fix: improve detection of mimetype

backup_to_cloud/main.py

@@ -8,7 +8,7 @@
 
 from .settings import EntryType, get_settings
 from .upload import backup
-from .utils import list_files, log
+from .utils import get_mimetype, list_files, log, ZIP_MIMETYPE
 
 
 def _main():
@@ -22,7 +22,8 @@ def _main():
             if not entry.zip:
                 files = list_files(entry.root_path, entry.filter)
                 for file in files:
-                    backup(file, entry.folder)
+                    mimetype = get_mimetype(file)
+                    backup(file, mimetype, entry.folder)
                 continue
 
             buffer = BytesIO()
@@ -47,10 +48,11 @@ def _main():
                 for file in files:
                     myzip.write(file, Path(file).relative_to(min_file))
 
-            backup(buffer, entry.folder, filename=entry.zipname)
+            backup(buffer, ZIP_MIMETYPE, entry.folder, filename=entry.zipname)
 
         elif entry.type == EntryType.single_file:
-            backup(entry.root_path, entry.folder)
+            mimetype = get_mimetype(entry.root_path)
+            backup(entry.root_path,mimetype, entry.folder)
 
 
 def main():

backup_to_cloud/upload.py

@@ -10,7 +10,7 @@
 FileData = Union[Path, str, BytesIO]
 
 
-def backup(file_data: FileData, folder=None, filename=None):
+def backup(file_data: FileData, mimetype, folder=None, filename=None):
     if isinstance(file_data, (str, Path)):
         filepath = Path(file_data)
         if not filepath.exists():
@@ -21,7 +21,6 @@ def backup(file_data: FileData, folder=None, filename=None):
         del filepath
 
     service = get_google_drive_services()
-    mimetype = mimetypes.guess_type(filename)[0]
     file_metadata = {"name": filename, "mimeType": mimetype}
     query = "name = %r" % (filename)
 
@@ -36,13 +35,12 @@ def backup(file_data: FileData, folder=None, filename=None):
         raise RuntimeError("Files should not be more than one")
 
     if ids:
-        return save_version(service, file_data, ids[0], filename)
-    return save_new_file(service, file_data, folder, filename)
+        return save_version(service, file_data, mimetype, ids[0], filename)
+    return save_new_file(service, file_data, mimetype, folder, filename)
 
 
-def save_new_file(gds, file_data: BytesIO, folder=None, filename=None):
+def save_new_file(gds, file_data: BytesIO, mimetype, folder=None, filename=None):
     log("saving new file: %s", filename)
-    mimetype = get_mimetype(filename)
     file_metadata = {"name": filename, "mimeType": mimetype}
 
     if folder:
@@ -55,13 +53,12 @@ def save_new_file(gds, file_data: BytesIO, folder=None, filename=None):
     return res
 
 
-def save_version(gds, file_data: BytesIO, file_id: str, filename=None):
+def save_version(gds, file_data: BytesIO, mimetype, file_id: str, filename=None):
     log("saving new version of %s", filename)
     file_metadata = {
         "name": filename,
         "published": True,
     }
-    mimetype = get_mimetype(filename)
     media = MediaIoBaseUpload(file_data, mimetype=mimetype)
     response = (
         gds.files()

backup_to_cloud/utils.py

@@ -12,6 +12,7 @@
 from .paths import CREDENTIALS_PATH, LOG_PATH, TOKEN_PATH
 
 SCOPES = ["https://www.googleapis.com/auth/drive"]
+ZIP_MIMETYPE = "application/octet-stream"
 
 
 class TokenError(Exception):
@@ -63,8 +64,21 @@ def get_creds_from_token():
     return creds
 
 
-def get_mimetype(filename: str) -> str:
-    return mimetypes.guess_type(filename)[0] or "application/octet-stream"
+def get_mimetype(filepath: str) -> str:
+    mimetype = mimetypes.guess_type(filepath)[0]
+    if mimetype:
+        return mimetype
+
+    data = Path(filepath).read_bytes()
+    try:
+        data.decode()
+    except UnicodeDecodeError:
+        try:
+            data.decode("utf-8")
+        except UnicodeDecodeError:
+            return "application/octet-stream"
+
+    return "text/plain"
 
 
 def list_files(root_path, regex_filter):