modgraph

modgraph is a tool to explore a collection of tracker module files as a module->sample graph. It can function both as a CLI app, and as a library to use in notebooks. In fact, this file is a notebook!

Using through CLI

!python -m modgraph --help

usage: modgraph [-h] [-f {csv,d2}] [-r RANK] files [files ...]

positional arguments:
  files                 module files to analyze

options:
  -h, --help            show this help message and exit
  -f {csv,d2}, --format {csv,d2}
                        output format
  -r RANK, --rank RANK  min number of repeats for sample to be included

Example:

!python -m modgraph *.it --rank 6 --format csv

mod_path,sample_name,sample_hash
catherine on the waves.it,tambourin.steel.quiet     ,e1b32f84b2b788f0a58e277f4e152df5
catherine on the waves.it,piano.001                 ,8ef52cdf9c20c9ada9df7bf4d3b59fc3
dallying sadly in space.it,                          ,e1b32f84b2b788f0a58e277f4e152df5
drifting to plutonia.it,tambourine.steel.quiet    ,e1b32f84b2b788f0a58e277f4e152df5
heavenly fantasy.it,tambourin.steel.quiet     ,e1b32f84b2b788f0a58e277f4e152df5
neverending illusion.it,piano.001                 ,8ef52cdf9c20c9ada9df7bf4d3b59fc3
"so close to you, my angel.it",piano.001                 ,8ef52cdf9c20c9ada9df7bf4d3b59fc3
sorrow.it,                          ,8ef52cdf9c20c9ada9df7bf4d3b59fc3
sylvia.it,piano.001                 ,8ef52cdf9c20c9ada9df7bf4d3b59fc3
tender storm.it,tambourin.steel.quiet     ,e1b32f84b2b788f0a58e277f4e152df5
why (enhanced version).it,piano.001                 ,8ef52cdf9c20c9ada9df7bf4d3b59fc3
why (enhanced version).it,tambourin.steel.quiet     ,e1b32f84b2b788f0a58e277f4e152df5

Using as a library

import pandas as pd
from modgraph import modgraph
from glob import glob

# digest your library into a mod_path -> sample_hash mapping
df = pd.DataFrame(modgraph(glob("*.it")))
df = df.set_index(['mod_path', 'sample_hash']).sort_index()
df

def most_used(df, cutoff):
    df = df.groupby("sample_hash")
    df = df.agg({"sample_name": [("name", lambda g: g.mode()[0]), "count"]})
    df = df.sort_values(("sample_name", "count"), ascending=False)
    df = df[df[("sample_name", "count")] >= cutoff]
    return df

most_used(df, cutoff=3).plot(kind="barh", x=('sample_name', 'name'))

<AxesSubplot: ylabel='(sample_name, name)'>