Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -6,18 +6,34 @@ import azure.cognitiveservices.speech as speechsdk
|
|
6 |
dialects = {"Palestinian/Jordanian": "P", "Syrian": "S", "Lebanese": "L", "Egyptian": "E"}
|
7 |
|
8 |
translator_en2ar = pipeline(task="translation", model="guymorlan/English2Dialect")
|
|
|
9 |
transliterator = pipeline(task="translation", model="guymorlan/DialectTransliterator")
|
10 |
|
11 |
speech_config = speechsdk.SpeechConfig(subscription=os.environ.get('SPEECH_KEY'), region=os.environ.get('SPEECH_REGION'))
|
12 |
|
13 |
-
def translate_english(input_text):
|
14 |
if not input_text:
|
15 |
return "", "", "", ""
|
16 |
|
17 |
inputs = [f"{val} {input_text}" for val in dialects.values()]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
18 |
result = translator_en2ar(inputs)
|
19 |
|
20 |
-
|
|
|
|
|
|
|
|
|
|
|
21 |
|
22 |
def translate_arabic(input_text):
|
23 |
if not input_text:
|
@@ -34,10 +50,9 @@ def get_audio(input_text):
|
|
34 |
speech_synthesis_result = speech_synthesizer.speak_text_async(input_text).get()
|
35 |
return f"{input_text}.wav"
|
36 |
|
37 |
-
def get_transliteration(input_text):
|
38 |
-
if not
|
39 |
return ""
|
40 |
-
|
41 |
result = transliterator([input_text])
|
42 |
return result[0]["translation_text"]
|
43 |
|
@@ -55,6 +70,10 @@ with gr.Blocks(title = "English to Levantine Arabic", css=css, theme="default")
|
|
55 |
input_text = gr.Textbox(label="Input", placeholder="Enter English text", lines=1)
|
56 |
gr.Examples(["I wanted to go to the store yesterday, but it rained", "How are you feeling today?", "Let's go to your place"], input_text)
|
57 |
btn = gr.Button("Translate", label="Translate")
|
|
|
|
|
|
|
|
|
58 |
gr.Markdown("Built by [Guy Mor-Lan](mailto:guy.mor@mail.huji.ac.il). Pronunciation model is specifically tailored to urban Palestinian Arabic. Text-to-speech uses Microsoft Azure's API and may provide different result from the transliterated pronunciation.")
|
59 |
|
60 |
with gr.Column():
|
@@ -67,9 +86,19 @@ with gr.Blocks(title = "English to Levantine Arabic", css=css, theme="default")
|
|
67 |
audio = gr.Audio(label="Audio - Palestinian", interactive=False)
|
68 |
audio_button = gr.Button("Get Audio", label="Click Here to Get Audio")
|
69 |
audio_button.click(get_audio, inputs=[pal], outputs=[audio])
|
70 |
-
btn.click(translate_english,inputs=[input_text], outputs=[pal, sy, lb, eg])
|
71 |
-
input_text.submit(translate_english, inputs=[input_text], outputs=[pal, sy, lb, eg])
|
72 |
-
pal.change(get_transliteration, inputs=[pal,], outputs=[pal_translit])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
73 |
with gr.Tab("Transliterate"):
|
74 |
with gr.Row():
|
75 |
with gr.Column():
|
@@ -81,10 +110,4 @@ with gr.Blocks(title = "English to Levantine Arabic", css=css, theme="default")
|
|
81 |
translit = gr.Textbox(label="Transliteration", lines=1, elem_id="liter")
|
82 |
btn.click(get_transliteration, inputs=input_text, outputs=[translit])
|
83 |
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
demo.launch()
|
89 |
-
|
90 |
-
|
|
|
6 |
dialects = {"Palestinian/Jordanian": "P", "Syrian": "S", "Lebanese": "L", "Egyptian": "E"}
|
7 |
|
8 |
translator_en2ar = pipeline(task="translation", model="guymorlan/English2Dialect")
|
9 |
+
translator_ar2en = pipeline(task="translation", model="guymorlan/Shami2English")
|
10 |
transliterator = pipeline(task="translation", model="guymorlan/DialectTransliterator")
|
11 |
|
12 |
speech_config = speechsdk.SpeechConfig(subscription=os.environ.get('SPEECH_KEY'), region=os.environ.get('SPEECH_REGION'))
