Index and persist a subset of postcode data

Cargo.toml

@@ -9,3 +9,4 @@ edition = "2021"
 csv = "1.1.6"
 serde = { version = "1.0.145", features = ["derive"] }
 serde_repr = "0.1.9"
+extindex = "0.5.0"

src/main.rs

@@ -1,7 +1,8 @@
-use csv::Reader;
 use csv::ReaderBuilder;
+use extindex::{Builder, Entry, SerdeWrapper, Reader as ExtReader};
 use std::error::Error;
 use std::io;
+use std::path::Path;
 use std::process;
 use serde::{Deserialize, Serialize};
 use serde;
@@ -21,7 +22,7 @@ enum UserType {
     Large,
 }
 
-#[derive(Serialize_repr, Deserialize_repr, PartialEq, Debug)]
+#[derive(Serialize_repr, Deserialize_repr, PartialEq, Debug, Clone)]
 #[repr(u8)]
 enum PositionalQuality {
     MatchedAddressPostcodeMean = 1,
@@ -43,9 +44,9 @@ enum PositionalQuality {
 //  9 No grid reference available.
 }
 
-#[derive(Debug, Deserialize)]
+#[derive(Debug, Deserialize, Serialize)]
 struct Postcode {
-    postcode: String,
+    pub postcode: String,
     status: Status,
     usertype: UserType,
     #[serde(deserialize_with = "csv::invalid_option")]
@@ -66,6 +67,31 @@ struct Postcode {
     incode: String,
 }
 
+#[derive(Serialize, Deserialize, Debug)]
+struct SmolPostcode {
+    postcode: String,
+    status: Status,
+    usertype: UserType,
+    positional_quality_indicator: PositionalQuality,
+    country: String,
+    latitude: Option<String>,
+    longitude: Option<String>,
+}
+
+impl SmolPostcode {
+    fn from_postcode(full: Postcode) -> Self {
+        SmolPostcode {
+            postcode: full.postcode.to_owned(),
+            status: full.status,
+            usertype: full.usertype,
+            positional_quality_indicator: full.positional_quality_indicator.to_owned(),
+            country: full.country.to_owned(),
+            latitude: full.latitude.to_owned(),
+            longitude: full.longitude.to_owned(),
+        }
+    }
+}
+
 fn read() -> Result<(), Box<dyn Error>> {
     let mut rdr = ReaderBuilder::new()
         .has_headers(false)
@@ -80,9 +106,50 @@ fn read() -> Result<(), Box<dyn Error>> {
     Ok(())
 }
 
+/*
+ * Thought this would be easy..
+ *
+    let postcodes = csv_iter(io::stdin());
+    for p in postcodes.take(10) {
+        println!("Postcode: {}", p);
+    }
+fn csv_iter(file: impl io::Read + 'static) -> impl Iterator<Item=String> {
+    let mut reader = ReaderBuilder::new()
+        .has_headers(false)
+        .from_reader(file);
+    let iter = reader.deserialize();
+    iter
+        .filter_map(|r: Result<Postcode, csv::Error>| r.ok())
+        .map(|postcode| postcode.postcode.to_owned())
+        .collect()
+}
+ */
+
 fn main() {
+    build();
     if let Err(err) = read() {
         println!("error running example: {}", err);
         process::exit(1);
     }
 }
+
+
+fn build() {
+    let index_file_path = Path::new("./postcodes.db");
+
+    let builder: Builder<String, SerdeWrapper<SmolPostcode>> = Builder::new(index_file_path);
+
+    let mut csv_reader = ReaderBuilder::new()
+        .has_headers(false)
+        .from_reader(io::stdin());
+    let entries = csv_reader.deserialize()
+        .filter_map(|r: Result<Postcode, csv::Error>| r.ok())
+        .map(|full: Postcode| SmolPostcode::from_postcode(full))
+        .map(|smol| Entry::new(smol.postcode.to_owned(), SerdeWrapper(smol)));
+
+    builder.build(entries.into_iter()).unwrap();
+
+    let reader = ExtReader::<String, SerdeWrapper<SmolPostcode>>::open(index_file_path).unwrap();
+    let here = reader.find(&"LS27 8BW".to_string()).unwrap().expect("Not found");
+    println!("Here: {:?}", here.value().0);
+}