Merge pull request #21 from radical-cybertools/seavea-hackathon-2025

SEAVEA Hackathon 2025

src/seavea-hackathon-2025/README.md

@@ -0,0 +1,31 @@
+# Docker container with RADICAL-Cybertools (RCT) tutorial notebooks
+
+[![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/radical-cybertools/tutorials/main)
+
+> [!IMPORTANT]
+> [Binder](https://mybinder.readthedocs.io) lets you run and test our RCT
+> notebooks. You should **not** expect Binder to match the performance
+> achievable by RCT on local or high performance computing (HPC) platforms.
+> If performance is a consideration, please consider executing our RCT Docker
+> container locally or in a suitable HPC environment.
+
+## How to run
+
+The RCT Tutorials container is based on
+[jupyter/minimal-notebook](https://github.com/jupyter/docker-stacks) image.
+We prepared a container image with all the latest tutorials in this
+repository, and uploaded this image to
+[DockerHub](https://hub.docker.com/u/radicalcybertools)
+(`radicalcybertools/tutorials:seavea-hackathon-2025`). The following command
+pulls and runs the RCT Tutorials container locally (**NOTE**: `docker` is
+required to be installed locally):
+
+```shell
+docker run --rm -it -p 8888:8888 radicalcybertools/tutorials:seavea-hackathon-2025
+```
+
+## Documentation
+
+* RADICAL-EnTK:  https://radicalentk.readthedocs.io/
+* RADICAL-Pilot: https://radicalpilot.readthedocs.io/
+

src/seavea-hackathon-2025/drafts/entk-darshan/darshan-enabled.ipynb

@@ -0,0 +1,212 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "d7060a86-eda9-46d7-a6a7-dc1dd0eeaa46",
+   "metadata": {},
+   "source": [
+    "# RADICAL-EnTK with Darshan to generate workflow provenance graph\n",
+    "\n",
+    "## Darshan\n",
+    "\n",
+    "https://wordpress.cels.anl.gov/darshan/\n",
+    "\n",
+    "Darshan is an HPC I/O characterization tool. It is designed to capture an accurate picture of application I/O behavior, including properties such as patterns of access within files, with minimum overhead. Darshan can be used to investigate and tune the I/O behavior of complex HPC applications."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b0f73a4a-2388-4203-bb5d-efd04aa5ce60",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%%capture\n",
+    "# install darshan or ensure that it is installed\n",
+    "!./scripts/darshan-install.sh"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5c6b8ba2-48f8-496e-8f32-bec3cb2c27dc",
+   "metadata": {},
+   "source": [
+    "## Example of enabling Darshan\n",
+    "\n",
+    "In the example, Darshan is enabled for a specific `re.Task` and for all tasks within a specific `re.Stage`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "db41fa7d-c485-4b10-bc2e-d90c252a61b3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "\n",
+    "import radical.entk  as re\n",
+    "import radical.pilot as rp\n",
+    "\n",
+    "from radical.entk.tools import (cache_darshan_env,\n",
+    "                                with_darshan,\n",
+    "                                enable_darshan,\n",
+    "                                get_provenance_graph)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "05556167-3618-4476-913a-195e52d2c27d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%env RADICAL_REPORT_ANIME=FALSE\n",
+    "%env RADICAL_REPORT=TRUE\n",
+    "%env RADICAL_LOG_LVL=DEBUG"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "bebd98ac-c4eb-42cd-be6c-781cc6d5af91",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%%capture\n",
+    "# prepare darshan environment\n",
+    "#    runtime root dir  (if not provided then env variable $DARSHAN_RUNTIME_ROOT should be set)\n",
+    "#    modules           (optional)\n",
+    "#    env variables     (optional)\n",
+    "cache_darshan_env(darshan_runtime_root='/usr/local',\n",
+    "                  modules=[], env={})"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3def0316-c6b2-414e-8cc3-113e227c9e25",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "TASK_01_OUTPUT = 'output_01.dat'\n",
+    "\n",
+    "\n",
+    "def get_stage_0():\n",
+    "\n",
+    "    # hello-RP task\n",
+    "    task_00 = re.Task({\n",
+    "        'executable': 'radical-pilot-hello.sh',\n",
+    "        'arguments' : [10],\n",
+    "        'cpu_reqs'  : {'cpu_processes'  : 1,\n",
+    "                       'cpu_threads'    : 4,\n",
+    "                       'cpu_thread_type': rp.OpenMP}\n",
+    "    })\n",
+    "\n",
+    "    # R/W data\n",
+    "    task_01   = re.Task({\n",
+    "        'executable'       : '/bin/sh',\n",
+    "        'arguments'        : ['-c', f'cat input.dat | wc > {TASK_01_OUTPUT}'],\n",
+    "        'upload_input_data': ['/etc/passwd > input.dat'],\n",
+    "        'copy_output_data' : [f'{TASK_01_OUTPUT} > $SHARED/{TASK_01_OUTPUT}']\n",
+    "    })\n",
+    "\n",
+    "    stage_0 = re.Stage()\n",
+    "    # --- enable Darshan for task \"task_01\" only\n",
+    "    stage_0.add_tasks([task_00, enable_darshan(task_01)])\n",
+    "    return stage_0\n",
+    "\n",
+    "\n",
+    "# --- enable Darshan for the whole \"stage_1\" using decorator\n",
+    "@with_darshan\n",
+    "def get_stage_1():\n",
+    "\n",
+    "    # R/W data and task depends on the task from the previous stage\n",
+    "    task_10 = re.Task({\n",
+    "        'executable'     : '/bin/sh',\n",
+    "        'arguments'      : ['-c',\n",
+    "                            f\"sed -r 's/\\s+//g' {TASK_01_OUTPUT} \" +  # noqa: W605\n",
+    "                            '| grep -o . | sort | uniq -c > output_10.dat'],\n",
+    "        'copy_input_data': [f'$SHARED/{TASK_01_OUTPUT} > {TASK_01_OUTPUT}']\n",
+    "    })\n",
+    "\n",
+    "    stage_1 = re.Stage()\n",
+    "    stage_1.add_tasks([task_10])\n",
+    "    return stage_1"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e49bc965-f69c-4416-8426-5a1b2c1b96aa",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "amgr = re.AppManager()\n",
+    "amgr.resource_desc = {'resource': 'local.localhost',\n",
+    "                      'cpus'    : 8,\n",
+    "                      'walltime': 15}\n",
+    "\n",
+    "pipeline = re.Pipeline()\n",
+    "pipeline.add_stages([get_stage_0(), get_stage_1()])\n",
+    "\n",
+    "amgr.workflow = [pipeline]\n",
+    "amgr.run()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "abb5ec82-ba30-460b-963f-b5401394f58e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "sid = amgr.sid\n",
+    "\n",
+    "!ls -al ~/radical.pilot.sandbox/\"$sid\"/pilot.0000/task.000001/darshan_logs"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8d41123e-23a6-4057-a737-192cc41fb5ab",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from pprint import pformat\n",
+    "\n",
+    "print(pformat(get_provenance_graph(pipelines=[pipeline],\n",
+    "                                   output_file='entk_provenance.json')))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "6396c7bf-731c-472d-9276-104d68c7e736",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

