fast api

Dockerfile

@@ -0,0 +1,12 @@
+FROM python:3.10-slim
+
+WORKDIR /code
+
+COPY app /code/app
+
+RUN pip install --upgrade pip
+RUN pip install -r /code/app/requirements.txt
+
+EXPOSE 7860
+
+CMD python app/prepare.py && uvicorn app.main:app --host 0.0.0.0 --port 7860

app/main.py

@@ -0,0 +1,137 @@
+import os
+import nltk
+import uvicorn
+from fastapi import FastAPI, File, UploadFile
+from pydantic import BaseModel
+from sumy.parsers.plaintext import PlaintextParser
+from sumy.nlp.tokenizers import Tokenizer
+from sumy.summarizers.lsa import LsaSummarizer
+from groq import Groq
+
+GROQ_API_KEY = "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"
+groq_client = Groq(api_key=GROQ_API_KEY)
+
+def save_to_file(content: str, filename: str) -> str:
+    with open(filename, 'w', encoding='utf-8') as file:
+        file.write(content)
+    return filename
+
+def transcribe_and_summarize(audio_path: str):
+    with open(audio_path, "rb") as audio_file:
+        response = groq_client.audio.transcriptions.create(
+            model="whisper-large-v3",
+            file=audio_file,
+            response_format="text"
+        )
+    transcription = response
+
+    parser = PlaintextParser.from_string(transcription, Tokenizer("english"))
+    summarizer = LsaSummarizer()
+    summary_sentences = summarizer(parser.document, 5)
+    summarized_text = " ".join(str(s) for s in summary_sentences)
+
+    original_tokens = len(nltk.word_tokenize(transcription))
+    summarized_tokens = len(nltk.word_tokenize(summarized_text))
+    token_info = f"Asli: {original_tokens} token | Ringkasan: {summarized_tokens} token"
+
+    summarized_file = save_to_file(summarized_text, "summarized_transcription.txt")
+
+    return transcription, summarized_text, summarized_file, audio_path, token_info
+
+def summarize_soap(dialogue: str):
+    prompt_soap = f"""
+    Anda adalah asisten medis yang membantu dokter dalam menyusun catatan SOAP berdasarkan percakapan dokter dan pasien.
+    Ringkaskan dalam bentuk paragraf tanpa adanya bullet point dan gunakan bahasa Indonesia.
+    Harap buat ringkasan dalam format berikut:
+    Subjective:
+    Objective:
+    Assessment:
+    Plan:
+
+    ### Percakapan:
+    {dialogue}
+
+    Tolong jangan tambahkan informasi tambahan selain yang berkaitan dengan diagnosis, obat, hasil lab, dan radiologi.
+    """
+    response_soap = groq_client.chat.completions.create(
+        model="llama3-8b-8192",
+        messages=[{"role": "user", "content": prompt_soap}]
+    )
+    return response_soap.choices[0].message.content
+
+def generate_soap(transcription: str):
+    soap_content = summarize_soap(transcription)
+    soap_file = save_to_file(soap_content, "soap_summary.txt")
+    return soap_content, soap_file
+
+def detect_medical_tags(dialogue: str):
+    prompt_tags = f"""
+    Identifikasi dan berikan luaran dalam bahasa Indonesia tags berikut dari percakapan dengan format:
+    Diagnosis:
+    Obat:
+    Hasil Lab:
+    Radiologi:
+
+    ### Percakapan:
+    {dialogue}
+
+    Tolong jangan tambahkan informasi tambahan selain yang berkaitan dengan diagnosis, obat, hasil lab, dan radiologi.
+    """
+    response_tags = groq_client.chat.completions.create(
+        model="llama3-8b-8192",
+        messages=[{"role": "user", "content": prompt_tags}]
+    )
+    return response_tags.choices[0].message.content
+
+def generate_tags(transcription: str):
+    tags_content = detect_medical_tags(transcription)
+    tags_file = save_to_file(tags_content, "medical_tags.txt")
+    return tags_content, tags_file
+
+app = FastAPI(title="Medical Transcription Pipeline (Groq API)")
+
+@app.post("/full_process")
+async def full_process(audio: UploadFile = File(...)):
+    filename = audio.filename
+    temp_audio_path = f"temp_{filename}"
+    with open(temp_audio_path, "wb") as f:
+        f.write(await audio.read())
+
+    transcription, summarized_text, summarized_file, audio_path, token_info = transcribe_and_summarize(temp_audio_path)
+
+    soap_content, soap_file = generate_soap(transcription)
+    tags_content, tags_file = generate_tags(transcription)
+
+    os.remove(temp_audio_path)  # bersihkan file temporer setelah selesai digunakan
+
+    return {
+        "transcription": transcription,
+        "summarized_text": summarized_text,
+        "summarized_file": summarized_file,
+        "audio_path": audio_path,
+        "token_info": token_info,
+        "soap_content": soap_content,
+        "soap_file": soap_file,
+        "tags_content": tags_content,
+        "tags_file": tags_file
+    }
+
+class TranscriptionInput(BaseModel):
+    dialogue: str
+
+@app.post("/soap_tags")
+async def soap_tags(data: TranscriptionInput):
+    transcript_text = data.dialogue
+
+    soap_content, soap_file = generate_soap(transcript_text)
+    tags_content, tags_file = generate_tags(transcript_text)
+
+    return {
+        "soap_content": soap_content,
+        "soap_file": soap_file,
+        "tags_content": tags_content,
+        "tags_file": tags_file
+    }
+
+if __name__ == "__main__":
+    uvicorn.run("main:app", host="0.0.0.0", port=8000, reload=True)

app/nltk.py

@@ -0,0 +1,4 @@
+import nltk
+
+nltk.download("punkt")
+nltk.download("punkt_tab")

app/requirements.txt

@@ -0,0 +1,8 @@
+fastapi==0.115.8
+faster_whisper==1.1.1
+huggingface_hub==0.28.1
+soundfile==0.13.1
+sumy==0.11.0
+nltk==3.9.1
+uvicorn==0.34.0
+groq==0.18.0