init asr project

Dockerfile

@@ -0,0 +1,10 @@
+FROM  python:3.8
+RUN mkdir /app
+WORKDIR /app
+COPY requirements.txt .
+
+RUN pip install -r requirements.txt
+
+COPY . .
+
+ENTRYPOINT ["streamlit", "run", "app.py", "--server.port=4000", "--server.address=0.0.0.0", "--client.showErrorDetails=false"]

app.py

@@ -0,0 +1,75 @@
+import streamlit as st
+from PIL import Image
+import os
+import requests
+
+
+ASR_API = os.environ['ASR_API']
+
+def request_to_asr_service(audiofile):
+
+    # file_path = "/media/mohammadkrb/hddExt/personal_projects/vidabia/audio_tests/epit_sample.mp3"
+    # file_data = open(file_path, 'rb')
+
+    files = {'file': (audiofile)}
+
+    response = requests.post(ASR_API, files=files)
+    return response.json()
+
+st.set_page_config(
+    page_title="Automatic Speech Recognition",
+    page_icon="🗣",
+    layout="centered",
+    initial_sidebar_state="auto",
+)
+
+upload_path = "uploads/"
+download_path = "downloads/"
+os.makedirs(upload_path, exist_ok=True)
+os.makedirs(download_path, exist_ok=True)
+# @st.cache(persist=True,allow_output_mutation=True,show_spinner=False,suppress_st_warning=True)
+# def asr_inference_wav2vec2(uploaded_file):
+#     asr = Wave2Vec2Inference("facebook/wav2vec2-base-960h")
+#     text = asr.file_to_text(uploaded_file)
+#     return text
+
+@st.cache(persist=True,allow_output_mutation=True,show_spinner=False,suppress_st_warning=True)
+def save_text(text, downloaded_txt_file):
+    with open(downloaded_txt_file, 'w') as outtxt:
+        outtxt.write(text)
+    print(downloaded_txt_file)
+
+@st.cache(persist=True,allow_output_mutation=True,show_spinner=False,suppress_st_warning=True)
+def download_success():
+    st.balloons()
+    st.success('✅ Download Successful !!')
+
+main_image = Image.open('static/main_banner.png')
+
+st.image(main_image,use_column_width='auto')
+st.title("🗣 Automatic Speech Recognition")
+st.info('✨ Supports ALL Audio Formats (mp3, wav, ogg, ...).')
+
+uploaded_file = st.file_uploader("Upload audio file", type=["wav"])
+if uploaded_file is not None:
+    with open(os.path.join(upload_path,uploaded_file.name),"wb") as f:
+        f.write((uploaded_file).getbuffer())
+    with st.spinner(f"Converting speech to text... 💫"):
+        resp = request_to_asr_service(uploaded_file)
+        text = resp['transcript']
+        # text = asr_inference_wav2vec2(upload_path + uploaded_file.name)
+        st.info(text)
+        downloaded_txt_file = os.path.abspath(os.path.join(download_path,str("processed_"+uploaded_file.name.split(".")[0] + ".txt")))
+        save_text(text, downloaded_txt_file)
+        with open(downloaded_txt_file, "rb") as file:
+            if st.download_button(
+                                    label="Download ASR Output 🗣",
+                                    data=file,
+                                    file_name=str("ASR_output_"+uploaded_file.name.split(".")[0]+ ".txt"),
+                                    mime='text/plain'
+                                 ):
+                download_success()
+# else:
+    # st.warning("Please upload your  file. Any other audio format is currently not supported")
+
+st.markdown("<br><hr><center>Made with ❤️ by <a href='https://ahdsoft.ir'><strong>AHD Co</strong></a></center><hr>", unsafe_allow_html=True)

requiremets.txt

@@ -0,0 +1,2 @@
+streamlit
+PIL