Spaces:
Runtime error
Runtime error
Commit
·
f89e973
1
Parent(s):
216427b
latest ver
Browse files
app.py
CHANGED
|
@@ -2,13 +2,20 @@ import gradio as gr
|
|
| 2 |
import torch
|
| 3 |
from transformers import pipeline
|
| 4 |
import os
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
|
| 6 |
-
# ✅
|
| 7 |
-
|
|
|
|
| 8 |
|
| 9 |
-
# ✅
|
|
|
|
|
|
|
|
|
|
| 10 |
MODEL_NAME = "biodatlab/whisper-th-small-combined"
|
| 11 |
-
lang = "th"
|
| 12 |
device = 0 if torch.cuda.is_available() else "cpu"
|
| 13 |
|
| 14 |
pipe = pipeline(
|
|
@@ -18,31 +25,95 @@ pipe = pipeline(
|
|
| 18 |
device=device,
|
| 19 |
)
|
| 20 |
|
| 21 |
-
# ✅ ฟังก์ชันแปลงเสียงเป็นข้อความ
|
| 22 |
def transcribe_audio(audio):
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
-
# ✅
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
|
| 38 |
-
|
| 39 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
|
| 41 |
-
#
|
| 42 |
transcribe_btn.click(
|
| 43 |
fn=transcribe_audio,
|
| 44 |
inputs=audio_input,
|
| 45 |
-
outputs=[transcribed_text,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
)
|
| 47 |
|
|
|
|
| 48 |
demo.launch()
|
|
|
|
| 2 |
import torch
|
| 3 |
from transformers import pipeline
|
| 4 |
import os
|
| 5 |
+
import tempfile
|
| 6 |
+
import shutil
|
| 7 |
+
from docx import Document
|
| 8 |
+
import time
|
| 9 |
|
| 10 |
+
# ✅ ตรวจสอบ ffmpeg
|
| 11 |
+
if not shutil.which("ffmpeg"):
|
| 12 |
+
raise EnvironmentError("ffmpeg not found. Please install ffmpeg and ensure it's in PATH.")
|
| 13 |
|
| 14 |
+
# ✅ ลบ path ffmpeg เฉพาะ local เพราะ Spaces มี ffmpeg ติดตั้งแล้ว
|
| 15 |
+
# os.environ["PATH"] += os.pathsep + r"C:\ffmpeg\ffmpeg-master-latest-win64-gpl\ffmpeg-master-latest-win64-gpl\bin"
|
| 16 |
+
|
| 17 |
+
# ✅ โหลดโมเดล small
|
| 18 |
MODEL_NAME = "biodatlab/whisper-th-small-combined"
|
|
|
|
| 19 |
device = 0 if torch.cuda.is_available() else "cpu"
|
| 20 |
|
| 21 |
pipe = pipeline(
|
|
|
|
| 25 |
device=device,
|
| 26 |
)
|
| 27 |
|
| 28 |
+
# ✅ ฟังก์ชันแปลงเสียงเป็นข้อความ (return text และ processing time)
|
| 29 |
def transcribe_audio(audio):
|
| 30 |
+
start_time = time.time() # บันทึกเวลาเริ่มต้น
|
| 31 |
+
if not audio:
|
| 32 |
+
return "กรุณาอัปโหลดไฟล์เสียงก่อน", "ไม่ได้ประมวลผล"
|
| 33 |
+
try:
|
| 34 |
+
result = pipe(audio, generate_kwargs={"language": "<|th|>", "task": "transcribe"}, batch_size=14)
|
| 35 |
+
text = result["text"]
|
| 36 |
+
end_time = time.time() # บันทึกเวลาสิ้นสุด
|
| 37 |
+
processing_time = end_time - start_time
|
| 38 |
+
return text, f"ใช้เวลา: {processing_time:.2f} วินาที"
|
| 39 |
+
except Exception as e:
|
| 40 |
+
return f"เกิดข้อผิดพลาด: {str(e)}", "เกิดข้อผิดพลาด"
|
| 41 |
|
| 42 |
+
# ✅ ฟังก์ชันสร้างไฟล์สำหรับดาวน์โหลด (.txt หรือ .docx)
|
| 43 |
+
def create_download_file(text, file_format):
|
| 44 |
+
