Upload 3 files

app.py

@@ -0,0 +1,34 @@
+import gradio as gr
+import requests
+import json
+from utils import get_story, get_voice, get_music
+
+with gr.Blocks() as performance:
+    with gr.Tab("Story Generation"):
+        chatbot = gr.Chatbot(label='GPT4', elem_id="chatbot")
+        input_text = gr.Textbox(lines=2, label="시작 문장")
+        input_number = gr.Number(label='문장수')
+        state = gr.State([])
+        with gr.Row():
+            with gr.Column():
+                b1 = gr.Button().style(full_width=True)
+        b1.click(get_story, [input_text, input_number, chatbot, state], [chatbot, state])
+
+    with gr.Tab("Voice Generation"):
+        input_text = gr.Textbox(lines=10, label="문장")
+        input_gender = gr.Radio(["남성", "여성"], label="성별")
+        input_age = gr.Radio(["어린이", "청소년", "청년", "중년"], label="연령대")
+        with gr.Row():
+            with gr.Column():
+                b2 = gr.Button().style(full_width=True)
+        b2.click(get_voice, [input_text, input_gender, input_age], [gr.Audio(label="결과 음성 파일", type="filepath")])
+
+    with gr.Tab("Music Generation"):
+        input_text = gr.Textbox(lines=10, label="문장")
+        input_duration = gr.Number(label="음악 시간")
+        with gr.Row():
+            with gr.Column():
+                b3 = gr.Button().style(full_width=True)
+        b3.click(get_music, [input_text, input_duration], gr.Audio(label="결과 음악 파일", type='filepath'))
+
+performance.queue(max_size=5).launch()

