From 142afd106ebf2e2e4aad02144be6c6acd296dc92 Mon Sep 17 00:00:00 2001
From: atl-ggregson <147005258+atl-ggregson@users.noreply.github.com>
Date: Thu, 16 Jan 2025 14:44:59 +1100
Subject: [PATCH] perf: Cache CSV stream schema  (#363)

The stream's `schema` property is accessed multiple times for each
record (see `Stream._generate_record_messages()` for instance). Since
the schema should be static this change caches it, resulting in a
significant performance improvement.

Testing with a sample 2,000,000 row dataset (`people-2000000` from
https://github.com/datablist/sample-csv-files) reduced the read time
from 441 seconds to 48 seconds; about a 10x improvement in throughput.

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
---
 pyproject.toml    | 2 +-
 tap_csv/client.py | 7 +++++--
 2 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/pyproject.toml b/pyproject.toml
index 5aed6b9..fdadda8 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "tap-csv"
-version = "1.1.0"
+version = "1.2.0"
 description = "Singer tap for CSV, built with the Meltano SDK for Singer Taps."
 authors = ["Pat Nadolny"]
 keywords = [
diff --git a/tap_csv/client.py b/tap_csv/client.py
index cae3892..f4b3739 100644
--- a/tap_csv/client.py
+++ b/tap_csv/client.py
@@ -6,6 +6,7 @@
 import os
 import typing as t
 from datetime import datetime, timezone
+from functools import cached_property
 
 from singer_sdk import typing as th
 from singer_sdk.streams import Stream
@@ -121,12 +122,14 @@ def get_rows(self, file_path: str) -> t.Iterable[list]:
         with open(file_path, encoding=encoding) as f:
             yield from csv.reader(f, dialect="tap_dialect")
 
-    @property
+    @cached_property
     def schema(self) -> dict:
         """Return dictionary of record schema.
 
         Dynamically detect the json schema for the stream.
-        This is evaluated prior to any records being retrieved.
+
+        This property is accessed multiple times for each record
+        so it's important to cache the result.
         """
         properties: list[th.Property] = []
         self.primary_keys = self.file_config.get("keys", [])