From 5472dc2c4297b213c3acdfb3582b5caed17d9dcc Mon Sep 17 00:00:00 2001
From: Maximilian Moser <maximilian.moser@tuwien.ac.at>
Date: Fri, 19 Jan 2024 15:20:29 +0100
Subject: [PATCH] Add pickle as an output file format for formatscaper
* also set it as default, because it's much faster and smaller than yaml
---
formatscaper/formatscaper.py | 27 ++++++++++++++++++++++-----
1 file changed, 22 insertions(+), 5 deletions(-)
diff --git a/formatscaper/formatscaper.py b/formatscaper/formatscaper.py
index 0895b1f..9db9f8e 100755
--- a/formatscaper/formatscaper.py
+++ b/formatscaper/formatscaper.py
@@ -2,6 +2,7 @@
import argparse
import dataclasses
+import pickle
import re
import subprocess
import sys
@@ -43,8 +44,15 @@ parser.add_argument(
parser.add_argument(
"--output",
"-o",
- default="results.yml",
- help="file in which to store the identified format for each file (default: results.yml)", # noqa
+ default="results.{FORMAT}",
+ help="file in which to store the identified format for each file (default: results.{FORMAT})", # noqa
+)
+parser.add_argument(
+ "--output-format",
+ "-F",
+ default="pickle",
+ choices=["pickle", "yaml"],
+ help="format of the results (default: pickle)",
)
parser.add_argument(
"--parallel",
@@ -215,12 +223,21 @@ if endangered_files:
# store the results to files
+output_file_name = args.output.format(FORMAT=args.output_format)
try:
- with open(args.output, "w") as output_file:
- yaml.dump([dataclasses.asdict(res) for res in all_results], output_file)
+ simple_results = [dataclasses.asdict(res) for res in all_results]
+ file_mode = "w" if args.output_format == "yaml" else "wb"
+ with open(output_file_name, file_mode) as output_file:
+ if args.output_format == "yaml":
+ yaml.dump(simple_results, output_file)
+ elif args.output_format == "pickle":
+ pickle.dump(simple_results, output_file)
except OSError:
- print(f"WARN: couldn't store the results ({args.output})", file=sys.stderr)
+ print(
+ f"WARN: couldn't store the results to file ({output_file_name})",
+ file=sys.stderr,
+ )
try:
updated_formats = [dataclasses.asdict(f) for f in formats.values()]
--
GitLab