Cache temp files created from base64 data (#3197)

* changes

* added workflow

* fix action

* fix action

* fix action

* changelg

* formatting

* fix

* Delete benchmark-queue.yml

* Delete benchmark_queue.py

* changelog

* lint

* fix tests

* fix tests

* fix for python 3.7

* formatting

CHANGELOG.md

@@ -55,6 +55,7 @@ No changes to highlight.
 
 ## Full Changelog:
 * Fix demos page css and add close demos button by [@aliabd](https://github.com/aliabd) in [PR 3151](https://github.com/gradio-app/gradio/pull/3151)
+* Caches temp files from base64 input data by giving them a deterministic path based on the contents of data by [@abidlabs](https://github.com/abidlabs) in [PR 3197](https://github.com/gradio-app/gradio/pull/3197) 
 * Better warnings (when there is a mismatch between the number of output components and values returned by a function, or when the `File` component or `UploadButton` component includes a `file_types` parameter along with `file_count=="dir"`) by [@abidlabs](https://github.com/abidlabs) in [PR 3194](https://github.com/gradio-app/gradio/pull/3194) 
 * Raises a `gr.Error` instead of a regular Python error when you use `gr.Interface.load()` to load a model and there's an error querying the HF API by [@abidlabs](https://github.com/abidlabs) in [PR 3194](https://github.com/gradio-app/gradio/pull/3194)
 

gradio/components.py

@@ -1808,13 +1808,9 @@ class Video(
             x.get("is_file", False),
         )
         if is_file:
-            file = self.make_temp_copy_if_needed(file_name)
+            file_name = Path(self.make_temp_copy_if_needed(file_name))
-            file_name = Path(file)
         else:
-            file = processing_utils.decode_base64_to_file(
+            file_name = Path(self.base64_to_temp_file_if_needed(file_data, file_name))
-                file_data, file_path=file_name
-            )
-            file_name = Path(file.name)
 
         uploaded_format = file_name.suffix.replace(".", "")
         modify_format = self.format is not None and uploaded_format != self.format
@@ -2041,20 +2037,26 @@ class Audio(
             else:
                 temp_file_path = self.make_temp_copy_if_needed(file_name)
         else:
-            temp_file_obj = processing_utils.decode_base64_to_file(
+            temp_file_path = self.base64_to_temp_file_if_needed(file_data, file_name)
-                file_data, file_path=file_name
-            )
-            temp_file_path = temp_file_obj.name
 
         sample_rate, data = processing_utils.audio_from_file(
             temp_file_path, crop_min=crop_min, crop_max=crop_max
         )
 
+        # Need a unique name for the file to avoid re-using the same audio file if
+        # a user submits the same audio file twice, but with different crop min/max.
+        temp_file_path = Path(temp_file_path)
+        output_file_name = str(
+            temp_file_path.with_name(
+                f"{temp_file_path.stem}-{crop_min}-{crop_max}{temp_file_path.suffix}"
+            )
+        )
+
         if self.type == "numpy":
             return sample_rate, data
         elif self.type == "filepath":
-            processing_utils.audio_to_file(sample_rate, data, temp_file_path)
+            processing_utils.audio_to_file(sample_rate, data, output_file_name)
-            return temp_file_path
+            return output_file_name
         else:
             raise ValueError(
                 "Unknown type: "
@@ -2075,8 +2077,8 @@ class Audio(
         if x.get("is_file"):
             sample_rate, data = processing_utils.audio_from_file(x["name"])
         else:
-            file_obj = processing_utils.decode_base64_to_file(x["data"])
+            file_name = self.base64_to_temp_file_if_needed(x["data"])
-            sample_rate, data = processing_utils.audio_from_file(file_obj.name)
+            sample_rate, data = processing_utils.audio_from_file(file_name)
         leave_one_out_sets = []
         tokens = []
         masks = []
@@ -2117,14 +2119,14 @@ class Audio(
     def get_masked_inputs(self, tokens, binary_mask_matrix):
         # create a "zero input" vector and get sample rate
         x = tokens[0]["data"]
-        file_obj = processing_utils.decode_base64_to_file(x)
+        file_name = self.base64_to_temp_file_if_needed(x)
-        sample_rate, data = processing_utils.audio_from_file(file_obj.name)
+        sample_rate, data = processing_utils.audio_from_file(file_name)
         zero_input = np.zeros_like(data, dtype="int16")
         # decode all of the tokens
         token_data = []
         for token in tokens:
-            file_obj = processing_utils.decode_base64_to_file(token["data"])
+            file_name = self.base64_to_temp_file_if_needed(token["data"])
-            _, data = processing_utils.audio_from_file(file_obj.name)
+            _, data = processing_utils.audio_from_file(file_name)
             token_data.append(data)
         # construct the masked version
         masked_inputs = []
@@ -4046,10 +4048,7 @@ class Model3D(
         if is_file:
             temp_file_path = self.make_temp_copy_if_needed(file_name)
         else:
-            temp_file = processing_utils.decode_base64_to_file(
+            temp_file_path = self.base64_to_temp_file_if_needed(file_data, file_name)
-                file_data, file_path=file_name
-            )
-            temp_file_path = temp_file.name
 
         return temp_file_path
 

gradio/processing_utils.py

@@ -353,6 +353,13 @@ class TempFileManager:
             sha1.update(data)
         return sha1.hexdigest()
 
+    def hash_base64(self, base64_encoding: str, chunk_num_blocks: int = 128) -> str:
+        sha1 = hashlib.sha1()
+        for i in range(0, len(base64_encoding), chunk_num_blocks * sha1.block_size):
+            data = base64_encoding[i : i + chunk_num_blocks * sha1.block_size]
+            sha1.update(data.encode("utf-8"))
+        return sha1.hexdigest()
+
     def get_prefix_and_extension(self, file_path_or_url: str) -> Tuple[str, str]:
         file_name = Path(file_path_or_url).name
         prefix, extension = file_name, None
@@ -374,6 +381,12 @@ class TempFileManager:
         file_hash = self.hash_url(url)
         return prefix + file_hash + extension
 
+    def get_temp_base64_path(self, base64_encoding: str, prefix: str) -> str:
+        extension = get_extension(base64_encoding)
+        extension = "." + extension if extension else ""
+        base64_hash = self.hash_base64(base64_encoding)
+        return prefix + base64_hash + extension
+
     def make_temp_copy_if_needed(self, file_path: str) -> str:
         """Returns a temporary file path for a copy of the given file path if it does
         not already exist. Otherwise returns the path to the existing temp file."""
@@ -408,6 +421,27 @@ class TempFileManager:
         self.temp_files.add(full_temp_file_path)
         return full_temp_file_path
 
+    def base64_to_temp_file_if_needed(
+        self, base64_encoding: str, file_name: str | None = None
+    ) -> str:
+        """Converts a base64 encoding to a file and returns the path to the file if
+        the file doesn't already exist. Otherwise returns the path to the existing file."""
+        f = tempfile.NamedTemporaryFile(delete=False)
+        temp_dir = Path(f.name).parent
+        prefix = self.get_prefix_and_extension(file_name)[0] if file_name else ""
+
+        temp_file_path = self.get_temp_base64_path(base64_encoding, prefix=prefix)
+        f.name = str(temp_dir / temp_file_path)
+        full_temp_file_path = str(utils.abspath(f.name))
+
+        if not Path(full_temp_file_path).exists():
+            data, _ = decode_base64_to_binary(base64_encoding)
+            with open(full_temp_file_path, "wb") as fb:
+                fb.write(data)
+
+        self.temp_files.add(full_temp_file_path)
+        return full_temp_file_path
+
 
 def download_tmp_copy_of_file(
     url_path: str, access_token: str | None = None, dir: str | None = None

test/test_components.py

@@ -770,9 +770,9 @@ class TestAudio:
         """
         x_wav = deepcopy(media_data.BASE64_AUDIO)
         audio_input = gr.Audio()
-        output = audio_input.preprocess(x_wav)
+        output1 = audio_input.preprocess(x_wav)
-        assert output[0] == 8000
+        assert output1[0] == 8000
-        assert output[1].shape == (8046,)
+        assert output1[1].shape == (8046,)
         assert filecmp.cmp(
             "test/test_files/audio_sample.wav",
             audio_input.serialize("test/test_files/audio_sample.wav")["name"],
@@ -796,7 +796,9 @@ class TestAudio:
         assert audio_input.preprocess(None) is None
         x_wav["is_example"] = True
         x_wav["crop_min"], x_wav["crop_max"] = 1, 4
-        assert audio_input.preprocess(x_wav) is not None
+        output2 = audio_input.preprocess(x_wav)
+        assert output2 is not None
+        assert output1 != output2
 
         audio_input = gr.Audio(type="filepath")
         assert isinstance(audio_input.preprocess(x_wav), str)

test/test_processing_utils.py

@@ -194,6 +194,30 @@ class TestTempFileManager:
         )
         assert len(temp_file_manager.temp_files) == 2
 
+    def test_base64_to_temp_file_if_needed(self):
+        temp_file_manager = processing_utils.TempFileManager()
+
+        base64_file_1 = media_data.BASE64_IMAGE
+        base64_file_2 = media_data.BASE64_AUDIO["data"]
+
+        f = temp_file_manager.base64_to_temp_file_if_needed(base64_file_1)
+        try:  # Delete if already exists from before this test
+            os.remove(f)
+        except OSError:
+            pass
+
+        f = temp_file_manager.base64_to_temp_file_if_needed(base64_file_1)
+        assert len(temp_file_manager.temp_files) == 1
+
+        f = temp_file_manager.base64_to_temp_file_if_needed(base64_file_1)
+        assert len(temp_file_manager.temp_files) == 1
+
+        f = temp_file_manager.base64_to_temp_file_if_needed(base64_file_2)
+        assert len(temp_file_manager.temp_files) == 2
+
+        for file in temp_file_manager.temp_files:
+            os.remove(file)
+
     @pytest.mark.flaky
     @patch("shutil.copyfileobj")
     def test_download_temp_copy_if_needed(self, mock_copy):