server

server-rs/src/data/crime_by_year.rs

@@ -1,5 +1,5 @@
-//! Per-LSOA per-crime-type per-year crime counts, loaded from a side parquet
-//! and used by the right pane to plot crime-over-time. Filtering is not
+//! Per-postcode per-crime-type per-year crime counts, loaded from a side
+//! parquet and used by the right pane to plot crime-over-time. Filtering is not
 //! supported — this data is display-only.
 
 use std::path::Path;
@@ -23,8 +23,8 @@ pub struct YearPoint {
     pub count: f32,
 }
 
-/// One per crime type: ordered list of (year, count) for a single LSOA.
-pub struct LsoaCrimeSeries {
+/// One per crime type: ordered list of (year, count) for a single postcode.
+pub struct PostcodeCrimeSeries {
     /// Index into `crime_types`.
     pub type_idx: u16,
     pub points: Vec<YearPoint>,
@@ -35,8 +35,8 @@ pub struct CrimeByYearData {
     pub crime_types: Vec<String>,
     /// All years available for each crime type, same order as `crime_types`.
     pub years_by_type: Vec<Vec<i32>>,
-    /// LSOA code → all available per-type series for that LSOA.
-    pub series_by_lsoa: FxHashMap<String, Vec<LsoaCrimeSeries>>,
+    /// Postcode → all available per-type series for that postcode.
+    pub series_by_postcode: FxHashMap<String, Vec<PostcodeCrimeSeries>>,
 }
 
 impl CrimeByYearData {
@@ -44,7 +44,7 @@ impl CrimeByYearData {
         Self {
             crime_types: Vec::new(),
             years_by_type: Vec::new(),
-            series_by_lsoa: FxHashMap::default(),
+            series_by_postcode: FxHashMap::default(),
         }
     }
 
@@ -67,20 +67,20 @@ impl CrimeByYearData {
                 format!("Failed to read crime-by-year parquet at {}", path.display())
             })?;
 
-        let lsoa_col = df
-            .column("LSOA code")
-            .context("crime-by-year parquet missing 'LSOA code' column")?
+        let postcode_col = df
+            .column("postcode")
+            .context("crime-by-year parquet missing 'postcode' column")?
             .str()
-            .context("'LSOA code' column is not a string")?;
-        let lsoa_values: Vec<String> = lsoa_col
+            .context("'postcode' column is not a string")?;
+        let postcode_values: Vec<String> = postcode_col
             .into_iter()
             .enumerate()
             .map(|(row, value)| {
                 let value =
-                    value.with_context(|| format!("crime-by-year row {row} has null LSOA code"))?;
+                    value.with_context(|| format!("crime-by-year row {row} has null postcode"))?;
                 let trimmed = value.trim();
                 if trimmed.is_empty() {
-                    bail!("crime-by-year row {row} has blank LSOA code");
+                    bail!("crime-by-year row {row} has blank postcode");
                 }
                 Ok(trimmed.to_string())
             })
@@ -106,7 +106,8 @@ impl CrimeByYearData {
 
         let crime_types: Vec<String> = crime_type_cols.iter().map(|(t, _)| t.clone()).collect();
 
-        let mut series_by_lsoa: FxHashMap<String, Vec<LsoaCrimeSeries>> = FxHashMap::default();
+        let mut series_by_postcode: FxHashMap<String, Vec<PostcodeCrimeSeries>> =
+            FxHashMap::default();
         let mut years_by_type: Vec<Vec<i32>> = Vec::with_capacity(crime_type_cols.len());
         let row_count = df.height();
 
@@ -161,10 +162,10 @@ impl CrimeByYearData {
                 }
                 points.sort_by_key(|p| p.year);
 
-                series_by_lsoa
-                    .entry(lsoa_values[row].clone())
+                series_by_postcode
+                    .entry(postcode_values[row].clone())
                     .or_default()
-                    .push(LsoaCrimeSeries {
+                    .push(PostcodeCrimeSeries {
                         type_idx: type_idx as u16,
                         points,
                     });
@@ -173,7 +174,7 @@ impl CrimeByYearData {
         }
 
         info!(
-            lsoas = series_by_lsoa.len(),
+            postcodes = series_by_postcode.len(),
             crime_types = crime_types.len(),
             "Crime-by-year data loaded"
         );
@@ -181,7 +182,7 @@ impl CrimeByYearData {
         Ok(Self {
             crime_types,
             years_by_type,
-            series_by_lsoa,
+            series_by_postcode,
         })
     }
 }

