Early version of profiler harness

.github/workflows/profiler.yml

@@ -0,0 +1,30 @@
+name: Performance check
+
+on:
+  pull_request:
+
+jobs:
+  run_profiler:
+    name: Run Profiler
+    runs-on: ubuntu-latest
+    permissions: # these permissions must be set for AWS auth to work!
+      id-token: write
+      contents: read
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Configure AWS Credentials
+        uses: aws-actions/configure-aws-credentials@v1
+        with:
+          aws-region: us-west-2
+          role-to-assume: ${{ secrets.AWS_ROLE_TO_ASSUME }}
+          role-session-name: PushDockerImage
+
+      - name: Run all tests
+        run: |
+          python -m venv profiler_env
+          source profiler_env/bin/activate
+          ./tools/perf_checker/perf_checker.sh

tools/perf_checker/benchmark1.py

@@ -0,0 +1,22 @@
+from time import perf_counter
+
+import cellxgene_census
+
+import tiledbsoma as soma
+
+census_S3_latest = dict(census_version="2023-10-23")
+
+
+def main():
+    t1 = perf_counter()
+    with cellxgene_census.open_soma(**census_S3_latest) as census:
+        with census["census_data"]["homo_sapiens"].axis_query(
+            measurement_name="RNA",
+            obs_query=soma.AxisQuery(value_filter="""tissue_general == 'eye'"""),
+        ) as query:
+            query.to_anndata(X_name="raw")
+    t2 = perf_counter()
+    print(f"End to end time {t2 - t1}")
+
+
+main()

tools/perf_checker/perf_checker.sh

@@ -0,0 +1,36 @@
+.sh
+#!/bin/sh
+set -euox pipefail
+
+python -m venv perf
+source perf/bin/activate
+pip install psutil
+pip install gitpython
+pip install somacore
+pip install tiledbsoma
+pip install cellxgene_census
+
+
+sudo wget https://s3.amazonaws.com/mountpoint-s3-release/latest/x86_64/mount-s3.deb
+sudo apt install -y ./mount-s3.deb
+mkdir ./mount-s3
+mkdir ./s3_cache
+mount-s3 census-profiler-tests ./mount-s3 --cache ./s3_cache  --metadata-ttl 300
+
+# new benchmarks must be added to this list
+declare -a benchmarks=("benchmark1.py")
+
+git clone https://github.com/single-cell-data/TileDB-SOMA.git
+pip install TileDB-SOMA/profiler
+cd TileDB-SOMA
+git checkout temp_profiler_for_harnes
+cd ../
+
+pip list | grep profiler
+arraylength=${#benchmarks[@]}
+for (( i=0; i<${arraylength}; i++ ))
+do
+  python  ./TileDB-SOMA/profiler "python ${benchmarks[$i]}" -t time
+  pwd
+  python ./tools/perf_checker/profile_report.py ${benchmarks[$i]}
+done

tools/perf_checker/profile_report.py

@@ -0,0 +1,33 @@
+import profiler.data
+import argparse
+
+parser = argparse.ArgumentParser()
+parser.add_argument("benchmark", type=str)
+args = parser.parse_args()
+
+# Processes the set of previously written logs
+
+threshold = 1.10  # Percent difference
+
+db = data.FileBasedProfileDB("./mount-s3")
+actual_max_ts = 0
+dt = db.find(f"python {parser.benchmark}")
+last_two = dt[-2:]
+c = 0
+
+for s in last_two:
+    new_db = sorted(dt, key=lambda ProfileData: ProfileData.timestamp)
+
+    L = []
+    L[0] = dt[0].user_time_sec + dt[0].elapsed_time
+    L[1] = dt[1].user_time_sec + dt[1].elapsed_time
+    for i in range(0, len(dt)):
+        print(f"{i} dt[{i}].user_time_sec = {dt[i].user_time_sec} ts {dt[i].timestamp}")
+    print(f"Prev = {L[0]} Curr = {L[1]}")
+
+    if threshold * float(L[1]) < float(L[0]) or float(L[1]) > threshold * float(L[0]):
+        raise SystemExit(f"Potential performance degradation detected {L[0]} va {L[1]}")
+    print("No recent performance degradation detected")
+    print(
+        f"Prev TBD version = {dt[0].tiledbsoma_version} Curr TBD version = {dt[1].tiledbsoma_version}"
+    )