handle file name conflicts in media on .anki2 import

anki/collection.py

@@ -630,7 +630,6 @@ where c.nid == f.id
         ids = self.db.list("""
 select id from notes where mid not in """ + ids2str(self.models.ids()))
         if ids:
-            print self.db.list("select distinct mid from notes where id in " + ids2str(ids))
             problems.append(
                 ngettext("Deleted %d note with missing note type.",
                          "Deleted %d notes with missing note type.", len(ids))

anki/importing/anki2.py

@@ -2,8 +2,9 @@
 # Copyright: Damien Elmes <anki@ichi2.net>
 # License: GNU AGPL, version 3 or later; http://www.gnu.org/licenses/agpl.html
 
+import os
 from anki import Collection
-from anki.utils import intTime
+from anki.utils import intTime, splitFields, joinFields, checksum
 from anki.importing.base import Importer
 from anki.lang import _
 from anki.lang import ngettext
@@ -48,7 +49,7 @@ class Anki2Importer(Importer):
         self._prepareModels()
         self._importNotes()
         self._importCards()
-        self._importMedia()
+        self._importStaticMedia()
         self._postImport()
         self.dst.db.execute("vacuum")
         self.dst.db.execute("analyze")
@@ -86,6 +87,8 @@ class Anki2Importer(Importer):
                 # rewrite internal ids, models, etc
                 note[2] = lmid
                 note[4] = usn
+                # update media references in case of dupes
+                note[6] = self._mungeMedia(mid, note[6])
                 add.append(note)
                 dirty.append(note[0])
                 # note we have the added note
@@ -265,8 +268,60 @@ insert or ignore into revlog values (?,?,?,?,?,?,?,?,?)""", revlog)
     # Media
     ######################################################################
 
-    def _importMedia(self):
-        self.src.media.copyTo(self.dst.media.dir())
+    def _importStaticMedia(self):
+        # Import any '_foo' prefixed media files regardless of whether
+        # they're used on notes or not
+        for fname in os.listdir(self.src.media.dir()):
+            if fname.startswith("_") and not self.dst.media.have(fname):
+                self._writeDstMedia(fname, self._srcMediaData(fname))
+
+    def _mediaData(self, fname, dir=None):
+        if not dir:
+            dir = self.src.media.dir()
+        path = os.path.join(dir, fname)
+        try:
+            return open(path).read()
+        except IOError, OSError:
+            return
+
+    def _srcMediaData(self, fname):
+        "Data for FNAME in src collection."
+        return self._mediaData(fname, self.src.media.dir())
+
+    def _dstMediaData(self, fname):
+        "Data for FNAME in dst collection."
+        return self._mediaData(fname, self.dst.media.dir())
+
+    def _writeDstMedia(self, fname, data):
+        path = os.path.join(self.dst.media.dir(), fname)
+        open(path, "wb").write(data)
+
+    def _mungeMedia(self, mid, fields):
+        fields = splitFields(fields)
+        def repl(match):
+            fname = match.group(2)
+            srcData = self._srcMediaData(fname)
+            dstData = self._dstMediaData(fname)
+            if not srcData:
+                # file was not in source, ignore
+                return match.group(0)
+            # if model-local file exists from a previous import, use that
+            name, ext = os.path.splitext(fname)
+            lname = "%s_%s%s" % (name, mid, ext)
+            if self.dst.media.have(lname):
+                return match.group(0).replace(fname, lname)
+            # if missing or the same, pass unmodified
+            elif not dstData or srcData == dstData:
+                # need to copy?
+                if not dstData:
+                    self._writeDstMedia(fname, srcData)
+                return match.group(0)
+            # exists but does not match, so we need to dedupe
+            self._writeDstMedia(lname, srcData)
+            return match.group(0).replace(fname, lname)
+        for i in range(len(fields)):
+            fields[i] = self.dst.media.transformNames(fields[i], repl)
+        return joinFields(fields)
 
     # Post-import cleanup
     ######################################################################

anki/media.py

@@ -120,6 +120,11 @@ If the same name exists, compare checksums."""
                     l.append(fname)
         return l
 
