From 36f5b4f043ac211986172b8d32895bfedd43f750 Mon Sep 17 00:00:00 2001
From: Kenny Daniel <platypii@gmail.com>
Date: Sun, 26 May 2024 06:00:20 -0700
Subject: [PATCH] Move decompressPage to avoid circular dependency chain

---
 src/column.js     | 30 +-----------------------------
 src/datapage.js   | 28 ++++++++++++++++++++++++++++
 src/datapageV2.js |  2 +-
 3 files changed, 30 insertions(+), 30 deletions(-)

diff --git a/src/column.js b/src/column.js
index dfafc4b..5e856e1 100644
--- a/src/column.js
+++ b/src/column.js
@@ -1,10 +1,9 @@
 import { assembleLists } from './assemble.js'
 import { convertWithDictionary } from './convert.js'
-import { readDataPage, readDictionaryPage } from './datapage.js'
+import { decompressPage, readDataPage, readDictionaryPage } from './datapage.js'
 import { readDataPageV2 } from './datapageV2.js'
 import { parquetHeader } from './header.js'
 import { getMaxDefinitionLevel, getMaxRepetitionLevel } from './schema.js'
-import { snappyUncompress } from './snappy.js'
 import { concat } from './utils.js'
 
 /**
@@ -120,30 +119,3 @@ export function getColumnOffset({ dictionary_page_offset, data_page_offset }) {
   }
   return Number(columnOffset)
 }
-
-/**
- * @param {Uint8Array} compressedBytes
- * @param {number} uncompressed_page_size
- * @param {import('./types.js').CompressionCodec} codec
- * @param {import('./types.js').Compressors | undefined} compressors
- * @returns {Uint8Array}
- */
-export function decompressPage(compressedBytes, uncompressed_page_size, codec, compressors) {
-  /** @type {Uint8Array} */
-  let page
-  const customDecompressor = compressors?.[codec]
-  if (codec === 'UNCOMPRESSED') {
-    page = compressedBytes
-  } else if (customDecompressor) {
-    page = customDecompressor(compressedBytes, uncompressed_page_size)
-  } else if (codec === 'SNAPPY') {
-    page = new Uint8Array(uncompressed_page_size)
-    snappyUncompress(compressedBytes, page)
-  } else {
-    throw new Error(`parquet unsupported compression codec: ${codec}`)
-  }
-  if (page?.length !== uncompressed_page_size) {
-    throw new Error(`parquet decompressed page length ${page?.length} does not match header ${uncompressed_page_size}`)
-  }
-  return page
-}
diff --git a/src/datapage.js b/src/datapage.js
index 4e05b4d..1597f0f 100644
--- a/src/datapage.js
+++ b/src/datapage.js
@@ -1,6 +1,7 @@
 import { bitWidth, byteStreamSplit, readRleBitPackedHybrid } from './encoding.js'
 import { readPlain } from './plain.js'
 import { getMaxDefinitionLevel, getMaxRepetitionLevel } from './schema.js'
+import { snappyUncompress } from './snappy.js'
 
 /**
  * Read a data page from uncompressed reader.
@@ -107,3 +108,30 @@ function readDefinitionLevels(reader, daph, schemaPath) {
 
   return { definitionLevels, numNulls }
 }
+
+/**
+ * @param {Uint8Array} compressedBytes
+ * @param {number} uncompressed_page_size
+ * @param {import('./types.js').CompressionCodec} codec
+ * @param {import('./types.js').Compressors | undefined} compressors
+ * @returns {Uint8Array}
+ */
+export function decompressPage(compressedBytes, uncompressed_page_size, codec, compressors) {
+  /** @type {Uint8Array} */
+  let page
+  const customDecompressor = compressors?.[codec]
+  if (codec === 'UNCOMPRESSED') {
+    page = compressedBytes
+  } else if (customDecompressor) {
+    page = customDecompressor(compressedBytes, uncompressed_page_size)
+  } else if (codec === 'SNAPPY') {
+    page = new Uint8Array(uncompressed_page_size)
+    snappyUncompress(compressedBytes, page)
+  } else {
+    throw new Error(`parquet unsupported compression codec: ${codec}`)
+  }
+  if (page?.length !== uncompressed_page_size) {
+    throw new Error(`parquet decompressed page length ${page?.length} does not match header ${uncompressed_page_size}`)
+  }
+  return page
+}
diff --git a/src/datapageV2.js b/src/datapageV2.js
index 38a45f3..6bb924a 100644
--- a/src/datapageV2.js
+++ b/src/datapageV2.js
@@ -1,4 +1,4 @@
-import { decompressPage } from './column.js'
+import { decompressPage } from './datapage.js'
 import { deltaBinaryUnpack, deltaByteArray, deltaLengthByteArray } from './delta.js'
 import { bitWidth, byteStreamSplit, readRleBitPackedHybrid } from './encoding.js'
 import { readPlain } from './plain.js'