Provides a more structured formatter for python's argparse module.
pip install argformat
Argformat provides the argformat.StructuredFormatter class that can be set as formatter_class of ArgumentParser.
The StructuredFormatter changes the output format in the following ways:
- It adds
(default = <value>)to the-h/--helpoutput if a default value is given. - It outputs the help message in column format.
- It ignores the default width of 80 characters for the
-h/--helpoutput (warning: output may be distorted on smaller terminals).
Below we provide the difference in output for the project (DeepCASE) that uses the argformat library.
usage: deepcase.py [-h] [--offset OFFSET] [--time TIME] [--all] [--breach] [--incident] [--ignore] [-f FEATURES] [-i DIM_INPUT]
[-o DIM_OUTPUT] [-m MAX_SEQUENCES] [-n MAX_EVENTS] [-c COMPLEXITY] [-b BATCH_SIZE] [-d DEVICE] [-e EPOCHS] [-r]
[-s] [--train TRAIN] [--epsilon EPSILON] [--min-samples MIN_SAMPLES] [--threshold THRESHOLD]
[--load-context LOAD_CONTEXT] [--load-interpreter LOAD_INTERPRETER] [--save-context SAVE_CONTEXT]
[--save-interpreter SAVE_INTERPRETER]
file [test] [malicious]
DeepCASE: providing contextual analysis of security alerts
optional arguments:
-h, --help show this help message and exit
Input:
file read preprocessed input file
test read preprocessed test file
malicious read preprocessed malicious file
--offset OFFSET offset for items to load
--time TIME max time length of input sequence
--all perform experiment on all data
--breach read breaches
--incident read incidents
--ignore ignore incident and breach info
ContextBuilder:
-f FEATURES, --features FEATURES
maximum number of expected features
-i DIM_INPUT, --dim-input DIM_INPUT
length of input sequence
-o DIM_OUTPUT, --dim-output DIM_OUTPUT
length of output sequence
-m MAX_SEQUENCES, --max-sequences MAX_SEQUENCES
maximum number of sequences ro read from input
-n MAX_EVENTS, --max-events MAX_EVENTS
maximum number of events to read from input
-c COMPLEXITY, --complexity COMPLEXITY
complexity of the model
ContextBuilder training:
-b BATCH_SIZE, --batch-size BATCH_SIZE
batch size
-d DEVICE, --device DEVICE
train using given device (cpu|cuda|auto)
-e EPOCHS, --epochs EPOCHS
number of epochs to train with
-r, --random train with random selection
-s, --silent supress printing progress
--train TRAIN training samples to use (or ratio of if 0 <= TRAIN <= 1)
Interpreter:
--epsilon EPSILON interpreter epsilon for clustering
--min-samples MIN_SAMPLES
interpreter min_samples for clustering
--threshold THRESHOLD
interpreter confidence threshold for fingerprinting
Model I/O parameters:
--load-context LOAD_CONTEXT
load context builder from LOAD file
--load-interpreter LOAD_INTERPRETER
load interpreter from LOAD file
--save-context SAVE_CONTEXT
save context builder to SAVE file
--save-interpreter SAVE_INTERPRETER
save interpreter to SAVE file
usage: deepcase.py [-h] [--offset OFFSET] [--time TIME] [--all] [--breach] [--incident] [--ignore] [-f FEATURES] [-i DIM_INPUT]
[-o DIM_OUTPUT] [-m MAX_SEQUENCES] [-n MAX_EVENTS] [-c COMPLEXITY] [-b BATCH_SIZE] [-d DEVICE] [-e EPOCHS] [-r] [-s]
[--train TRAIN] [--epsilon EPSILON] [--min-samples MIN_SAMPLES] [--threshold THRESHOLD]
[--load-context LOAD_CONTEXT] [--load-interpreter LOAD_INTERPRETER] [--save-context SAVE_CONTEXT]
[--save-interpreter SAVE_INTERPRETER]
file [test] [malicious]
DeepCASE: providing contextual analysis of security alerts
optional arguments:
-h, --help show this help message and exit
Input:
file read preprocessed input file
test read preprocessed test file (optional)
malicious read preprocessed malicious file (optional)
--offset OFFSET offset for items to load
--time TIME max time length of input sequence (default = 86400)
--all perform experiment on all data
--breach read breaches
--incident read incidents
--ignore ignore incident and breach info
ContextBuilder:
-f, --features FEATURES maximum number of expected features (default = 280)
-i, --dim-input DIM_INPUT length of input sequence (default = 10)
-o, --dim-output DIM_OUTPUT length of output sequence (default = 1)
-m, --max-sequences MAX_SEQUENCES maximum number of sequences ro read from input (default = inf)
-n, --max-events MAX_EVENTS maximum number of events to read from input (default = inf)
-c, --complexity COMPLEXITY complexity of the model (default = 128)
ContextBuilder training:
-b, --batch-size BATCH_SIZE batch size (default = 128)
-d, --device DEVICE train using given device (cpu|cuda|auto) (default = auto)
-e, --epochs EPOCHS number of epochs to train with (default = 10)
-r, --random train with random selection
-s, --silent supress printing progress
--train TRAIN training samples to use (or ratio of if 0 <= TRAIN <= 1) (default = 0.5)
Interpreter:
--epsilon EPSILON interpreter epsilon for clustering (default = 0.1)
--min-samples MIN_SAMPLES interpreter min_samples for clustering (default = 5)
--threshold THRESHOLD interpreter confidence threshold for fingerprinting (default = 0.2)
Model I/O parameters:
--load-context LOAD_CONTEXT load context builder from LOAD file
--load-interpreter LOAD_INTERPRETER load interpreter from LOAD file
--save-context SAVE_CONTEXT save context builder to SAVE file
--save-interpreter SAVE_INTERPRETER save interpreter to SAVE file
Simply import the StructuredFormatter and set it as the argparse formatter_class:
import argparse
from argformat import StructuredFormatter
# Parse arguments
parser = argparse.ArgumentParser(
prog = <Your program>,
description = <Your description>,
formatter_class = StructuredFormatter,
)
# Add your arguments normally...