+    def transformNames(self, txt, func):
+        for reg in self.regexps:
+            txt = re.sub(reg, func, txt)
+        return txt
+
     def strip(self, txt):
         for reg in self.regexps:
             txt = re.sub(reg, "", txt)
@@ -206,6 +211,9 @@ If the same name exists, compare checksums."""
             cnt += 1
         return cnt
 
+    def have(self, fname):
+        return os.path.exists(os.path.join(self.dir(), fname))
+
     # Media syncing - changes and removal
     ##########################################################################
 

tests/test_importing.py

@@ -8,7 +8,6 @@ from anki.errors import *
 from anki.importing import Anki1Importer, Anki2Importer, TextImporter, \
     SupermemoXmlImporter, MnemosyneImporter
 from anki.notes import Note
-
 from anki.db import *
 
 testDir = os.path.dirname(__file__)
@@ -27,7 +26,7 @@ def test_anki2():
     srcCards = src.cardCount()
     srcRev = src.db.scalar("select count() from revlog")
     # add a media file for testing
-    open(os.path.join(src.media.dir(), "foo.jpg"), "w").write("foo")
+    open(os.path.join(src.media.dir(), "_foo.jpg"), "w").write("foo")
     src.close()
     # create a new empty deck
     dst = getEmptyDeck()
@@ -53,6 +52,53 @@ def test_anki2():
     assert len(os.listdir(dst.media.dir())) == 1
     #print dst.path
 
+def test_anki2_mediadupes():
+    tmp = getEmptyDeck()
+    # add a note that references a sound
+    n = tmp.newNote()
+    n['Front'] = "[sound:foo.mp3]"
+    mid = n.model()['id']
+    tmp.addNote(n)
+    # add that sound to media folder
+    open(os.path.join(tmp.media.dir(), "foo.mp3"), "w").write("foo")
+    tmp.close()
+    # it should be imported correctly into an empty deck
+    empty = getEmptyDeck()
+    imp = Anki2Importer(empty, tmp.path)
+    imp.run()
+    assert os.listdir(empty.media.dir()) == ["foo.mp3"]
+    # and importing again will not duplicate, as the file content matches
+    empty.remCards(empty.db.list("select id from cards"))
+    imp = Anki2Importer(empty, tmp.path)
+    imp.run()
+    assert os.listdir(empty.media.dir()) == ["foo.mp3"]
+    n = empty.getNote(empty.db.scalar("select id from notes"))
+    assert "foo.mp3" in n.fields[0]
+    # if the local file content is different, and import should trigger a
+    # rename
+    empty.remCards(empty.db.list("select id from cards"))
+    open(os.path.join(empty.media.dir(), "foo.mp3"), "w").write("bar")
+    imp = Anki2Importer(empty, tmp.path)
+    imp.run()
+    assert sorted(os.listdir(empty.media.dir())) == [
+        "foo.mp3", "foo_%s.mp3" % mid]
+    n = empty.getNote(empty.db.scalar("select id from notes"))
+    assert "_" in n.fields[0]
+    # if the localized media file already exists, we rewrite the note and
+    # media
+    empty.remCards(empty.db.list("select id from cards"))
+    open(os.path.join(empty.media.dir(), "foo.mp3"), "w").write("bar")
+    imp = Anki2Importer(empty, tmp.path)
+    imp.run()
+    assert sorted(os.listdir(empty.media.dir())) == [
+        "foo.mp3", "foo_%s.mp3" % mid]
+    assert sorted(os.listdir(empty.media.dir())) == [
+        "foo.mp3", "foo_%s.mp3" % mid]
+    n = empty.getNote(empty.db.scalar("select id from notes"))
+    assert "_" in n.fields[0]
+
+    #print dst.path
+
 def test_anki1():
     # get the deck path to import
     tmp = getUpgradeDeckPath()
@@ -60,7 +106,7 @@ def test_anki1():
     mdir = tmp.replace(".anki2", ".media")
     if not os.path.exists(mdir):
         os.mkdir(mdir)
-    open(os.path.join(mdir, "foo.jpg"), "w").write("foo")
+    open(os.path.join(mdir, "_foo.jpg"), "w").write("foo")
     # create a new empty deck
     dst = getEmptyDeck()
     # import src into dst