utils.py

@@ -0,0 +1,256 @@
+# ---
+# jupyter:
+#   jupytext:
+#     formats: ipynb,py:light
+#     text_representation:
+#       extension: .py
+#       format_name: light
+#       format_version: '1.5'
+#       jupytext_version: 1.14.1
+#   kernelspec:
+#     display_name: Python 3 (ipykernel)
+#     language: python
+#     name: python3
+# ---
+
+OPENAPI_KEY = "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"
+CLOVA_VOICE_Client_ID = "yulxvnhzer"
+CLOVA_VOICE_Client_Secret = "EykVrsTYScAkp1dMghZAWp1oL5uB7T6dG01h7Xo7"
+PAPAGO_Translate_Client_ID = "vlujenu5w4"
+PAPAGO_Translate_Client_Secret = "1TvXphvjgjSHY2lk8Wbsk2TwH4PVx1bZmN006NjZ"
+mubert_pat = "c29zczMyNjQuMTk3MzY2MTcuNDQ1ZDQ0NmQ1Y2Y1NTRiNWYwYzUyOTBmZjc2NmEzYTdjMmRhZmVkYi4xLjM.44ed0f5d055012782e08bf7276aefe81fe2e3c7ce233721ce1fb78e93ae5eeae"
+SUMMARY_Client_ID = "2x59wtfeyj"
+SUMMARY_Client_Secret = "in5CTQIsj303cZ3EHgPYjRucMFJvQNUl9YDQTAU5"
+
+# +
+import os
+import sys
+import time
+import urllib.request
+import json
+import random
+import requests
+from voice import voice_dict
+
+OPENAPI_KEY = os.getenv('OPENAPI_KEY')
+CLOVA_VOICE_Client_ID = os.getenv('CLOVA_VOICE_Client_ID')
+CLOVA_VOICE_Client_Secret = os.getenv('CLOVA_VOICE_Client_Secret')
+PAPAGO_Translate_Client_ID = os.getenv('PAPAGO_Translate_Client_ID')
+PAPAGO_Translate_Client_Secret = os.getenv('PAPAGO_Translate_Client_Secret')
+mubert_pat = os.getenv('mubert_pat')
+SUMMARY_Client_ID = os.getenv('SUMMARY_Client_ID')
+SUMMARY_Client_Secret = os.getenv('SUMMARY_Client_Secret')
+
+
+def get_story(first_sentence:str, num_sentences:int):
+    response = requests.post("https://api.openai.com/v1/chat/completions", 
+                            headers={"Content-Type": "application/json", "Authorization": f"Bearer {OPENAPI_KEY}"},
+                            data=json.dumps({
+                                "model": "gpt-3.5-turbo",
+                                "messages": [{"role": "system", "content": "You are a helpful assistant."}, 
+                                            {"role": "user", "content": f"""I will provide the first sentence of the novel, and please write {num_sentences} sentences continuing the story in a first-person protagonist's perspective in Korean. Don't number the sentences.
+                                            \n\nFirst sentence: {first_sentence}"""}]
+                            }))
+    return response.json()['choices'][0]['message']['content']
+
+def get_voice(input_text:str, gender:str="female", age_group:str="youth", filename="voice.mp3"):
+    """
+    gender: female or male
+    age_group: child, teenager, youth, middle_aged
+    """
+    speaker = random.choice(voice_dict[gender][age_group])
+    data = {"speaker":speaker, "text":input_text}
+    url = "https://naveropenapi.apigw.ntruss.com/tts-premium/v1/tts"
+    headers = {
+        "X-NCP-APIGW-API-KEY-ID": CLOVA_VOICE_Client_ID,
+        "X-NCP-APIGW-API-KEY": CLOVA_VOICE_Client_Secret,
+    }
+    response = requests.post(url, headers=headers, data=data)
+    if response.status_code == 200:
+        print("TTS mp3 저장")
+        response_body = response.content
+        with open(filename, 'wb') as f:
+            f.write(response_body)
+    else:
+        print("Error Code: " + str(response.status_code))
+        print("Error Message: " + str(response.json()))
+    return filename
+    
+def translate_text(text:str):
+    
+    encText = urllib.parse.quote(text)
+    data = f"source=ko&target=en&text={encText}"
+    url = "https://naveropenapi.apigw.ntruss.com/nmt/v1/translation"
+    
+    request = urllib.request.Request(url)
+    request.add_header("X-NCP-APIGW-API-KEY-ID", PAPAGO_Translate_Client_ID)
+    request.add_header("X-NCP-APIGW-API-KEY", PAPAGO_Translate_Client_Secret)
+    
+    try:
+        response = urllib.request.urlopen(request, data=data.encode("utf-8"))
+        response_body = response.read()
+        return json.loads(response_body.decode('utf-8'))['message']['result']['translatedText']
+    except urllib.error.HTTPError as e:
+        return f"Error Code: {e.code}"
+
+
+# -
+
+def get_summary(input_text:str, summary_count:int = 5):
+    if len(input_text) > 2000:
+        input_text = input_text[:2000]
+    input_text = input_text.strip()
+        
+    data = {
+          "document": {
+            "content": input_text
+          },
+          "option": {
+            "language": "ko",
+            "model": "general",
+            "tone": "0",
+            "summaryCount": summary_count
+          }
+        }
+    url = "https://naveropenapi.apigw.ntruss.com/text-summary/v1/summarize"
+    headers = {
+        "X-NCP-APIGW-API-KEY-ID": SUMMARY_Client_ID,
+        "X-NCP-APIGW-API-KEY": SUMMARY_Client_Secret,
+        "Content-Type": "application/json"
+    }
+    response = requests.post(url, headers=headers, data=json.dumps(data))
+    if response.status_code == 200:
+        return ' '.join(response.json()['summary'].split('\n'))
+    else:
+        print("Error Code: " + str(response.status_code))
+        print("Error Message: " + str(response.json()))
+
+
+def get_music(text, duration=300):
+    print('original text length: ', len(text))
+    summary = get_summary(text, 3)
+    print('summary text length: ', len(summary))
+    translated_text = translate_text(summary)
+    print('translated_text length: ', len(translated_text))
