add memory usage/estimation quick parser utility (#26)

CHANGELOG.rst

@@ -1,9 +1,9 @@
 =========
 Changelog
 =========
-`Version 0.4.1`
-===============
-* 
+`Unreleased`
+================================================================================================
+* 🧮 Evaluates memory usage for specific nodes from `callback.log` files
 
 `Version 0.4.0: Goodbye Singularity Hub <https://github.com/FCP-INDI/cpac/releases/tag/v0.4.0>`_
 ================================================================================================

requirements.txt

@@ -6,4 +6,5 @@ setuptools
 spython >= 0.0.81
 tabulate >= 0.8.6
 tornado
-websocket-client
+websocket-client
+rich

src/cpac/helpers/cpac_parse_resources.py

@@ -0,0 +1,93 @@
+#!/usr/bin/env python
+'''cpac_parse_resources.py
+
+`cpac_parse resources` is intended to be run outside a C-PAC container
+'''
+
+from rich.console import Console
+from rich.table import Table
+
+from argparse import ArgumentParser
+import pandas as pd
+import numpy as np
+import json
+
+
+runti = 'runtime_memory_gb'
+estim = 'estimated_memory_gb'
+
+field = {'runtime': runti,
+         'estimate': estim,
+         'efficiency': 'efficiency'}
+
+
+def display(df):
+    console = Console()
+
+    table = Table(show_header=True, header_style="bold magenta")
+    table.add_column("Task ID", style="dim", width=40)
+    table.add_column("Memory Used")
+    table.add_column("Memory Estimated")
+    table.add_column("Memory Efficiency")
+
+    for _, d in df.iterrows():
+        tmp = list()
+        tmp += [d['id']]
+        tmp += [d[runti]]
+        tmp += [d[estim]]
+        tmp += ["{0:.2f} %".format(100*d[runti] * 1.0 / d[estim])]
+
+        tmp = ["{0:.4f}".format(t) if isinstance(t, float) else str(t)
+               for t in tmp]
+        table.add_row(*tmp)
+        del tmp
+
+    console.print(table)
+
+
+def load_runtime_stats(callback):
+    with open(callback) as fhandle:
+        logs = [json.loads(log) for log in fhandle.readlines()]
+
+    pruned_logs = []
+    for log in logs:
+        if runti not in log.keys():
+            continue
+
+        tmp = {}
+        for k in ['id', runti, estim]:
+            tmp[k] = log[k]
+        tmp['efficiency'] = tmp[runti] / tmp[estim] * 100
+
+        pruned_logs += [tmp]
+        del tmp
+
+    return pd.DataFrame.from_dict(pruned_logs)
+
+
+def query(usage, f, g, c):
+    order = True if g == 'lowest' else False
+    usage.sort_values(by=field[f], ascending=order, inplace=True)
+    usage.reset_index(inplace=True, drop=True)
+    return usage[0:c]
+
+
+if __name__ == '__main__':
+    parser = ArgumentParser(__file__)
+    parser.add_argument("callback",
+                        help="callback.log file found in the 'log' "
+                             "directory of the specified derivatives path")
+    parser.add_argument("--filter_field", "-f", action="store",
+                        choices=['runtime', 'estimate', 'efficiency'],
+                        default='efficiency')
+    parser.add_argument("--filter_group", "-g", action="store",
+                        choices=['lowest', 'highest'], default='lowest')
+    parser.add_argument("--filter_count", "-n", action="store", type=int,
+                        default=10)
+
+    res = parser.parse_args()
+    usage = load_runtime_stats(res.callback)
+
+    filtered_usage = query(usage, res.filter_field, res.filter_group,
+                           res.filter_count)
+    display(filtered_usage)