ukam-os-builder 0.1.0.dev5__tar.gz → 0.1.0.dev6__tar.gz

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ukam-os-builder
-Version: 0.1.0.dev5
+Version: 0.1.0.dev6
 Summary: Download, process and transform OS address data (NGD or ABP) for UK address matching
 Project-URL: Homepage, https://github.com/moj-analytical-services/prepare_ngd_for_address_matching
 Project-URL: Repository, https://github.com/moj-analytical-services/prepare_ngd_for_address_matching
@@ -221,13 +221,10 @@ Each file contains:
 | `filename` | VARCHAR | Source file name (for example `add_gb_builtaddress.parquet`) |
 | `classificationcode` | VARCHAR | Property classification code (for example RD06 for residential) |
 | `parentuprn` | BIGINT | Parent UPRN for hierarchical addresses |
-| `rootuprn` | BIGINT | Root UPRN at the top of the hierarchy |
-| `hierarchylevel` | INTEGER | Level in the address hierarchy (1 = root) |
+| `lowertierlocalauthoritygsscode` | VARCHAR | Lower-tier local authority GSS code |
 | `floorlevel` | VARCHAR | Floor level identifier |
-| `lowestfloorlevel` | DOUBLE | Lowest floor number |
-| `highestfloorlevel` | DOUBLE | Highest floor number |
 
-Metadata columns (`classificationcode`, `parentuprn`, `rootuprn`, `hierarchylevel`, `floorlevel`, `lowestfloorlevel`, `highestfloorlevel`) are enriched via UPRN lookup from core address files. This means Royal Mail addresses and alternate address records receive metadata from their corresponding Built, Historic, or Pre-Build records.
+Metadata used in output (`classificationcode`, `parentuprn`, `lowertierlocalauthoritygsscode`, `floorlevel`) is enriched via UPRN lookup from core address files. This means Royal Mail addresses and alternate address records receive metadata from their corresponding Built, Historic, or Pre-Build records. `lowertierlocalauthoritygsscode` is always sourced from Built Address via UPRN lookup.
 
 </details>
 

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/README.md

@@ -195,13 +195,10 @@ Each file contains:
 | `filename` | VARCHAR | Source file name (for example `add_gb_builtaddress.parquet`) |
 | `classificationcode` | VARCHAR | Property classification code (for example RD06 for residential) |
 | `parentuprn` | BIGINT | Parent UPRN for hierarchical addresses |
-| `rootuprn` | BIGINT | Root UPRN at the top of the hierarchy |
-| `hierarchylevel` | INTEGER | Level in the address hierarchy (1 = root) |
+| `lowertierlocalauthoritygsscode` | VARCHAR | Lower-tier local authority GSS code |
 | `floorlevel` | VARCHAR | Floor level identifier |
-| `lowestfloorlevel` | DOUBLE | Lowest floor number |
-| `highestfloorlevel` | DOUBLE | Highest floor number |
 
-Metadata columns (`classificationcode`, `parentuprn`, `rootuprn`, `hierarchylevel`, `floorlevel`, `lowestfloorlevel`, `highestfloorlevel`) are enriched via UPRN lookup from core address files. This means Royal Mail addresses and alternate address records receive metadata from their corresponding Built, Historic, or Pre-Build records.
+Metadata used in output (`classificationcode`, `parentuprn`, `lowertierlocalauthoritygsscode`, `floorlevel`) is enriched via UPRN lookup from core address files. This means Royal Mail addresses and alternate address records receive metadata from their corresponding Built, Historic, or Pre-Build records. `lowertierlocalauthoritygsscode` is always sourced from Built Address via UPRN lookup.
 
 </details>
 

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "ukam-os-builder"
-version = "0.1.0.dev5"
+version = "0.1.0.dev6"
 description = "Download, process and transform OS address data (NGD or ABP) for UK address matching"
 readme = "README.md"
 requires-python = ">=3.10"

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/tests/test_inspect_results.py