|
13 |
|
14 |
+
def translate_english(input_text, include):
|
15 |
if not input_text:
|
16 |
return "", "", "", ""
|
17 |
|
18 |
inputs = [f"{val} {input_text}" for val in dialects.values()]
|
19 |
+
|
20 |
+
sy, lb, eg = "Syrian" in include, "Lebanese" in include, "Egyptian" in include
|
21 |
+
# remove 2nd element if sy is false
|
22 |
+
if not eg:
|
23 |
+
inputs.pop()
|
24 |
+
if not lb:
|
25 |
+
inputs.pop()
|
26 |
+
if not sy:
|
27 |
+
inputs.pop()
|
28 |
+
|
29 |
result = translator_en2ar(inputs)
|
30 |
|
31 |
+
pal_out = result[0]["translation_text"]
|
32 |
+
sy_out = result[1]["translation_text"] if sy else ""
|
33 |
+
lb_out = result[1 + sy]["translation_text"] if lb else ""
|
34 |
+
eg_out = result[1 + sy + lb]["translation_text"] if eg else ""
|
35 |
+
|
36 |
+
return pal_out, sy_out, lb_out, eg_out
|
37 |
|
38 |
def translate_arabic(input_text):
|
39 |
if not input_text:
|
|
|
50 |
speech_synthesis_result = speech_synthesizer.speak_text_async(input_text).get()
|
51 |
return f"{input_text}.wav"
|
52 |
|
53 |
+
def get_transliteration(input_text, include=["Transliteration"]):
|
54 |
+
if "Transliteration" not in include:
|
55 |
return ""
|
|
|
56 |
result = transliterator([input_text])
|
57 |
return result[0]["translation_text"]
|
58 |
|
|
|
70 |
input_text = gr.Textbox(label="Input", placeholder="Enter English text", lines=1)
|
71 |
gr.Examples(["I wanted to go to the store yesterday, but it rained", "How are you feeling today?", "Let's go to your place"], input_text)
|
72 |
btn = gr.Button("Translate", label="Translate")
|
73 |
+
with gr.Row():
|
74 |
+
include = gr.CheckboxGroup(["Transliteration", "Syrian", "Lebanese", "Egyptian"],
|
75 |
+
label="Disable features to speed up translation",
|
76 |
+
value=["Transliteration", "Syrian", "Lebanese", "Egyptian"])
|
77 |
gr.Markdown("Built by [Guy Mor-Lan](mailto:guy.mor@mail.huji.ac.il). Pronunciation model is specifically tailored to urban Palestinian Arabic. Text-to-speech uses Microsoft Azure's API and may provide different result from the transliterated pronunciation.")
|
78 |
|
79 |
with gr.Column():
|
|
|
86 |
audio = gr.Audio(label="Audio - Palestinian", interactive=False)
|
87 |
audio_button = gr.Button("Get Audio", label="Click Here to Get Audio")
|
88 |
audio_button.click(get_audio, inputs=[pal], outputs=[audio])
|
89 |
+
btn.click(translate_english,inputs=[input_text, include], outputs=[pal, sy, lb, eg])
|
90 |
+
input_text.submit(translate_english, inputs=[input_text, include], outputs=[pal, sy, lb, eg])
|
91 |
+
pal.change(get_transliteration, inputs=[pal, include], outputs=[pal_translit])
|
92 |
+
with gr.Tab('Ar -> En'):
|
93 |
+
with gr.Row():
|
94 |
+
with gr.Column():
|
95 |
+
input_text = gr.Textbox(label="Input", placeholder="Enter Levantine Arabic text", lines=1, elem_id="trans")
|
96 |
+
gr.Examples(["خلينا ندور على مطعم تاني", "قديش حق البندورة؟"], input_text)
|
97 |
+
btn = gr.Button("Translate", label="Translate")
|
98 |
+
gr.Markdown("Built by [Guy Mor-Lan](mailto:guy.mor@mail.huji.ac.il).")
|
99 |
+
with gr.Column():
|
100 |
+
eng = gr.Textbox(label="English", lines=1, elem_id="liter")
|
101 |
+
btn.click(translate_arabic,inputs=input_text, outputs=[eng])
|
102 |
with gr.Tab("Transliterate"):
|
103 |
with gr.Row():
|
104 |
with gr.Column():
|
|
|
110 |
translit = gr.Textbox(label="Transliteration", lines=1, elem_id="liter")
|
111 |
btn.click(get_transliteration, inputs=input_text, outputs=[translit])
|
112 |
|
|
|
|
|
|
|
|
|
113 |
demo.launch()
|
|
|
|