Merge pull request #2885 from finos/strptime-wasm

Better `datetime`/`date` string parsing performance

cmake/arrow.txt.in

@@ -8,7 +8,7 @@ cmake_policy(SET CMP0097 NEW)
 include(ExternalProject)
 ExternalProject_Add(apachearrow
   GIT_REPOSITORY    https://github.com/apache/arrow.git
-  GIT_TAG           apache-arrow-17.0.0
+  GIT_TAG           apache-arrow-18.1.0
   GIT_SUBMODULES    ""
   GIT_SHALLOW       TRUE
   SOURCE_DIR        "${CMAKE_BINARY_DIR}/arrow-src"
@@ -20,5 +20,5 @@ ExternalProject_Add(apachearrow
   TEST_COMMAND      ""
   # This patch is to work around https://github.com/apache/arrow/issues/44384
   # It can be removed when a version of Arrow is released with https://github.com/apache/arrow/pull/44385
-  PATCH_COMMAND     "${CMAKE_COMMAND}" -E chdir <SOURCE_DIR> git apply "${CMAKE_SOURCE_DIR}/patches/fix_arrow_libtool.patch"
+  PATCH_COMMAND     "${CMAKE_COMMAND}" -E chdir <SOURCE_DIR> git apply "${CMAKE_SOURCE_DIR}/patches/fix_arrow_libtool.patch" && "${CMAKE_COMMAND}" -E chdir <SOURCE_DIR> git apply "${CMAKE_SOURCE_DIR}/patches/arrow_strptime.patch"
 )

cpp/perspective/patches/arrow_strptime.patch

@@ -0,0 +1,13 @@
+diff --git a/cpp/src/arrow/util/value_parsing.h b/cpp/src/arrow/util/value_parsing.h
+index 609906052..1e3dfae7c 100644
+--- a/cpp/src/arrow/util/value_parsing.h
++++ b/cpp/src/arrow/util/value_parsing.h
+@@ -804,7 +804,7 @@ static inline bool ParseTimestampStrptime(const char* buf, size_t length,
+   std::string clean_copy(buf, length);
+   struct tm result;
+   memset(&result, 0, sizeof(struct tm));
+-#ifdef _WIN32
++#if defined(_WIN32) || defined(__EMSCRIPTEN__)
+   char* ret = arrow_strptime(clean_copy.c_str(), format, &result);
+ #else
+   char* ret = strptime(clean_copy.c_str(), format, &result);

tools/perspective-bench/basic_suite.mjs

@@ -62,6 +62,8 @@ perspective_bench.suite(
             metadata = { version, version_idx };
         }
 
+        globalThis.__SEND__ = (x) => process.send(x);
+
         await all_benchmarks.table_suite(client, metadata);
         await all_benchmarks.view_suite(client, metadata);
         await all_benchmarks.to_data_suite(client, metadata);

tools/perspective-bench/cross_platform_suite.mjs

@@ -158,8 +158,10 @@ export async function table_suite(perspective, metadata) {
             const table = await perspective.table(
                 new_superstore_table(metadata)
             );
+
             const view = await table.view();
             const csv = await view.to_csv();
+            const arrow = await view.to_arrow();
             const json = await view.to_json();
             const columns = await view.to_columns();
             if (check_version_gte(metadata.version, "2.10.9")) {
@@ -169,7 +171,8 @@ export async function table_suite(perspective, metadata) {
             if (check_version_gte(metadata.version, "3.0.0")) {
                 await table.delete();
             }
-            return { csv, columns, json };
+
+            return { csv, arrow, table, json, columns };
         } catch (e) {
             console.error(e);
         }
@@ -184,8 +187,8 @@ export async function table_suite(perspective, metadata) {
                 await table.delete();
             }
         },
-        async test() {
-            return await perspective.table(new_superstore_table(metadata));
+        async test({ table, arrow }) {
+            return await perspective.table(arrow.slice());
         },
     });
 
@@ -198,7 +201,7 @@ export async function table_suite(perspective, metadata) {
                 await table.delete();
             }
         },
-        async test({ table, csv }) {
+        async test({ csv }) {
             return await perspective.table(csv);
         },
     });

tools/perspective-bench/puppeteer_suite.mjs

@@ -0,0 +1,87 @@
+// ┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
+// ┃ ██████ ██████ ██████       █      █      █      █      █ █▄  ▀███ █       ┃
+// ┃ ▄▄▄▄▄█ █▄▄▄▄▄ ▄▄▄▄▄█  ▀▀▀▀▀█▀▀▀▀▀ █ ▀▀▀▀▀█ ████████▌▐███ ███▄  ▀█ █ ▀▀▀▀▀ ┃
+// ┃ █▀▀▀▀▀ █▀▀▀▀▀ █▀██▀▀ ▄▄▄▄▄ █ ▄▄▄▄▄█ ▄▄▄▄▄█ ████████▌▐███ █████▄   █ ▄▄▄▄▄ ┃
+// ┃ █      ██████ █  ▀█▄       █ ██████      █      ███▌▐███ ███████▄ █       ┃
+// ┣━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┫
+// ┃ Copyright (c) 2017, the Perspective Authors.                              ┃
+// ┃ ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌ ┃
+// ┃ This file is part of the Perspective library, distributed under the terms ┃
+// ┃ of the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0). ┃
+// ┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
+
+import * as all_benchmarks from "./cross_platform_suite.mjs";
+import * as perspective_bench from "./src/js/benchmark.mjs";
+import * as puppeteer from "puppeteer";
+
+import * as fs from "node:fs";
+import * as path from "node:path";
+import * as url from "node:url";
+import * as process from "node:process";
+
+const __dirname = url.fileURLToPath(new URL(".", import.meta.url)).slice(0, -1);
+
+/**
+ * We use the `dependencies` of this package for the benchmark candidate
+ * module list, so that we only need specify the dependencies and benchmark
+ * candidates in one place.
+ */
+const VERSIONS = [
+    "@finos/perspective",
+    "perspective-3-0-0",
+    "perspective-2-10-0",
+];
+
+perspective_bench.suite(
+    [...VERSIONS],
+    path.join(__dirname, "dist/benchmark-js.arrow"),
+    async function (path, version_idx) {
+        let client, metadata;
+        console.log(path);
+        const browser = await puppeteer.launch({
+            headless: true,
+            protocolTimeout: 100_000_000,
+        });
+        const page = await browser.newPage();
+
+        await page.goto("http://localhost:8081/empty.html");
+
+        async function test_suite(suite) {
+            const items = await page.evaluate(
+                async ([version, suite]) => {
+                    const { default: perspective } = await import(
+                        `/tools/perspective-bench/node_modules/${version}/dist/esm/perspective.inline.js`
+                    );
+                    const benchmarks = await import(
+                        "/tools/perspective-bench/cross_platform_suite.mjs"
+                    );
+
+                    const metadata = {
+                        version: "3.2.0",
+                        version_idx: 0,
+                    };
+                    const total = [];
+                    window.__SEND__ = (x) => {
+                        total.push(x);
+                    };
+
+                    await benchmarks[suite](
+                        await perspective.worker(),
+                        metadata
+                    );
+
+                    return total;
+                },
+                [path, suite]
+            );
+
+            for (const { obs_records, stats } of items) {
+                process.send({ obs_records, stats });
+            }
+        }
+
+        await test_suite("table_suite");
+        await test_suite("view_suite");
+        await test_suite("to_data_suite");
+    }
+);