@@ -26,7 +26,7 @@ def test_inspect_flatfile_variants_uses_config_defaults(tmp_path: Path) -> None:
                     (1001::BIGINT, 'A'::VARCHAR),
                     (1001::BIGINT, 'B'::VARCHAR),
                     (1002::BIGINT, 'C'::VARCHAR)
-            ) AS t(uprn, address_concat)
+            ) AS t(unique_id, address_concat)
         ) TO '{parquet_path.as_posix()}' (FORMAT PARQUET)
         """
     )
@@ -61,7 +61,7 @@ def test_inspect_flatfile_variants_supports_abp_pattern(tmp_path: Path) -> None:
                     (2001::BIGINT, 'A'::VARCHAR),
                     (2002::BIGINT, 'B'::VARCHAR),
                     (2002::BIGINT, 'C'::VARCHAR)
-            ) AS t(uprn, address_concat)
+            ) AS t(unique_id, address_concat)
         ) TO '{parquet_path.as_posix()}' (FORMAT PARQUET)
         """
     )

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/tests/test_public_api_integration.py

@@ -71,7 +71,7 @@ def test_package_root_inspect_flatfile_variants(tmp_path: Path) -> None:
                     (4001::BIGINT, 'A'::VARCHAR),
                     (4001::BIGINT, 'B'::VARCHAR),
                     (4002::BIGINT, 'C'::VARCHAR)
-            ) AS t(uprn, address_concat)
+            ) AS t(unique_id, address_concat)
         ) TO '{parquet_path.as_posix()}' (FORMAT PARQUET)
         """
     )

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/tests/test_smoke.py

@@ -172,17 +172,14 @@ def test_flatfile_single_chunk(temp_settings: Settings) -> None:
     column_names = [row[0] for row in schema]
 
     expected_columns = [
-        "uprn",
+        "unique_id",
         "address_concat",
         "postcode",
         "filename",
         "classificationcode",
         "parentuprn",
-        "rootuprn",
-        "hierarchylevel",
+        "lowertierlocalauthoritygsscode",
         "floorlevel",
-        "lowestfloorlevel",
-        "highestfloorlevel",
     ]
     for col in expected_columns:
         assert col in column_names, f"Column {col} should exist in output"
@@ -231,9 +228,9 @@ def test_deduplication(temp_settings: Settings) -> None:
     # Verify no exact duplicates
     con = duckdb.connect()
     result = con.execute(f"""
-        SELECT uprn, address_concat, COUNT(*) as cnt
+        SELECT unique_id, address_concat, COUNT(*) as cnt
         FROM read_parquet('{output_files[0].as_posix()}')
-        GROUP BY uprn, address_concat
+        GROUP BY unique_id, address_concat
         HAVING COUNT(*) > 1
     """).fetchall()
 

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/ukam_os_builder/__init__.py

@@ -8,7 +8,7 @@ from ukam_os_builder.os_builder.inspect_results import (
     inspect_flatfile_variants,
 )
 
-__version__ = "0.1.0.dev5"
+__version__ = "0.1.0.dev6"
 
 __all__ = [
     "create_config_and_env",

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/ukam_os_builder/data_sources/abp/transform/runner.py

@@ -170,7 +170,7 @@ def _transform_to_flatfile_chunk(
     logger.debug("Combination and deduplication in %.2f seconds", perf_counter() - t0)
 
     # Get chunk metrics
-    chunk_metrics = con.execute("SELECT COUNT(DISTINCT uprn), COUNT(*) FROM result").fetchone()
+    chunk_metrics = con.execute("SELECT COUNT(DISTINCT unique_id), COUNT(*) FROM result").fetchone()
     chunk_uprns = chunk_metrics[0]
     chunk_rows = chunk_metrics[1]
 
@@ -244,7 +244,7 @@ def transform_to_flatfile(
         con = create_duckdb_connection(settings)
         output_path = output_paths[0]
         stats = con.execute(f"""
-            SELECT COUNT(DISTINCT uprn), COUNT(*)
+            SELECT COUNT(DISTINCT unique_id), COUNT(*)
             FROM read_parquet('{output_path.as_posix()}')
         """).fetchone()
         total_uprns = stats[0]

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/ukam_os_builder/data_sources/abp/transform/stages/combine.py

@@ -62,7 +62,7 @@ def combine_and_dedupe(con: duckdb.DuckDBPyConnection) -> duckdb.DuckDBPyRelatio
             FROM deduped_filtered
         )
         SELECT
-            sr.uprn,
+            sr.uprn AS unique_id,
             sr.postcode,
             sr.address_concat,
             cb.classification_code,

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/ukam_os_builder/data_sources/ngd/to_flatfile.py

