From c7d242233403594d7a5422344a49a16c2c738222 Mon Sep 17 00:00:00 2001
From: Munyoki Kilyungi
Date: Fri, 3 Feb 2023 12:58:13 +0300
Subject: Add perf scripts that compare fetching a trait using GN API vs LMDB

* scripts/performance/(README,org, timeit_gn2.py, timeit_lmdb.py): New
files.

Signed-off-by: Munyoki Kilyungi <me@bonfacemunyoki.com>
---
 scripts/performance/timeit_lmdb.py | 36 ++++++++++++++++++++++++++++++++++++
 1 file changed, 36 insertions(+)
 create mode 100644 scripts/performance/timeit_lmdb.py

(limited to 'scripts/performance/timeit_lmdb.py')

diff --git a/scripts/performance/timeit_lmdb.py b/scripts/performance/timeit_lmdb.py
new file mode 100644
index 00000000..17514667
--- /dev/null
+++ b/scripts/performance/timeit_lmdb.py
@@ -0,0 +1,36 @@
+import sys
+import timeit
+
+print(timeit.timeit(
+"""
+PATH = "/home/munyoki/tmp/dataset/HLCPublish/10001/"
+env = lmdb.open(PATH)
+
+BLOB_HASH_DIGEST = 32
+
+# def index_matrix(row_pointers):
+
+with env.begin(write=False) as txn:
+    current_hash = txn.get(b"current")
+    matrix_hash = txn.get(current_hash + b":matrix")
+    row_pointers = txn.get(matrix_hash +
+                           b":row-pointers")
+    nrows, = struct.unpack("<Q",
+                           txn.get(matrix_hash + b":nrows"))
+    metadata = txn.get(matrix_hash + b":metadata")
+    sample_data = []
+    for i in range(0, (nrows-1)*32, 32):
+        sample_data.append(
+            json.loads(txn.get(row_pointers[i:i+32]).decode())
+        )
+    print(sample_data)
+    print(metadata.decode())
+""",
+    setup="""
+import struct
+import array
+import json
+import lmdb
+""",
+    number=int(sys.argv[1])
+))
-- 
cgit v1.2.3