetches the wikidata identifier associated with the wikipedia article
pub fn fetch_wiki_item_by_article_title(title: String, language_code: String) -> u64 {
let uri = format!("https://{}.wikipedia.org/w/api.php?action=query&prop=pageprops&format=json&titles={}", &language_code, &title);
@ -172,31 +148,6 @@ mod tests_lib {
);
}
#[test]
fn test_format_infobox_from_used_properties() {
assert_eq!(
format_infobox_from_used_properties(
"Earth".to_string(),
"Q2".to_string(),
vec![18, 170, 571],
vec![18, 31, 138, 170, 361, 571]
),
concat![
"┌──────────────────┬─────────────────────┐\n",
"│ Infobox requires │ Wikidata Earth (Q2) │\n",
"├──────────────────┼─────────────────────┤\n",
"│ P18 │ P18 │\n", // property: image
"│ P170 │ P170 │\n", // property: creator
"│ P571 │ P571 │\n", // property: inception
"│ │ P31 │\n", // property: instance of
"│ │ P138 │\n", // property: inception
"│ │ P361 │\n", // property: part of
"└──────────────────┴─────────────────────┘\n",
]
);
}
#[tokio::test]
async fn test_fetch_name_for_wiki_property() {
let name = fetch_name_for_wiki_property(31).await;

53
src/main.rs
Unescape View File

@ -1,5 +1,6 @@
use wikipedia_infobox_analyzer::wikipedia_infobox_analyzer::*;
use ascii_table::{Align, AsciiTable};
use dirs_next::cache_dir;
use std::collections::HashMap;
use std::fs;
@ -112,6 +113,31 @@ impl Cache {
}
}
/// Creates an ascii table out of the properties
fn format_property_usage(name: String, qid: String, properties_template: Vec<u64>, properties_item: Vec<u64>, ) -> String {
let mut ascii_table = AsciiTable::default();
ascii_table.set_max_width(80);
ascii_table
.column(0)
.set_header("Infobox requires")
.set_align(Align::Left);
ascii_table
.column(1)
.set_header(format!("Wikidata {name} ({qid})"))
.set_align(Align::Left);
// Display the property columns as a table
let data: Vec<Vec<String>> = join_columns(&properties_template, &properties_item)
.iter()
.map(|&m| {
let (v1, v2) = m;
vec![ v1.map_or("".to_owned(), |v| format!("P{v}").to_owned())
, v2.map_or("".to_owned(), |v| format!("P{v}").to_owned())]
}
).collect();
ascii_table.format(data)
}
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
@ -129,7 +155,7 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
let properties_template = extract_used_properties_from_template(template);
let properties_item = fetch_properties_for_wiki_item(qid);
let table = format_infobox_from_used_properties(
let table = format_property_usage(
title.clone(),
format!("Q{qid}"),
properties_template,
@ -168,4 +194,29 @@ mod tests_main {
// Cleanup
let _ = cache_new.delete();
}
#[test]
fn test_format_property_usage() {
assert_eq!(
format_property_usage(
"Earth".to_string(),
"Q2".to_string(),
vec![18, 170, 571],
vec![18, 31, 138, 170, 361, 571]
),
concat![
"┌──────────────────┬─────────────────────┐\n",
"│ Infobox requires │ Wikidata Earth (Q2) │\n",
"├──────────────────┼─────────────────────┤\n",
"│ P18 │ P18 │\n", // property: image
"│ P170 │ P170 │\n", // property: creator
"│ P571 │ P571 │\n", // property: inception
"│ │ P31 │\n", // property: instance of
"│ │ P138 │\n", // property: inception
"│ │ P361 │\n", // property: part of
"└──────────────────┴─────────────────────┘\n",
]
);
}
}

Write Preview
Loading…
Cancel
Save