andras/perfect-postcode
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages 2 Wiki Activity Actions

Deploy again

Browse source
This commit is contained in:
Andras Schmelczer 2026-02-19 22:24:06 +00:00
parent ffe080adef
commit 787428f1a5
18 changed files with 717 additions and 223 deletions
Download patch file Download diff file Expand all files Collapse all files

75
server-rs/src/data/property.rs
View file

@ -437,82 +437,11 @@ impl PropertyData {
tracing::info!(rows = prop_count, "Properties joined with postcodes");
// Load online listings (buy + rent) — these have their own lat/lon.
// Normalize column names from finder output to server-expected names.
// strict=false: columns already using the new name are silently skipped.
// Expects the new finder parquet format with human-readable column names.
let load_listings = |path: &Path, label: &str| -> anyhow::Result<DataFrame> {
tracing::info!("Loading {} listings from {:?}", label, path);
let mut lf = LazyFrame::scan_parquet(path, Default::default())
let lf = LazyFrame::scan_parquet(path, Default::default())
.with_context(|| format!("Failed to scan {label} listings parquet"))?;
let schema = lf
.collect_schema()
.with_context(|| format!("Failed to read {label} listings schema"))?;
// Rename raw finder columns → server-expected names (no-op if already renamed)
let lf = lf.rename(
[
"postcode",
"address",
"latitude",
"longitude",
"bedrooms",
"bathrooms",
"total_rooms",
"tenure",
"property_type",
"property_sub_type",
"price_qualifier",
"floorspace_sqm",
"url",
"features",
],
[
"Postcode",
"Address per Property Register",
"lat",
"lon",
"Bedrooms",
"Bathrooms",
"Number of bedrooms & living rooms",
"Leashold/Freehold",
"Property type",
"Property sub-type",
"Price qualifier",
"Total floor area (sqm)",
"Listing URL",
"Listing features",
],
false,
);
// Derive missing columns for raw finder output that doesn't have them
let listing_status = if label == "buy" {
"For sale"
} else {
"For rent"
};
let lf = if schema.get("Listing status").is_none() {
lf.with_column(lit(listing_status).alias("Listing status"))
} else {
lf
};
let lf = if schema.get("Asking price").is_none() && schema.get("price").is_some() {
if label == "buy" {
lf.with_column(col("price").alias("Asking price"))
} else {
// Normalize rent to monthly: weekly×52/12, yearly÷12
lf.with_column(
when(col("price_frequency").eq(lit("weekly")))
.then(col("price").cast(DataType::Float64) * lit(52.0 / 12.0))
.when(col("price_frequency").eq(lit("yearly")))
.then(col("price").cast(DataType::Float64) / lit(12.0))
.otherwise(col("price").cast(DataType::Float64))
.cast(DataType::Int64)
.alias("Asking rent (monthly)"),
)
}
} else {
lf
};
// Join with postcodes for area features (listings have their own lat/lon)
let pc_no_coords = postcode_df.clone().lazy().drop(["lat", "lon"]);