move claims to claims table

This commit is contained in:
Elijah McMorris 2023-12-13 23:07:42 -08:00
parent f0716081fe
commit 60cbb47a10
Signed by: NexVeridian
SSH key fingerprint: SHA256:bsA1SKZxuEcEVHAy3gY1HUeM5ykRJl0U0kQHQn0hMg8
2 changed files with 59 additions and 40 deletions

View file

@ -1,15 +1,13 @@
use anyhow::{Error, Result}; use anyhow::{Error, Ok, Result};
use dotenv_codegen::dotenv; use dotenv_codegen::dotenv;
use serde_json::{from_str, Value}; use serde_json::{from_str, Value};
use std::fs::File; use std::fs::File;
use std::io::{BufRead, BufReader}; use std::io::{BufRead, BufReader};
use surrealdb::engine::remote::ws::Ws; use surrealdb::{engine::remote::ws::Ws, opt::auth::Root, Surreal};
use surrealdb::opt::auth::Root; use wikidata::Entity;
use surrealdb::Surreal;
mod utils; mod utils;
use utils::*; use utils::*;
use wikidata::Entity;
#[tokio::main] #[tokio::main]
async fn main() -> Result<(), Error> { async fn main() -> Result<(), Error> {
@ -23,7 +21,7 @@ async fn main() -> Result<(), Error> {
db.use_ns("wikidata").use_db("wikidata").await?; db.use_ns("wikidata").use_db("wikidata").await?;
let file = File::open("data/w.json")?; let file = File::open("data/ex2.json")?;
let reader = BufReader::new(file); let reader = BufReader::new(file);
for line in reader.lines() { for line in reader.lines() {
@ -35,10 +33,13 @@ async fn main() -> Result<(), Error> {
let json: Value = from_str(&line)?; let json: Value = from_str(&line)?;
let data = Entity::from_json(json).expect("Failed to parse JSON"); let data = Entity::from_json(json).expect("Failed to parse JSON");
let (id, data) = EntityMini::from_entity(data); let (id, claims, data) = EntityMini::from_entity(data);
let _: Option<EntityMini> = db.delete(&id).await?; let _: Option<EntityMini> = db.delete(&id).await?;
let _: Option<EntityMini> = db.create(&id).content(data.clone()).await?; let _: Option<EntityMini> = db.create(&id).content(data.clone()).await?;
let _: Option<Claims> = db.delete(&claims.0).await?;
let _: Option<Claims> = db.create(&claims.0).content(claims.1).await?;
} }
Ok(()) Ok(())

View file

@ -1,54 +1,69 @@
use dotenv_codegen::dotenv;
use serde::{Deserialize, Serialize}; use serde::{Deserialize, Serialize};
use surrealdb::sql::Thing; use surrealdb::sql::Thing;
use wikidata::ClaimValueData; use wikidata::{ClaimValue, ClaimValueData, Entity, Lang, Pid, WikiId};
use wikidata::{ClaimValue, Entity, Lang, Pid, WikiId};
#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
pub struct Claims {
pub claims: Vec<(Thing, ClaimValueData)>,
}
#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)] #[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
pub struct EntityMini { pub struct EntityMini {
// In English
pub label: String, pub label: String,
pub claims: Vec<(Thing, ClaimValueData)>, pub claims: Thing,
pub description: String, pub description: String,
} }
impl EntityMini { impl EntityMini {
pub fn from_entity(entity: Entity) -> (Thing, Self) { pub fn from_entity(entity: Entity) -> (Thing, (Thing, Claims), Self) {
let thing_claim = Thing {
id: get_id(&entity).id,
tb: "Claims".to_string(),
};
( (
get_id(&entity), get_id(&entity),
(
thing_claim.clone(),
Self::flatten_claims(entity.claims.clone()),
),
Self { Self {
label: get_name(&entity), label: get_name(&entity),
claims: Self::flatten_claims(entity.claims.clone()), claims: thing_claim,
description: get_description(&entity).unwrap_or("".to_string()), description: get_description(&entity).unwrap_or("".to_string()),
}, },
) )
} }
fn flatten_claims(claims: Vec<(Pid, ClaimValue)>) -> Vec<(Thing, ClaimValueData)> { fn flatten_claims(claims: Vec<(Pid, ClaimValue)>) -> Claims {
claims Claims {
.iter() claims: claims
.flat_map(|(pid, claim_value)| { .iter()
let mut flattened = vec![( .flat_map(|(pid, claim_value)| {
Thing { let mut flattened = vec![(
id: pid.0.into(), Thing {
tb: "Property".to_string(), id: pid.0.into(),
}, tb: "Property".to_string(),
claim_value.data.clone(), },
)]; claim_value.data.clone(),
)];
flattened.extend(claim_value.qualifiers.iter().map( flattened.extend(claim_value.qualifiers.iter().map(
|(qualifier_pid, qualifier_value)| { |(qualifier_pid, qualifier_value)| {
( (
Thing { Thing {
id: qualifier_pid.0.into(), id: qualifier_pid.0.into(),
tb: "Property".to_string(), tb: "Property".to_string(),
}, },
qualifier_value.clone(), qualifier_value.clone(),
) )
}, },
)); ));
flattened flattened
}) })
.collect() .collect(),
}
} }
} }
@ -66,11 +81,14 @@ fn get_id(entity: &Entity) -> Thing {
fn get_name(entity: &Entity) -> String { fn get_name(entity: &Entity) -> String {
entity entity
.labels .labels
.get(&Lang("en".to_string())) .get(&Lang(dotenv!("WIKIDATA_LANG").to_string()))
.expect("No label found") .expect("No label found")
.to_string() .to_string()
} }
fn get_description(entity: &Entity) -> Option<String> { fn get_description(entity: &Entity) -> Option<String> {
entity.descriptions.get(&Lang("en".to_string())).cloned() entity
.descriptions
.get(&Lang(dotenv!("WIKIDATA_LANG").to_string()))
.cloned()
} }