du alternative

tests-ng/pdu.py

@@ -0,0 +1,88 @@
+#!/usr/bin/env python3
+# author: deadc0de6 (https://github.com/deadc0de6)
+# Copyright (c) 2024, deadc0de6
+#
+# a naive python du alternative
+# ./tests-ng/pdu.py . --ignore='*/.git*' -H
+#
+
+
+import os
+import argparse
+import fnmatch
+from typing import List
+
+
+NAME = 'pdu'
+DEBUG = True
+
+
+def debug(txt: str):
+    """debug output"""
+    if not DEBUG:
+        return
+    print(f'[DEBUG] {txt}')
+
+
+def size_to_str(size: float, human: bool = False) -> str:
+    """size to string"""
+    div = 1024.
+    suf = ['B', 'K', 'M', 'G', 'T', 'P']
+    if not human or size < div:
+        return f'{size}'
+    for i in suf:
+        if size < div:
+            return f'{size:.1f}{i}'
+        size = size / div
+    sufix = suf[-1]
+    return f'{size:.1f}{sufix}'
+
+
+def must_ignore(path: str, patterns: List[str]) -> bool:
+    """must ignore path"""
+    if not patterns:
+        return False
+    lst = [fnmatch.fnmatch(path, patt) for patt in patterns]
+    debug(f'{path} -> {lst}')
+    return any(lst)
+
+
+def main(path: str, human: bool,
+         ign: List[str] = None):
+    """entry point"""
+    if not os.path.exists(path):
+        print(f'[ERROR] {path} does not exist')
+        return False
+    total = 0
+    for root, _, files in os.walk(path, topdown=True):
+        dirsz = 0
+        if must_ignore(root, ign):
+            debug(f'ignore root {root}')
+            continue
+        for file in files:
+            fpath = os.path.join(root, file)
+            if must_ignore(fpath, ign):
+                debug(f'ignore sub {fpath}')
+                continue
+            size = os.path.getsize(fpath)
+            dirsz += size
+            total += size
+        if root != path:
+            print(f'{size_to_str(dirsz, human=human)} {root}')
+    print(f'{size_to_str(total, human=human)} {path}')
+    return True
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(prog=NAME,
+                                     description='python du')
+    parser.add_argument('path')
+    parser.add_argument('-H', '--human',
+                        action='store_true')
+    parser.add_argument('-i', '--ignore',
+                        nargs='+')
+    parser.add_argument('-d', '--debug',
+                        action='store_true')
+    args = parser.parse_args()
+    DEBUG = args.debug
+    main(args.path, args.human, ign=args.ignore)

tests-ng/test-du.sh

@@ -37,14 +37,16 @@ cnt=$(wc -l "${out}" | awk '{print $1}')
 
 # bin size
 echo ">>> test du bin size raw <<<"
-expected=$(du -c --block=1 --apparent-size "${cur}/../cmd/gocatcli" | tail -1 | awk '{print $1}')
+#expected=$(du -c --block=1 --apparent-size "${cur}/../cmd/gocatcli" | tail -1 | awk '{print $1}')
+expected=$("${cur}/pdu.py" "${cur}/../cmd/gocatcli" | tail -1 | awk '{print $1}')
 size=$(grep '^.* *gocatcli/cmd/gocatcli$' "${out}" | awk '{print $1}')
 echo "size:${size} VS exp:${expected}"
 [ "${expected}" != "${size}" ] && (echo "bad bin size" && exit 1)
 
 # total size
 echo ">>> test du total size raw <<<"
-expected=$(du -c --block=1 --apparent-size "${cur}/../" | tail -1 | awk '{print $1}')
+#expected=$(du -c --block=1 --apparent-size "${cur}/../" | tail -1 | awk '{print $1}')
+expected=$("${cur}/pdu.py" "${cur}/../" | tail -1 | awk '{print $1}')
 #cat_file "${out}"
 size=$(tail -1 "${out}" | awk '{print $1}')
 echo "size:${size} VS exp:${expected}"

tests-ng/test-size.sh

@@ -30,8 +30,9 @@ out="${tmpd}/output.txt"
 echo ">>> test file size <<<"
 "${bin}" --debug ls -r -S -c "${catalog}" "${cur}/../internal/walker/walker.go" | sed -e 's/\x1b\[[0-9;]*m//g' > "${out}"
 # shellcheck disable=SC2126
-du --block=1 "${cur}/../internal/walker/walker.go"
-expected=$(du --block=1 --apparent-size "${cur}/../internal/walker/walker.go" | awk '{print $1}')
+#expected=$(du --block=1 --apparent-size "${cur}/../internal/walker/walker.go" | awk '{print $1}')
+# shellcheck disable=SC2012
+expected=$(ls -la "${cur}/../internal/walker/walker.go" | awk '{print $5}')
 cat_file "${out}"
 size=$(grep 'walker.go' "${out}" | awk '{print $4}')
 echo "size:${size} VS exp:${expected}"
@@ -40,16 +41,19 @@ echo "size:${size} VS exp:${expected}"
 echo ">>> test directory size <<<"
 "${bin}" --debug ls -r -S -c "${catalog}" | sed -e 's/\x1b\[[0-9;]*m//g' > "${out}"
 # shellcheck disable=SC2126
-expected=$(du -c --block=1 --apparent-size "${cur}/../internal/catcli" | tail -1 | awk '{print $1}')
-cat_file "${out}"
+#expected=$(du -c --block=1 --apparent-size "${cur}/../internal/catcli" | tail -1 | awk '{print $1}')
+"${cur}/pdu.py" "${cur}/../internal/catcli" | tail -1
+expected=$("${cur}/pdu.py" "${cur}/../internal/catcli" | tail -1 | awk '{print $1}')
+#cat_file "${out}"
 size=$(grep '^catcli' "${out}" | awk '{print $4}')
 echo "size:${size} VS exp:${expected}"
 [ "${size}" != "${expected}" ] && echo "expecting ${expected} (got ${size})" && exit 1
 
 echo ">>> test storage size <<<"
 "${bin}" --debug ls -S -c "${catalog}" | sed -e 's/\x1b\[[0-9;]*m//g' > "${out}"
 cat_file "${out}"
-expected=$(du -c --block=1 --apparent-size "${cur}/../internal" | tail -1 | awk '{print $1}')
+#expected=$(du -c --block=1 --apparent-size "${cur}/../internal" | tail -1 | awk '{print $1}')
+expected=$("${cur}/pdu.py" "${cur}/../internal" | tail -1 | awk '{print $1}')
 size=$(grep '^storage' "${out}" | awk '{print $3}')
 echo "size:${size} VS exp:${expected}"
 [ "${size}" != "${expected}" ] && echo "expecting ${expected} (got ${size})" && exit 1