if not text or text.startswith("กรุณา") or text.startswith("เกิดข้อผิดพลาด"):
|
| 45 |
+
return None
|
| 46 |
+
try:
|
| 47 |
+
if file_format == "Text (.txt)":
|
| 48 |
+
with tempfile.NamedTemporaryFile(suffix=".txt", delete=False, mode="w", encoding="utf-8") as f:
|
| 49 |
+
f.write(text)
|
| 50 |
+
return f.name
|
| 51 |
+
else: # Word (.docx)
|
| 52 |
+
doc = Document()
|
| 53 |
+
doc.add_paragraph(text)
|
| 54 |
+
with tempfile.NamedTemporaryFile(suffix=".docx", delete=False) as f:
|
| 55 |
+
doc.save(f.name)
|
| 56 |
+
return f.name
|
| 57 |
+
except Exception as e:
|
| 58 |
+
return None
|
| 59 |
+
|
| 60 |
+
# ✅ CSS สำหรับจัด Markdown ตรงกลางและทำให้ Textbox มีแถบเลื่อน
|
| 61 |
+
custom_css = """
|
| 62 |
+
.markdown {
|
| 63 |
+
text-align: center !important;
|
| 64 |
+
}
|
| 65 |
+
#transcribed-text textarea {
|
| 66 |
+
height: 250px !important;
|
| 67 |
+
overflow-y: auto !important;
|
| 68 |
+
resize: vertical !important;
|
| 69 |
+
}
|
| 70 |
+
"""
|
| 71 |
|
| 72 |
+
# ✅ UI Layout
|
| 73 |
+
with gr.Blocks(css=custom_css) as demo:
|
| 74 |
+
gr.Markdown("""
|
| 75 |
+
<div style="text-align: center;">
|
| 76 |
+
<h2> แปลงเสียงพูดภาษาไทยเป็นข้อความ </h2>
|
| 77 |
+
</div>
|
| 78 |
+
""")
|
| 79 |
+
with gr.Row():
|
| 80 |
+
with gr.Column(scale=1):
|
| 81 |
+
audio_input = gr.Audio(label="🎵 อัปโหลดไฟล์เสียง (MP3, WAV, M4A)", type="filepath")
|
| 82 |
+
download_format = gr.Dropdown(
|
| 83 |
+
choices=["Text (.txt)", "Word (.docx)"],
|
| 84 |
+
label="📄 เลือกฟอร์แมตไฟล์",
|
| 85 |
+
value="Text (.txt)"
|
| 86 |
+
)
|
| 87 |
+
transcribe_btn = gr.Button("🔄 แปลงเสียงเป็นข้อความ")
|
| 88 |
+
with gr.Column(scale=2):
|
| 89 |
+
transcribed_text = gr.Textbox(label="📜 ข้อความที่แปลงแล้ว", elem_id="transcribed-text")
|
| 90 |
+
processing_time_display = gr.Textbox(label="⏱️ เวลาที่ใช้", interactive=False)
|
| 91 |
+
with gr.Row():
|
| 92 |
+
copy_button = gr.Button("📋 คัดลอกข้อความ")
|
| 93 |
+
download_button = gr.DownloadButton(label="⬇️ ดาวน์โหลดไฟล์")
|
| 94 |
|
| 95 |
+
# Action
|
| 96 |
transcribe_btn.click(
|
| 97 |
fn=transcribe_audio,
|
| 98 |
inputs=audio_input,
|
| 99 |
+
outputs=[transcribed_text, processing_time_display],
|
| 100 |
+
show_progress=True
|
| 101 |
+
)
|
| 102 |
+
|
| 103 |
+
# Action คัดลอก (ใช้ JavaScript)
|
| 104 |
+
copy_button.click(
|
| 105 |
+
fn=None,
|
| 106 |
+
inputs=transcribed_text,
|
| 107 |
+
outputs=None,
|
| 108 |
+
js="function(text) {navigator.clipboard.writeText(text); gr.Info('คัดลอกข้อความแล้ว!'); return []}"
|
| 109 |
+
)
|
| 110 |
+
|
| 111 |
+
# Action ดาวน์โหลด
|
| 112 |
+
download_button.click(
|
| 113 |
+
fn=create_download_file,
|
| 114 |
+
inputs=[transcribed_text, download_format],
|
| 115 |
+
outputs=download_button
|
| 116 |
)
|
| 117 |
|
| 118 |
+
# รันใน Hugging Face Spaces
|
| 119 |
demo.launch()
|