src/seavea-hackathon-2025/drafts/entk-darshan/scripts/darshan-install.sh

@@ -0,0 +1,37 @@
+#!/bin/bash
+
+if ! test -z "$(which darshan-config)"; then
+  echo "Darshan installed"
+  exit 0
+fi
+
+SCRIPTS_DIR="$( cd -- "$( dirname -- "${BASH_SOURCE[0]:-$0}"; )/" &> /dev/null \
+                && pwd 2> /dev/null; )"
+
+# patch EnTK with Darshan related fixes
+"$SCRIPTS_DIR/entk-patched.sh"
+
+(sudo apt-get update  -y && \
+ sudo apt-get install -y \
+      autoconf build-essential cmake libtool tar) \
+ &> /dev/null || true
+
+DARSHAN_VER=3.4.6
+
+cd "$HOME" || true
+wget -q "https://web.cels.anl.gov/projects/darshan/releases/darshan-$DARSHAN_VER.tar.gz"
+tar -xvzf "darshan-$DARSHAN_VER.tar.gz"
+
+cd "$HOME/darshan-$DARSHAN_VER/darshan-runtime/" || true
+../prepare.sh
+MAKE=gmake ./configure \
+  --prefix="/usr/local/" \
+  --with-log-path-by-env=DARSHAN_LOG_DIR_PATH \
+  --with-jobid-env=NONE --without-mpi CC=gcc
+sudo make && sudo make install
+
+cd "$HOME/darshan-$DARSHAN_VER/darshan-util" || true
+../prepare.sh
+./configure --prefix="/usr/local/"
+sudo make && sudo make install
+

src/seavea-hackathon-2025/drafts/entk-darshan/scripts/entk-patched.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+SCRIPTS_DIR="$( cd -- "$( dirname -- "${BASH_SOURCE[0]:-$0}"; )/" &> /dev/null \
+                && pwd 2> /dev/null; )"
+
+REPO_DIR="$HOME/entk"
+git clone --single-branch --branch master \
+    https://github.com/radical-cybertools/radical.entk.git "$REPO_DIR"
+
+cd "$REPO_DIR" || true
+git apply "$SCRIPTS_DIR/entk.patch"
+pip install .
+

src/seavea-hackathon-2025/drafts/entk-darshan/scripts/entk.patch

@@ -0,0 +1,34 @@
+diff --git a/src/radical/entk/tools/darshan.py b/src/radical/entk/tools/darshan.py
+index c64a005a..f47ddc70 100644
+--- a/src/radical/entk/tools/darshan.py
++++ b/src/radical/entk/tools/darshan.py
+@@ -103,7 +103,7 @@ def cache_darshan_env(darshan_runtime_root: Optional[str] = None,
+         if out is not None:
+             out = out.strip()
+
+-        if ret or not out:
++        if ret or not out or 'DARSHAN_LOG_DIR_PATH' in out:
+             print(f'[WARNING] Darshan log path not set: "{err}"')
+             _darshan_log_path = '%(sandbox)s/darshan_logs'
+
+@@ -245,8 +245,9 @@ def get_parsed_data(log: str, target_counters: Union[str, List[str]]) -> set:
+         return data
+
+     grep_patterns = '-e ' + ' -e '.join(ru.as_list(target_counters))
+-    parser_cmd    = (f'darshan-parser {log} | grep {grep_patterns} | '
+-                     "awk '{print $5\":\"$6}'")
++    parser_cmd    = (f'env LD_PRELOAD="{_darshan_runtime_root}'
++                     f'/lib/libdarshan-util.so" darshan-parser {log} | '
++                     f'grep {grep_patterns} | ' + "awk '{print $5\":\"$6}'")
+     out, err, ret = ru.sh_callout(parser_cmd, env=_darshan_env, shell=True)
+     if ret:
+         print(f'[ERROR] Darshan not able to parse "{log}": {err}')
+@@ -282,7 +283,7 @@ def annotate_task_with_darshan(task: Task) -> None:
+
+     log_files = []
+     if '%(sandbox)s' in _darshan_log_path:
+-        for log in glob.glob(_darshan_log_path % {'sandbox': task.path}):
++        for log in glob.glob(_darshan_log_path % {'sandbox': task.path} + '/*'):
+             log_files.append(log)
+     else:
+         if _start_datetime is None:

src/seavea-hackathon-2025/environment.yml

@@ -0,0 +1,12 @@
+dependencies:
+  - openmpi
+  - pip
+  - setuptools
+  - wheel
+  - pip:
+      - git+https://github.com/radical-cybertools/radical.utils.git@devel
+      - git+https://github.com/radical-cybertools/radical.gtod.git@devel
+      - git+https://github.com/radical-cybertools/radical.pilot.git@devel
+      - git+https://github.com/radical-cybertools/radical.entk.git@devel
+      - git+https://github.com/radical-cybertools/radical.analytics.git@devel
+