wagner-austin
commited on
Commit
·
36e257e
1
Parent(s):
4ff8979
Initial Space – Gradio demo for Kazakh/Kyrgyz transliteration
Browse files- app.py +113 -0
- requirements.txt +4 -0
app.py
ADDED
@@ -0,0 +1,113 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
"""
|
2 |
+
A simple web interface to demonstrate the Turkish transliteration.
|
3 |
+
"""
|
4 |
+
import os
|
5 |
+
import gradio as gr
|
6 |
+
from turkic_translit.core import to_latin, to_ipa
|
7 |
+
import unicodedata as ud
|
8 |
+
|
9 |
+
def enable_submit(text):
|
10 |
+
return bool(text)
|
11 |
+
|
12 |
+
def transliterate(text, lang, include_arabic, output_format):
|
13 |
+
if not text:
|
14 |
+
return "", ""
|
15 |
+
try:
|
16 |
+
if output_format == "Latin":
|
17 |
+
result = to_latin(text, lang, include_arabic=include_arabic)
|
18 |
+
format_label = "Latin"
|
19 |
+
else:
|
20 |
+
result = to_ipa(text, lang)
|
21 |
+
format_label = "IPA"
|
22 |
+
result = ud.normalize("NFC", result)
|
23 |
+
stats_md = (f"**Bytes** — Cyrillic : {len(text.encode('utf8'))}, "
|
24 |
+
f"{format_label} : {len(result.encode('utf8'))}")
|
25 |
+
return result, stats_md
|
26 |
+
except Exception as e:
|
27 |
+
raise gr.Error(str(e))
|
28 |
+
|
29 |
+
# Create the Gradio interface
|
30 |
+
with gr.Blocks(title="Turkic Transliteration Demo") as demo:
|
31 |
+
gr.Markdown("# Turkic Transliteration Demo")
|
32 |
+
gr.Markdown("Enter Cyrillic text for Kazakh (kk) or Kyrgyz (ky) and see the Latin transliteration")
|
33 |
+
|
34 |
+
with gr.Row():
|
35 |
+
with gr.Column():
|
36 |
+
input_text = gr.Textbox(
|
37 |
+
label="Input Text (Cyrillic)",
|
38 |
+
placeholder="Enter Kazakh or Kyrgyz text in Cyrillic script...",
|
39 |
+
lines=5
|
40 |
+
)
|
41 |
+
lang = gr.Radio(
|
42 |
+
["kk", "ky"],
|
43 |
+
label="Language",
|
44 |
+
info="kk = Kazakh, ky = Kyrgyz",
|
45 |
+
value="kk"
|
46 |
+
)
|
47 |
+
output_format = gr.Radio(
|
48 |
+
["Latin", "IPA"],
|
49 |
+
label="Output Format",
|
50 |
+
info="Latin = Standard Latin alphabet, IPA = International Phonetic Alphabet",
|
51 |
+
value="Latin"
|
52 |
+
)
|
53 |
+
include_arabic = gr.Checkbox(False, label="Also transliterate Arabic script (Latin mode only)")
|
54 |
+
submit_btn = gr.Button("Transliterate", variant="primary", interactive=False)
|
55 |
+
input_text.change(
|
56 |
+
fn=enable_submit,
|
57 |
+
inputs=input_text,
|
58 |
+
outputs=submit_btn
|
59 |
+
)
|
60 |
+
|
61 |
+
with gr.Column():
|
62 |
+
output_text = gr.Textbox(
|
63 |
+
label="Transliteration Output",
|
64 |
+
lines=5,
|
65 |
+
interactive=False
|
66 |
+
)
|
67 |
+
stats = gr.Markdown(value="")
|
68 |
+
|
69 |
+
# Example inputs
|
70 |
+
examples = [
|
71 |
+
["Қазақ тілі - Түркі тілдерінің бірі.", "kk", "Latin"],
|
72 |
+
["Қазақ тілі - Түркі тілдерінің бірі.", "kk", "IPA"],
|
73 |
+
["Кыргыз тили - Түрк тилдеринин бири.", "ky", "Latin"],
|
74 |
+
["Кыргыз тили - Түрк тилдеринин бири.", "ky", "IPA"]
|
75 |
+
]
|
76 |
+
gr.Examples(examples, [input_text, lang, output_format])
|
77 |
+
|
78 |
+
# Set up the event
|
79 |
+
submit_btn.click(
|
80 |
+
fn=transliterate,
|
81 |
+
inputs=[input_text, lang, include_arabic, output_format],
|
82 |
+
outputs=[output_text, stats]
|
83 |
+
)
|
84 |
+
|
85 |
+
# Also update on input change for real-time feedback
|
86 |
+
input_text.change(
|
87 |
+
fn=transliterate,
|
88 |
+
inputs=[input_text, lang, include_arabic, output_format],
|
89 |
+
outputs=[output_text, stats]
|
90 |
+
)
|
91 |
+
|
92 |
+
lang.change(
|
93 |
+
fn=transliterate,
|
94 |
+
inputs=[input_text, lang, include_arabic, output_format],
|
95 |
+
outputs=[output_text, stats]
|
96 |
+
)
|
97 |
+
|
98 |
+
include_arabic.change(
|
99 |
+
fn=transliterate,
|
100 |
+
inputs=[input_text, lang, include_arabic, output_format],
|
101 |
+
outputs=[output_text, stats]
|
102 |
+
)
|
103 |
+
|
104 |
+
output_format.change(
|
105 |
+
fn=transliterate,
|
106 |
+
inputs=[input_text, lang, include_arabic, output_format],
|
107 |
+
outputs=[output_text, stats]
|
108 |
+
)
|
109 |
+
|
110 |
+
# Launch the app
|
111 |
+
if __name__ == "__main__":
|
112 |
+
demo.queue(concurrency_count=int(os.getenv("CONCURRENCY", "4"))).launch()
|
113 |
+
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
turkic_transliterate==0.1.0 # will pull PyICU, panphon, etc.
|
2 |
+
gradio>=4.30
|
3 |
+
sentencepiece>=0.2
|
4 |
+
rapidfuzz>=3.5
|