use polars and duckdb for lazy processing

src/app/vinyl/assets.py

@@ -2,7 +2,6 @@ from datetime import datetime
 from glob import glob
 
 import duckdb
-import pandas as pd
 import polars as pl
 import structlog
 from duckdb.typing import DATE, VARCHAR
@@ -110,73 +109,73 @@ def deals(context):
     )
 
 
-@asset(deps=[deals], io_manager_key="duckdb_io_manager")
-def new_deals(
-    context: OpExecutionContext,
-    # duckdb: DuckDBResource
-) -> pd.DataFrame:
+@asset(deps=[deals], io_manager_key="polars_parquet_io_manager")
+def new_deals(context: OpExecutionContext) -> pl.DataFrame:
     ic()
     storage_dir = context.instance.storage_directory()
     asset_key = "deals"
 
     with duckdb.connect() as con:
-        # with duckdb.get_connection() as con:
         con.create_function("PARSE_DATE", parse_date, [VARCHAR], DATE)
         return con.execute(
             f"""
-        WITH tmp_plato AS (
-            SELECT
-                source,
-                CAST(date AS DATE) AS date,
-                ean AS id,
-                _artist AS artist,
-                LOWER(title) AS title,
-                CAST(_date AS DATE) AS release,
-                CAST(_price AS FLOAT) AS price,
-                CONCAT('https://www.platomania.nl', url) AS url,
-            FROM read_parquet('{storage_dir}/{asset_key}/*/plato.parquet', union_by_name = true)
-        ), tmp_sounds AS (
+            WITH tmp_plato AS (
+                SELECT
+                    source,
+                    CAST(date AS DATE) AS date,
+                    ean AS id,
+                    _artist AS artist,
+                    LOWER(title) AS title,
+                    CAST(_date AS DATE) AS release,
+                    CAST(_price AS FLOAT) AS price,
+                    CONCAT('https://www.platomania.nl', url) AS url,
+                FROM read_parquet('{storage_dir}/{asset_key}/*/plato.parquet', union_by_name = true)
+            ), tmp_sounds AS (
+                SELECT
+                    source,
+                    date,
+                    id,
+                    LOWER(TRIM(COALESCE(artist, SPLIT(name, '-')[1]))) AS artist,
+                    LOWER(TRIM(COALESCE(title, ARRAY_TO_STRING(split(name, '-')[2:], '-')))) AS title,
+                    PARSE_DATE(release) AS release,
+                    CAST(price AS FLOAT) AS price,
+                    CONCAT('https://www.sounds.nl/detail/', id) AS url
+                FROM read_parquet('{storage_dir}/{asset_key}/*/sounds.parquet', union_by_name = true)
+            ), tmp_both AS (
+                SELECT * FROM tmp_plato UNION ALL SELECT * FROM tmp_sounds
+            ), tmp_rn AS (
+                SELECT
+                    *,
+                    ROW_NUMBER() OVER(PARTITION BY source, id, artist, title, price ORDER BY date DESC) as rn
+                FROM tmp_both
+            )
             SELECT
                 source,
                 date,
                 id,
-                LOWER(TRIM(COALESCE(artist, SPLIT(name, '-')[1]))) AS artist,
-                LOWER(TRIM(COALESCE(title, ARRAY_TO_STRING(split(name, '-')[2:], '-')))) AS title,
-                PARSE_DATE(release) AS release,
-                CAST(price AS FLOAT) AS price,
-                CONCAT('https://www.sounds.nl/detail/', id) AS url
-            FROM read_parquet('{storage_dir}/{asset_key}/*/sounds.parquet', union_by_name = true)
-        ), tmp_both AS (
-            SELECT * FROM tmp_plato UNION ALL SELECT * FROM tmp_sounds
-        ), tmp_rn AS (
-            SELECT
-                *,
-                ROW_NUMBER() OVER(PARTITION BY source, id, artist, title, price ORDER BY date DESC) as rn
-            FROM tmp_both
-        )
-        SELECT
-            source,
-            date,
-            id,
-            artist,
-            title,
-            release,
-            price,
-            url
-        FROM tmp_rn
-        WHERE rn = 1
-        ORDER BY date ASC
-        """
-        ).df()
+                artist,
+                title,
+                release,
+                price,
+                url
+            FROM tmp_rn
+            WHERE rn = 1
+            ORDER BY date ASC
+            """
+        ).pl()
 
 
 @asset(
-    # deps=[new_deals],
-    io_manager_key="duckdb_io_manager"
+    io_manager_key="polars_parquet_io_manager",
 )
-def works(new_deals: pd.DataFrame) -> pd.DataFrame:
-    return new_deals[["artist", "title"]].drop_duplicates()
-    # with duckdb.get_connection() as con:
-    #     return con.execute(
-    #         "SELECT DISTINCT artist, title, release FROM vinyl.public.new_deals"
-    #     ).df()
+def works(new_deals: pl.DataFrame) -> pl.DataFrame:
+    # Pandas
+    # columns = ["artist", "title"]
+    # return pl.from_pandas(new_deals[columns].to_pandas().drop_duplicates())
+
+    # Polars
+    # return new_deals[columns].unique(subset=columns)
+
+    # DuckDB
+    with duckdb.connect() as con:
+        return con.execute("SELECT DISTINCT artist, title, release FROM new_deals").pl()

src/app/vinyl/repo.py

@@ -23,8 +23,7 @@ vinyl = Definitions(
     assets=[deals, new_deals, works],
     resources={
         "polars_parquet_io_manager": PolarsParquetIOManager(),
-        "duckdb_io_manager": PandasDuckDBIOManager(database="vinyl.duckdb"),
-        # "duckdb": DuckDBResource(database="vinyl.duckdb")
+        "duckdb_io_manager": PandasDuckDBIOManager(database="vinyl"),
     },
     jobs=[deals_job, check_partititions_job, musicbrainz_lookup_job],
     schedules=[deals_schedule],