@@ -97,7 +97,6 @@ def _create_metadata_lookup_view(
             """)
 
     if not union_parts:
-        # No core files found - create empty lookup
         logger.warning("No core feature files found. Metadata lookup will be empty.")
         con.execute("""
             CREATE OR REPLACE TEMP VIEW uprn_metadata_lookup AS
@@ -112,37 +111,48 @@ def _create_metadata_lookup_view(
                 CAST(NULL AS DOUBLE) AS highestfloorlevel
             WHERE 1=0
         """)
-        return
-
-    union_sql = "\nUNION ALL\n".join(union_parts)
+    else:
+        union_sql = "\nUNION ALL\n".join(union_parts)
 
-    sql = f"""
-        CREATE OR REPLACE TEMP VIEW uprn_metadata_lookup AS
-        WITH core_data AS (
-            {union_sql}
-        ),
-        ranked AS (
+        sql = f"""
+            CREATE OR REPLACE TEMP VIEW uprn_metadata_lookup AS
+            WITH core_data AS (
+                {union_sql}
+            ),
+            ranked AS (
+                SELECT
+                    *,
+                    ROW_NUMBER() OVER (
+                        PARTITION BY uprn
+                        ORDER BY source_priority
+                    ) AS rn
+                FROM core_data
+            )
             SELECT
-                *,
-                ROW_NUMBER() OVER (
-                    PARTITION BY uprn
-                    ORDER BY source_priority
-                ) AS rn
-            FROM core_data
-        )
+                uprn,
+                classificationcode,
+                parentuprn,
+                rootuprn,
+                hierarchylevel,
+                floorlevel,
+                lowestfloorlevel,
+                highestfloorlevel
+            FROM ranked
+            WHERE rn = 1;
+        """
+        con.execute(sql)
+
+    built_path = parquet_dir / "add_gb_builtaddress.parquet"
+    built_sql = f"""
+        CREATE OR REPLACE TEMP VIEW builtaddress_ltla_lookup AS
         SELECT
-            uprn,
-            classificationcode,
-            parentuprn,
-            rootuprn,
-            hierarchylevel,
-            floorlevel,
-            lowestfloorlevel,
-            highestfloorlevel
-        FROM ranked
-        WHERE rn = 1;
+            CAST(uprn AS BIGINT) AS uprn,
+            MAX(CAST(lowertierlocalauthoritygsscode AS VARCHAR)) AS lowertierlocalauthoritygsscode
+        FROM read_parquet('{built_path.as_posix()}')
+        {where_clause}
+        GROUP BY CAST(uprn AS BIGINT)
     """
-    con.execute(sql)
+    con.execute(built_sql)
 
 
 def _create_core_feature_view(
@@ -183,6 +193,7 @@ def _create_core_feature_view(
             CAST(floorlevel AS VARCHAR) AS floorlevel,
             CAST(lowestfloorlevel AS DOUBLE) AS lowestfloorlevel,
             CAST(highestfloorlevel AS DOUBLE) AS highestfloorlevel,
+            CAST(NULL AS VARCHAR) AS lowertierlocalauthoritygsscode,
             -- Internal columns for deduplication (not in final output)
             CAST(description AS VARCHAR) AS feature_type,
             CAST(addressstatus AS VARCHAR) AS address_status,
@@ -222,6 +233,7 @@ def _create_core_feature_view(
             CAST(floorlevel AS VARCHAR) AS floorlevel,
             CAST(lowestfloorlevel AS DOUBLE) AS lowestfloorlevel,
             CAST(highestfloorlevel AS DOUBLE) AS highestfloorlevel,
+            CAST(NULL AS VARCHAR) AS lowertierlocalauthoritygsscode,
             -- Internal columns for deduplication (not in final output)
             CAST(description AS VARCHAR) AS feature_type,
             CAST(addressstatus AS VARCHAR) AS address_status,
@@ -277,6 +289,7 @@ def _create_altadd_view(
             CAST(floorlevel AS VARCHAR) AS floorlevel,
             CAST(lowestfloorlevel AS DOUBLE) AS lowestfloorlevel,
             CAST(highestfloorlevel AS DOUBLE) AS highestfloorlevel,
+            CAST(NULL AS VARCHAR) AS lowertierlocalauthoritygsscode,
             -- Internal columns for deduplication (not in final output)
             '{feature_type}' AS feature_type,
             CAST(addressstatus AS VARCHAR) AS address_status,
@@ -333,6 +346,7 @@ def _create_royal_mail_view(
             CAST(NULL AS VARCHAR) AS floorlevel,
             CAST(NULL AS DOUBLE) AS lowestfloorlevel,
             CAST(NULL AS DOUBLE) AS highestfloorlevel,
+            CAST(NULL AS VARCHAR) AS lowertierlocalauthoritygsscode,
             -- Internal columns for deduplication (not in final output)
             'Royal Mail Address' AS feature_type,
             CAST(NULL AS VARCHAR) AS address_status,
@@ -363,6 +377,7 @@ def _create_royal_mail_view(
             CAST(NULL AS VARCHAR) AS floorlevel,
             CAST(NULL AS DOUBLE) AS lowestfloorlevel,
             CAST(NULL AS DOUBLE) AS highestfloorlevel,
+            CAST(NULL AS VARCHAR) AS lowertierlocalauthoritygsscode,
             -- Internal columns for deduplication (not in final output)
             'Royal Mail Address' AS feature_type,
             CAST(NULL AS VARCHAR) AS address_status,
@@ -398,12 +413,14 @@ def _enrich_with_metadata(con: duckdb.DuckDBPyConnection) -> None:
             COALESCE(a.floorlevel, m.floorlevel) AS floorlevel,
             COALESCE(a.lowestfloorlevel, m.lowestfloorlevel) AS lowestfloorlevel,
             COALESCE(a.highestfloorlevel, m.highestfloorlevel) AS highestfloorlevel,
+            b.lowertierlocalauthoritygsscode AS lowertierlocalauthoritygsscode,
             -- Internal columns for deduplication
             a.feature_type,
             a.address_status,
             a.build_status
         FROM all_full_addresses a
-        LEFT JOIN uprn_metadata_lookup m ON a.uprn = m.uprn;
+        LEFT JOIN uprn_metadata_lookup m ON a.uprn = m.uprn
+        LEFT JOIN builtaddress_ltla_lookup b ON a.uprn = b.uprn;
     """
     con.execute(sql)
 
