another rool

reporting/file_usage_by_name.py

@@ -0,0 +1,69 @@
+#!/usr/bin/python
+import sys
+import os
+import pwd
+from argparse import ArgumentParser 
+
+MINSIZE=0
+
+def setup_options():
+    parser=ArgumentParser(description="List file usage")
+    parser.add_argument('startpath', action="append", nargs='+')
+
+    options=parser.parse_args()
+    return options
+
+def add_recurse(db,options,start):
+    for path,dirs,files in os.walk(start,followlinks=False):
+        files=clean_syms(files,path)
+        for f in files:
+            filename=os.path.join(path,f)
+            stats=os.stat(filename)
+            owner=stats.st_uid
+            size=stats.st_size
+            if not owner in db:
+                db[owner]=0
+            db[owner]+=size
+    return db
+
+def clean_syms(files,path):
+    nonsyms=[]
+    for f in files:
+        if not os.path.islink(os.path.join(path,f)):
+            nonsyms.append(f)
+    return nonsyms
+
+def print_db(db):
+    namedb=dict()
+    for u in db:
+        try:
+            namedb[pwd.getpwuid(u).pw_name]=humanize_size(db[u])
+        except:
+            namedb[str(u)]=humanize_size(db[u])
+    names=namedb.keys()
+    names.sort()
+    for n in names:   
+        print(n+": "+namedb[n])
+
+def humanize_size(size,precision=1):
+    if size==None:
+        return 'nan'
+    suffixes=['B','KB','MB','GB','TB']
+    suffixIndex = 0
+    defPrecision=0
+    while size > 1024:
+        suffixIndex += 1 #increment the index of the suffix
+        size = size/1024.0 #apply the division
+        defPrecision=precision
+    return "%.*f%s"%(defPrecision,size,suffixes[suffixIndex])
+
+def main():
+    options=setup_options();
+    db=dict()
+    for f in options.startpath[0]:
+        db=add_recurse(db,options,f)
+    print_db(db)
+    sys.exit(0)
+
+main()
+