postcodes/src/main.rs

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;
use serde_repr::*;

#[derive(Serialize, Deserialize, Debug)]
#[serde(rename_all = "lowercase")]
enum Status {
    Live,
    Terminated,
}

#[derive(Serialize, Deserialize, Debug)]
#[serde(rename_all = "lowercase")]
enum UserType {
    Small,
    Large,
}

#[derive(Serialize_repr, Deserialize_repr, PartialEq, Debug, Clone)]
#[repr(u8)]
enum PositionalQuality {
    MatchedAddressPostcodeMean = 1,
    LandlineMapsInspection = 2,
    ApproximateWithin50Meters = 3,
    PostcodeUnitMean = 4,
    ImputedByONS = 5,
    PostcodeSectorMean = 6,
    TerminatedPriorToGridlink = 8,
    NoGridReferenceAvailable = 9,
//  1 Within the building of the matched address closest to the postcode mean.
//  2 As for status value 1, except by visual inspection of Landline maps (Scotland only).
//  3 Approximate to within 50 metres.
//  4 Postcode unit mean (mean of matched addresses with the same postcode, but not snapped to a building).
//  5 Imputed by ONS, by reference to surrounding postcode grid references.
//  6 Postcode sector mean, (mainly PO Boxes).
//  7 Not used.
//  8 Postcode terminated prior to Gridlink® initiative, last known ONS postcode grid reference.
//  9 No grid reference available.
}

#[derive(Debug, Deserialize, Serialize)]
struct Postcode {
    pub postcode: String,
    status: Status,
    usertype: UserType,
    #[serde(deserialize_with = "csv::invalid_option")]
    easting: Option<u32>,
    #[serde(deserialize_with = "csv::invalid_option")]
    northing: Option<u32>,
    positional_quality_indicator: PositionalQuality,
    country: String,
    latitude: Option<String>,
    longitude: Option<String>,
    postcode_no_space: String,
    postcode_fixed_width_seven: String,
    postcode_fixed_width_eight: String,
    postcode_area: String,
    postcode_district: String,
    postcode_sector: String,
    outcode: String,
    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)
        .from_reader(io::stdin());
    let mut i: u32 = 0;
    for result in rdr.deserialize() {
        let record: Postcode = result?;
        // println!("{}:: {:?}", i, record);
        i += 1;
    }
    println!("Read {:?} postcodes", i);
    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())
        .filter(|full| match full.status { Status::Live => true, _ => false })
        .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);
}