Add progress reporting when finding duplicates.

1 files changed, 21 insertions, 8 deletions

diff --git a/dupfiles b/dupfiles
index 35742f1..1aa9dc0 100755
--- a/dupfiles
+++ b/dupfiles
@@ -80,8 +80,12 @@ class DuplicateFileFinder(object):
         for size, tuples in self.by_size.iteritems():
             if len(set((dev, ino) for dev, ino, pathname in tuples)) == 1:
                 # All duplicates are hardlinks to the same inode. Skip.
-                continue
-            result += self.find_duplicates([p for d, i, p in tuples])
+                done_bytes += len(tuples) * size
+            else:
+                new_dups = self.find_duplicates([p for d, i, p in tuples])
+                result += new_dups
+                done_bytes += len(new_dups) * size
+            self.duplicates_progress(done_bytes, total_bytes, start_time)
         self.progress.finished()
         return result
 
@@ -146,14 +150,10 @@ class DuplicateFileFinder(object):
 
     def duplicates_progress(self, done, total, started):
         duration = time.time() - started
-        if duration < 1:
-            speed = 0
-        else:
-            speed = done / duration
-        self.progress.write('%s/%s (%.1f%%) done (%s/s)' %
+        self.progress.write('%s/%s (%.1f%%) done (%s)' %
                             (self.human_size(done), self.human_size(total), 
                              100.0 * float(done) / float(total),
-                             self.human_size(speed)))
+                             self.human_duration(duration)))
 
     def human_size(self, size):
         tab = [(1024**3, 'GiB'),
@@ -164,6 +164,19 @@ class DuplicateFileFinder(object):
                 return '%.1f %s' % (float(size) / float(limit), unit)
         return '0 B'
 
+    def human_duration(self, duration):
+        units = [(3600, 'h'),
+                 (60, 'min'),
+                 (1, 's')]
+
+        parts = []
+        for limit, unit in units:
+            count = int(duration) / limit
+            duration %= limit
+            if count > 0:
+                parts.append('%d %s' % (count, unit))
+        return ' '.join(parts or ['0 s'])
+
 
 def make_hardlinks(duplicates):
     canonical = duplicates.pop()