@@ -419,11 +436,28 @@ def _create_custom_level_rows(con: duckdb.DuckDBPyConnection) -> None:
     dedup priority and never override official address data.
     """
     sql = """
-        INSERT INTO all_full_addresses_enriched
+        INSERT INTO all_full_addresses_enriched (
+            uprn,
+            address_concat,
+            postcode,
+            filename,
+            classificationcode,
+            parentuprn,
+            rootuprn,
+            hierarchylevel,
+            floorlevel,
+            lowestfloorlevel,
+            highestfloorlevel,
+            lowertierlocalauthoritygsscode,
+            feature_type,
+            address_status,
+            build_status
+        )
         WITH level_parsed AS (
             SELECT
                 uprn, address_concat, postcode, filename,
                 classificationcode, parentuprn, rootuprn,
+                lowertierlocalauthoritygsscode,
                 hierarchylevel, floorlevel, lowestfloorlevel, highestfloorlevel,
                 address_status, build_status,
                 CASE
@@ -464,6 +498,7 @@ def _create_custom_level_rows(con: duckdb.DuckDBPyConnection) -> None:
             floorlevel,
             lowestfloorlevel,
             highestfloorlevel,
+            lowertierlocalauthoritygsscode,
             'Custom Level' AS feature_type,
             address_status,
             build_status
@@ -523,17 +558,14 @@ def _create_dedup_view(con: duckdb.DuckDBPyConnection) -> None:
           WHERE feature_type NOT IN ('Non-Addressable Object')
         )
         SELECT
-          uprn,
+          uprn AS unique_id,
           address_concat,
           postcode,
           filename,
           classificationcode,
           parentuprn,
-          rootuprn,
-          hierarchylevel,
-          floorlevel,
-          lowestfloorlevel,
-          highestfloorlevel
+          lowertierlocalauthoritygsscode,
+                    floorlevel
         FROM ranked
         WHERE rn = 1;
     """

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/ukam_os_builder/os_builder/inspect_results.py

@@ -12,7 +12,7 @@ SourceType = Literal["ngd", "abp"]
 logger = logging.getLogger(__name__)
 
