Last active
June 28, 2024 09:50
-
-
Save juice500ml/b8584dd6b57a927c0e8a26bff65b6c78 to your computer and use it in GitHub Desktop.
Counting files
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from collections import defaultdict | |
import pkgutil | |
import os | |
import matplotlib as mpl | |
import matplotlib.pyplot as plt | |
from cycler import cycler | |
mpl.rcParams["axes.prop_cycle"] = cycler(color=["#648fff", "#785ef0", "#dc267f", "#fe6100", "#FFB000", "#648fff", "#dc267f", "#000000"]) | |
def _size_count(fname): | |
p = Path(fname) | |
if p.exists(): | |
with open(fname) as f: | |
return p.suffix, len(f.readlines()) | |
else: | |
return p.suffix, 0 | |
def _recursive_python_count(fname, cache): | |
if fname in cache: | |
return 0, 0 | |
else: | |
cache.add(fname) | |
line_count, file_count = 0, 0 | |
with open(fname) as f: | |
lines = f.readlines() | |
line_count += len(lines) | |
file_count += 1 | |
for line in lines: | |
if "import" in line and ("espnet." in line or "espnet2." in line or "espnetez." in line): | |
line = line.strip() | |
if line[:4] == "from": | |
pkg = line.split()[1] | |
lc, fc = _recursive_python_count(pkgutil.get_loader(pkg).get_filename(), cache) | |
line_count += lc | |
file_count += fc | |
return line_count, file_count | |
def _recursive_file_count(dirname, cache): | |
dirname = Path(dirname) | |
assert dirname.is_dir() | |
for f in dirname.rglob("*"): | |
f = Path(f).resolve() | |
if f.is_dir(): | |
for t, c in _recursive_file_count(f, cache): | |
yield t, c | |
else: | |
if str(f) not in cache: | |
cache.add(str(f)) | |
yield _size_count(str(f)) | |
def count_files(files, recursive=False): | |
line_counts = defaultdict(int) | |
file_counts = defaultdict(int) | |
cache = set() | |
for f in files: | |
p = Path(f).resolve() | |
if p.is_dir(): | |
lc, fc = count_files(list(p.rglob("*")), recursive=True) | |
for k, v in lc.items(): | |
line_counts[k] += v | |
for k, v in fc.items(): | |
file_counts[k] += v | |
else: | |
ty, cnt = _size_count(f) | |
if recursive and ty == ".py": | |
lc, fc = _recursive_python_count(f, cache) | |
line_counts[".py"] += lc | |
file_counts[".py"] += fc | |
else: | |
line_counts[ty] += cnt | |
file_counts[ty] += 1 | |
return line_counts, file_counts | |
def plot_bar(ez, esp, ylabel, title, fname): | |
mpl.rcParams["axes.prop_cycle"] = cycler(color=["#648fff", "#785ef0", "#dc267f", "#fe6100", "#FFB000", "#648fff", "#dc267f", "#000000"]) | |
ext = {".py": "Python", ".sh": "Bash", ".yaml": "Config", ".pl": "Perl", } | |
ez[".yaml"] += ez[".conf"] | |
esp[".yaml"] += esp[".conf"] | |
keys = (set(ez.keys()) | set(esp.keys())) - {".md", ".conf"} | |
both = {k: ez[k] + esp[k] for k in keys} | |
both = sorted(both.items(), key=lambda x: -x[1]) | |
bottom = [0, 0] | |
fig, ax = plt.subplots(figsize=(2., 3.5)) | |
for k, _ in both: | |
bar = ax.bar(["ESPnet", "ESPnet-EZ"], [esp[k], ez[k]], width=0.4, label=ext[k], bottom=bottom) | |
bottom[0] += esp[k] | |
bottom[1] += ez[k] | |
if both[-1][0] == k: | |
if sum(bottom) > 10000: | |
labels = [f"{b//1000}k" for b in bottom] | |
else: | |
labels = bottom | |
ax.bar_label(bar, labels=labels) | |
handles, labels = ax.get_legend_handles_labels() | |
ax.legend(handles[::-1], labels[::-1], title="Language", loc="upper right", frameon=False) | |
ax.set_ylabel(ylabel) | |
ax.set_title(title) | |
ax.set_ylim(0, int(max(bottom) * 1.1)) | |
if sum(bottom) > 10000: | |
ax.set_yticks([0, 50000, 100000, 150000, 200000, 250000], labels="0,50k,100k,150k,200k,250k".split(",")) | |
fig.savefig(fname, bbox_inches="tight") | |
# Minimum number of files | |
# espnetez | |
ez_lines, _ = count_files([ | |
"egs2/librispeech_100/ez1/conf/owsm_finetune_base.yaml", | |
"egs2/librispeech_100/ez1/main.py", | |
# "egs2/librispeech_100/ez1/train.sh", | |
], False) | |
# espnet | |
orig_lines, _ = count_files([ | |
"egs2/librispeech_100/s2t1/conf/tuning/train_s2t_ebf_lr1e-3_warmup5k.yaml", | |
"egs2/librispeech_100/s2t1/conf/decode_s2t.yaml", | |
"egs2/librispeech_100/s2t1/local/data_prep.py", | |
"egs2/librispeech_100/s2t1/local/data.sh", | |
# "egs2/librispeech_100/s2t1/run.sh", | |
], False) | |
plot_bar(ez_lines, orig_lines, "# of lines", "Newly written code", "new_lines.pdf") | |
# Relevant files | |
ez_lines, ez_files = count_files([ | |
"egs2/librispeech_100/ez1/conf/owsm_finetune_base.yaml", | |
"egs2/librispeech_100/ez1/main.py", | |
# "egs2/librispeech_100/ez1/train.sh", | |
], True) | |
orig_lines, orig_files = count_files([ | |
"egs2/librispeech_100/s2t1/conf/tuning/train_s2t_ebf_lr1e-3_warmup5k.yaml", | |
"egs2/librispeech_100/s2t1/conf/decode_s2t.yaml", | |
"egs2/librispeech_100/s2t1/local/data_prep.py", | |
"egs2/librispeech_100/s2t1/local/data.sh", | |
# "egs2/librispeech_100/s2t1/run.sh", | |
"egs2/librispeech_100/s2t1", | |
"espnet2/bin/s2st_train.py", | |
"espnet2/bin/aggregate_stats_dirs.py", | |
"espnet2/bin/split_scps.py", | |
"espnet2/bin/launch.py", | |
], True) | |
plot_bar(ez_lines, orig_lines, "# of lines", "Dependent code", "all_lines.pdf") | |
plot_bar(ez_files, orig_files, "# of files", "Dependent code", "all_files.pdf") | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment