Make streaming compatible

whisper_adversarial_examples.py

@@ -28,7 +28,7 @@ Adversarial examples fooling whisper models
 """
 
 
-_DL_URLS = {
+_DL_URLS0 = {
     "targeted": {
         "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/75f06ad3-4f86-4f4b-b748-ea0e94f23379?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
     },
@@ -62,6 +62,40 @@ _DL_URLS = {
 }
 
 
+_DL_URLS = {
+    "targeted": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/ca70a5b6-ec76-4e6f-8a67-982e411da265?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "untargeted-35": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/516787a5-4832-4432-9138-9f01cccc4875?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "untargeted-40": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/ed7127c6-9769-4db5-ab5a-98e9ce15a6ae?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "language-armenian": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/57a8301c-a3de-4f34-a321-6cbdec5b7d55?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "language-lithuanian": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/b8dc1e63-d308-45e8-b16c-98ca4ac3e939?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "language-czech": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/8e5246e6-dfad-4d4c-aa1e-091cf24d975c?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "language-danish": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/15a27ffe-8ad3-4a92-adfc-ac1c6a7b230b?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "language-indonesian": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/ad3366b1-21a4-4ad4-9755-8a1d3775db62?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "language-italian": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/1729f188-ae9f-4a29-a8da-9597c1f2d0cc?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+    "language-english": {
+        "all": "https://data.mendeley.com/api/datasets/96dh52hz9r/draft/files/7d09cf90-af7d-4d33-914a-3002ea956a53?a=ee30841f-1832-41ec-bdac-bf3e5b67073c"
+    },
+}
+
+
 class AdvWhisperASRConfig(datasets.BuilderConfig):
     """BuilderConfig for AdvWhisperASR."""
 
@@ -158,7 +192,7 @@ class AdvWhisperASR(datasets.GeneratorBasedBuilder):
                     name=lang+"."+target[0],
                     gen_kwargs={
                         "local_extracted_archive": local_extracted_archive.get("all"),
-                        "files": dl_manager.iter_files(local_extracted_archive.get("all")),
+                        "files": dl_manager.iter_archive(archive_path["all"]),
                         "path_audio": os.path.join(folders[self.config.name]+"-"+target[1],"whisper-medium",str(seeds[self.config.name]),"save")
                     },
                 ) for target in targets
@@ -167,7 +201,7 @@ class AdvWhisperASR(datasets.GeneratorBasedBuilder):
                     name="original",
                     gen_kwargs={
                         "local_extracted_archive": local_extracted_archive.get("all"),
-                        "files": dl_manager.iter_files(local_extracted_archive.get("all")),
+                        "files": dl_manager.iter_archive(archive_path["all"]),
                         "path_audio": folders[self.config.name]+"-original"
                     },
                 )
@@ -178,7 +212,7 @@ class AdvWhisperASR(datasets.GeneratorBasedBuilder):
                     name=model.replace("-","."),
                     gen_kwargs={
                         "local_extracted_archive": local_extracted_archive.get("all"),
-                        "files": dl_manager.iter_files(local_extracted_archive.get("all")),
+                        "files": dl_manager.iter_archive(archive_path["all"]),
                         "path_audio": os.path.join(folders[self.config.name],model,str(seeds[self.config.name]),"save")
                     },
                 ) for model in models
@@ -187,7 +221,7 @@ class AdvWhisperASR(datasets.GeneratorBasedBuilder):
                     name="original",
                     gen_kwargs={
                         "local_extracted_archive": local_extracted_archive.get("all"),
-                        "files": dl_manager.iter_files(local_extracted_archive.get("all")),
+                        "files": dl_manager.iter_archive(archive_path["all"]),
                         "path_audio": os.path.join(folders[self.config.name],"original")
                     },
                 )
@@ -198,34 +232,39 @@ class AdvWhisperASR(datasets.GeneratorBasedBuilder):
     def _generate_examples(self, files, local_extracted_archive,path_audio):
         """Generate examples from an extracted path."""
         key = 0
+        suffix = "_nat.wav" if "original" in path_audio else "_adv.wav"
         audio_data = {}
         transcripts = []
-        for path in files:
-            if path.endswith(".csv"):
-                with open(path,'r') as f:
-                    for line in f:
-                        if line:
-                            line = (line.decode("utf-8") if isinstance(line,bytes) else line)
-                            line=line.strip().split(",")
-                            id_ = line[0]
-                            transcript=line[-1]
-                            transcript = transcript[:-1] if transcript[-1]=='\n' else transcript
-                            suffix = "_nat.wav" if "original" in path_audio else "_adv.wav"
-                            audio_file = id_+suffix
-                            audio_file = os.path.join(local_extracted_archive,path_audio, audio_file)
-                            if os.path.exists(audio_file):
-                                with open(audio_file,"rb") as f:
-                                    audio_data[id_] = f.read()
-                                transcripts.append(
-                                    {
-                                        "id": id_,
-                                        "file": audio_file,
-                                        "text": transcript,
-                                    }
-                                )
+        for t in files:
+            path, f = t
+            if path.endswith(".wav"):
+                if path_audio in path and path.endswith(suffix):
+                    id_ = path.split("/")[-1][: -len(suffix)]
+                    audio_data[id_] = f.read()
+            elif path.endswith(".csv"):
+                for line in f:
+                    if line:
+                        line = (line.decode("utf-8") if isinstance(line,bytes) else line)
+                        line=line.strip().split(",")
+                        id_ = line[0]
+                        transcript=line[-1]
+                        transcript = transcript[:-1] if transcript[-1]=='\n' else transcript
+                        audio_file = id_+suffix
+                        audio_file = (
+                            os.path.join(local_extracted_archive,path_audio, audio_file)
+                            if local_extracted_archive else audio_file
+                        )
+                        transcripts.append(
+                            {
+                                "id": id_,
+                                "file": audio_file,
+                                "text": transcript,
+                            }
+                        )
         for transcript in transcripts:
-            audio = {"path": transcript["file"], "bytes": audio_data[transcript["id"]]}
-            yield key, {"audio": audio, **transcript}
-            key += 1
+            if transcript["id"] in audio_data:
+                audio = {"path": transcript["file"], "bytes": audio_data[transcript["id"]]}
+                yield key, {"audio": audio, **transcript}
+                key += 1
         audio_data = {}
         transcripts = []