 _DEFAULT_SELECT_COLUMNS = [
-    "uprn",
+    "unique_id",
     "address_concat",
     "postcode",
     "source",
@@ -128,9 +128,9 @@ def get_variant_statistics(
 
     stats = con.sql(f"""
         WITH variant_counts AS (
-            SELECT uprn, COUNT(*) AS variant_count
+            SELECT unique_id, COUNT(*) AS variant_count
             FROM read_parquet('{files_sql}')
-            GROUP BY uprn
+            GROUP BY unique_id
         )
         SELECT
             COUNT(*) AS total_uprns,
@@ -179,7 +179,7 @@ def get_random_uprn(
 
     select_columns = _choose_select_columns(con, files_sql, columns)
     random_uprn = con.sql(f"""
-        SELECT DISTINCT uprn
+        SELECT DISTINCT unique_id
         FROM read_parquet('{files_sql}')
         ORDER BY RANDOM()
         LIMIT 1
@@ -192,7 +192,7 @@ def get_random_uprn(
         SELECT
             {select_columns}
         FROM read_parquet('{files_sql}')
-        WHERE uprn = {int(random_uprn[0])}
+        WHERE unique_id = {int(random_uprn[0])}
         ORDER BY is_primary DESC NULLS LAST, source NULLS LAST, variant_label NULLS LAST
     """)
 
@@ -220,14 +220,14 @@ def get_random_large_uprn(
 
     selected = con.sql(f"""
         WITH variant_counts AS (
-            SELECT uprn, COUNT(*) AS variant_count
+            SELECT unique_id, COUNT(*) AS variant_count
             FROM read_parquet('{files_sql}')
             {where_filter}
-            GROUP BY uprn
-            ORDER BY variant_count DESC, uprn ASC
+            GROUP BY unique_id
+            ORDER BY variant_count DESC, unique_id ASC
             LIMIT {int(top_n)}
         )
-        SELECT uprn
+        SELECT unique_id
         FROM variant_counts
         ORDER BY RANDOM()
         LIMIT 1
@@ -240,7 +240,7 @@ def get_random_large_uprn(
         SELECT
             {select_columns}
         FROM read_parquet('{files_sql}')
-        WHERE uprn = {int(selected[0])}
+        WHERE unique_id = {int(selected[0])}
         {and_filter}
         ORDER BY is_primary DESC NULLS LAST, source NULLS LAST, variant_label NULLS LAST
     """)
@@ -269,7 +269,7 @@ def get_uprn_variants(
         SELECT
             {select_columns}
         FROM read_parquet('{files_sql}')
-        WHERE uprn = {int(uprn)}
+        WHERE unique_id = {int(uprn)}
         {and_filter}
         ORDER BY is_primary DESC NULLS LAST, source NULLS LAST, variant_label NULLS LAST
     """)
@@ -317,10 +317,10 @@ def inspect_flatfile_variants(
             WITH data AS (
                 SELECT * FROM read_parquet('{files_sql}')
             )
-            SELECT uprn, COUNT(*) AS variant_count
+            SELECT unique_id, COUNT(*) AS variant_count
             FROM data
-            GROUP BY uprn
-            ORDER BY variant_count DESC, uprn ASC
+            GROUP BY unique_id
+            ORDER BY variant_count DESC, unique_id ASC
             LIMIT 1 OFFSET {top_offset}
             """
         ).fetchone()
@@ -333,7 +333,7 @@ def inspect_flatfile_variants(
             f"""
             SELECT COUNT(*)
             FROM read_parquet('{files_sql}')
-            WHERE uprn = ?
+            WHERE unique_id = ?
             """,
             [target_uprn],
         ).fetchone()
@@ -343,7 +343,7 @@ def inspect_flatfile_variants(
         f"""
         SELECT *
         FROM read_parquet('{files_sql}')
-        WHERE uprn = ?
+        WHERE unique_id = ?
         ORDER BY 1
         """,
         [target_uprn],
@@ -358,7 +358,7 @@ def inspect_flatfile_variants(
             max_width=10_000
         )
         logger.info("Selected UPRN rows:")
-        con.sql(f"SELECT * FROM read_parquet('{files_sql}') WHERE uprn = {target_uprn}").show(
+        con.sql(f"SELECT * FROM read_parquet('{files_sql}') WHERE unique_id = {target_uprn}").show(
             max_width=10_000
         )
 

{ukam_os_builder-0.1.0.dev5 → ukam_os_builder-0.1.0.dev6}/uv.lock

@@ -1421,7 +1421,7 @@ wheels = [
 
 [[package]]
 name = "ukam-os-builder"
-version = "0.1.0.dev5"
+version = "0.1.0.dev6"
 source = { editable = "." }
 dependencies = [
     { name = "duckdb" },