server-rs/src/data/property.rs

@@ -831,10 +831,6 @@ pub struct PropertyData {
     /// Interned postcodes: reader is thread-safe, keys index into it.
     postcode_interner: lasso::RodeoReader,
     postcode_keys: Vec<lasso::Spur>,
-    /// Interned LSOA (2021) codes per row.
-    /// Used to look up per-LSOA side tables (e.g. crime time series).
-    lsoa_interner: lasso::RodeoReader,
-    lsoa_keys: Vec<lasso::Spur>,
     /// Rows for each postcode, keyed by the interned postcode key.
     postcode_row_index: FxHashMap<lasso::Spur, Vec<u32>>,
     /// Inverted index from address tokens to property rows.
@@ -881,11 +877,6 @@ impl PropertyData {
         self.postcode_interner.resolve(&self.postcode_keys[row])
     }
 
-    /// Get the LSOA (2021) code for a given row.
-    pub fn lsoa(&self, row: usize) -> &str {
-        self.lsoa_interner.resolve(&self.lsoa_keys[row])
-    }
-
     /// Get postcode components for field-level borrowing (avoids conflicting borrows with feature_data).
     pub fn postcode_parts(&self) -> (&lasso::RodeoReader, &[lasso::Spur]) {
         (&self.postcode_interner, &self.postcode_keys)
@@ -1541,15 +1532,6 @@ impl PropertyData {
             }
         }
 
-        // LSOA (2021) code per row, brought in via the postcode join. Used as a
-        // lookup key into per-LSOA side tables (e.g. crime time series).
-        match schema.get("lsoa21") {
-            Some(dtype) if matches!(dtype, DataType::String) || dtype.is_categorical() => {}
-            Some(dtype) => bail!("Column 'lsoa21' has unexpected type {:?}", dtype),
-            None => bail!("Required column 'lsoa21' not found in joined property data"),
-        }
-        select_exprs.push(col("lsoa21").cast(DataType::String));
-
         // Enum features as String
         for &name in &enum_names {
             select_exprs.push(col(name).cast(DataType::String));
@@ -1704,33 +1686,8 @@ impl PropertyData {
                     Ok(vec![None; row_count])
                 }
             };
-        let extract_required_trimmed_string_col =
-            |df: &DataFrame, name: &str| -> anyhow::Result<Vec<String>> {
-                let column = df
-                    .column(name)
-                    .with_context(|| format!("Required column '{name}' not found in parquet"))?;
-                let string_column = column
-                    .str()
-                    .with_context(|| format!("Column '{name}' is not a string column"))?;
-                string_column
-                    .into_iter()
-                    .enumerate()
-                    .map(|(row, value)| {
-                        let value = value.with_context(|| {
-                            format!("Required column '{name}' has null at row {row}")
-                        })?;
-                        let trimmed = value.trim();
-                        if trimmed.is_empty() {
-                            bail!("Required column '{name}' has blank value at row {row}");
-                        }
-                        Ok(trimmed.to_string())
-                    })
-                    .collect()
-            };
-
         let property_sub_type_raw = extract_optional_string_col(&df, "Property sub-type")?;
         let price_qualifier_raw = extract_optional_string_col(&df, "Price qualifier")?;
-        let lsoa_raw = extract_required_trimmed_string_col(&df, "lsoa21")?;
 
         tracing::info!("Building enum features");
         // enum_col_major: Vec<(values_list, encoded_as_f32)>
@@ -2041,14 +1998,6 @@ impl PropertyData {
         }
         let postcode_interner = postcode_rodeo.into_reader();
 
-        // Intern LSOA codes (permuted).
-        let mut lsoa_rodeo = lasso::Rodeo::default();
-        let mut lsoa_keys: Vec<lasso::Spur> = Vec::with_capacity(row_count);
-        for &perm_index in perm.iter() {
-            lsoa_keys.push(lsoa_rodeo.get_or_intern(&lsoa_raw[perm_index as usize]));
-        }
-        let lsoa_interner = lsoa_rodeo.into_reader();
-
         let row_to_poi_metric_idx: Vec<u32> = if poi_metrics.is_empty() {
             vec![NO_POI_METRIC_ROW; row_count]
         } else {
@@ -2220,8 +2169,6 @@ impl PropertyData {
             address_lengths,
             postcode_interner,
             postcode_keys,
-            lsoa_interner,
-            lsoa_keys,
             postcode_row_index,
             address_token_index,
             address_prefix_index,