support canon cr3 raws

photoapp/image.py

@@ -127,15 +127,51 @@ def hms_to_decimal(values):
     return values[0] + values[1] / 60 + values[2] / 3600
 
 
+def is_cr3(header):
+    # detect cannon's new cr3 format
+    # here's what we look for, expecting the first 100 bytes from the file:
+    # example cr3 file: IMG_0956.CR3
+    # 00000000  00 00 00 18 66 74 79 70  63 72 78 20 00 00 00 01  |....ftypcrx ....|    <--- the "ftypcrx "
+    # 00000010  63 72 78 20 69 73 6f 6d  00 00 71 a0 6d 6f 6f 76  |crx isom..q.moov|
+    # 00000020  00 00 68 c0 75 75 69 64  85 c0 b6 87 82 0f 11 e0  |..h.uuid........|
+    # 00000030  81 11 f4 ce 46 2b 6a 48  00 00 00 26 43 4e 43 56  |....F+jH...&CNCV|
+    # 00000040  43 61 6e 6f 6e 43 52 33  5f 30 30 31 2f 30 30 2e  |CanonCR3_001/00.|    <--- the "CanonCR3"
+    # 00000050  31 31 2e 30 30 2f 30 30  2e 30 30 2e 30 30 00 00  |11.00/00.00.00..|
+    # 00000060  00 5c 43 43                                       |.\CC|
+    # 00000064
+    tag1 = header[0:13]
+    tag2 = header[64:72]
+    return tag1 == b'\x00\x00\x00\x18ftypcrx \x00' and tag2 == b'CanonCR3'
+
+
+def is_file_cr3(fpath):
+    with open(fpath, "rb") as f:
+        header = f.read(100)
+    return is_cr3(header)
+
+
+def is_file_xmp(fname):
+    return fname.split(".")[-1].lower() == "xmp"
+
+
 def special_magic(fpath):
-    if fpath.split(".")[-1].lower() == "xmp":
-        return "application/octet-stream-xmp"
-    else:
-        return magic.from_file(fpath, mime=True)
+    """
+    magic but support for some new formats magic doesn't know about
+    """
+    if is_file_xmp(fpath):  # this one is cheap and uses the file name
+        return "application/octet-stream-xmp"  # duplicate check here avoids a file open() which i guess is nice
+    else:  # otherwise, we fall back to reading the file
+        with open(fpath, "rb") as f:
+            return special_magic_fobj(f, fpath)
 
 
 def special_magic_fobj(fobj, fname):
-    if fname.split(".")[-1].lower() == "xmp":
+    if is_file_xmp(fname):  # supported here too ...
         return "application/octet-stream-xmp"
     else:
-        return magic.from_buffer(fobj.read(1024), mime=True)
+        header = fobj.read(1024)
+        if is_cr3(header):
+            return "image/x-canon-cr3"
+        else:
+            ret = magic.from_buffer(header, mime=True)
+            return ret

photoapp/migrate.py

@@ -172,7 +172,7 @@ def migrate_gpsfix():
             continue
 
         # just bail if there's a CR2, the paired jpg is known not to have gps data in my dataset :)
-        if any(["image/x-canon-cr2" == i.format for i in p.files]):
+        if any([i.format in ("image/x-canon-cr2", "image/x-canon-cr3", ) for i in p.files]):
             continue
 
         # pick the jpg out of the set

photoapp/types.py

@@ -27,6 +27,8 @@ ftypes = dict(jpg=dict(category=fcategory.image,
                        mimes={"image/gif"}),
               cr2=dict(category=fcategory.raw,
                        mimes={"image/x-canon-cr2"}),
+              cr3=dict(category=fcategory.raw,
+                       mimes={"image/x-canon-cr3"}),
               xmp=dict(category=fcategory.raw,
                        mimes={"application/octet-stream-xmp"}),
               psd=dict(category=fcategory.raw,
@@ -66,7 +68,7 @@ known_mimes = set.union(*[i["mimes"] for i in ftypes.values()])
 regular_images = set([extension for extension, ftype in ftypes.items() if ftype["category"] == fcategory.image])
 regular_mimes = set().union(*[ftype["mimes"] for ftype in ftypes.values() if ftype["category"] == fcategory.image])
 # "derived" files, treated as black boxes, we can't open them because proprietary
-# cr2, xmp, etc
+# cr2, cr3, xmp, etc
 files_raw = set([extension for extension, ftype in ftypes.items() if ftype["category"] == fcategory.raw])
 # video types
 # mp4, mov, etc