+    if len(translated_text) > 200:
+        translated_text = translated_text[:200]
+        
+    r = requests.post('https://api-b2b.mubert.com/v2/TTMRecordTrack', 
+        json={
+            "method":"TTMRecordTrack",
+            "params":
+            {
+                "text":translated_text,
+                "pat":mubert_pat,
+                "mode":"track",
+                "duration":duration, 
+                "bitrate":128
+            }
+        })
+
+    rdata = json.loads(r.text)
+    if rdata['status'] == 1:
+        url = rdata['data']['tasks'][0]['download_link']
+        
+        done = False
+        while not done:
+            r = requests.post('https://api-b2b.mubert.com/v2/TrackStatus', 
+            json={
+                "method":"TrackStatus",
+                "params":
+                        {
+                            "pat":mubert_pat
+                        }
+            })
+            
+            if r.json()['data']['tasks'][0]['task_status_text'] == 'Done':
+                done = True
+                time.sleep(2)
+            
+        # return url
+        local_filename = "mubert_music.mp3"
+        headers = {
+            "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.3"
+        }
+        
+        download = False
+        while not download:
+            response = requests.get(url, stream=True, headers=headers)
+            
+            if response.status_code == 200:
+                download=True
+                time.sleep(1)
+
+        if response.status_code == 404:
+            print("파일이 존재하지 않습니다.")
+            return
+        elif response.status_code != 200:
+            print(f"파일 다운로드에 실패하였습니다. 에러 코드: {response.status_code}")
+            return
+
+        with open(local_filename, "wb") as f:
+            for chunk in response.iter_content(chunk_size=8192):
+                if chunk:
+                    f.write(chunk)
+        print(f"{local_filename} 파일이 저장되었습니다.")
+        return local_filename
+
+
+def get_story(first_sentence:str, num_sentences:int, chatbot=[], history=[]):
+    history.append(first_sentence)
+    # make a POST request to the API endpoint using the requests.post method, passing in stream=True
+    response = requests.post("https://api.openai.com/v1/chat/completions", 
+                            headers={"Content-Type": "application/json", "Authorization": f"Bearer {OPENAPI_KEY}"},
+                             stream=True,
+                            data=json.dumps({
+                                "stream": True,
+                                "model": "gpt-3.5-turbo",
+                                "messages": [{"role": "system", "content": "You are a helpful assistant."}, 
+                                            {"role": "user", "content": f"""I will provide the first sentence of the novel, and please write {num_sentences} sentences continuing the story in a first-person protagonist's perspective in Korean. Don't number the sentences.
+                                            \n\nFirst sentence: {first_sentence}"""}]
+                            }))
+
+    token_counter = 0 
+    partial_words = "" 
+    counter=0
+    for chunk in response.iter_lines():
+        #Skipping first chunk
+        if counter == 0:
+            counter+=1
+            continue
+        # check whether each line is non-empty
+        if chunk.decode() :
+            chunk = chunk.decode()
+          # decode each line as response data is in bytes
+            if len(chunk) > 12 and "content" in json.loads(chunk[6:])['choices'][0]['delta']:
+                partial_words = partial_words + json.loads(chunk[6:])['choices'][0]["delta"]["content"]
+                if token_counter == 0:
+                    history.append(" " + partial_words)
+                else:
+                    history[-1] = partial_words
+                chat = [(history[i], history[i + 1]) for i in range(0, len(history) - 1, 2) ]  # convert to tuples of list
+                token_counter+=1
+                yield chat, history, response
+
+
+def get_voice_filename(text, gender, age):
+    filename = None
+    if gender == '남성':
+        if age == "어린이":
+            filename = get_voice(text, gender="male", age_group="child", filename="voice.mp3")
+        elif age == "청소년":
+            filename = get_voice(text, gender="male", age_group="teenager", filename="voice.mp3")
+        elif age == "청년":
+            filename = get_voice(text, gender="male", age_group="youth", filename="voice.mp3")
+        elif age == "중년":
+            filename = get_voice(text, gender="male", age_group="middle_aged", filename="voice.mp3")
+    else:
+        if age == "어린이":
+            filename = get_voice(text, gender="female", age_group="child", filename="voice.mp3")
+        elif age == "청소년":
+            filename = get_voice(text, gender="female", age_group="teenager", filename="voice.mp3")
+        elif age == "청년":
+            filename = get_voice(text, gender="female", age_group="youth", filename="voice.mp3")
+        elif age == "중년":
+            filename = get_voice(text, gender="female", age_group="middle_aged", filename="voice.mp3")
+    return filename

voice.py

@@ -0,0 +1,16 @@
+voice_dict = {
+    "female": {
+        "child": ["ndain","ngaram", "nmeow", "vdain"],
+        "teenager": ["nminseo", "nbora", "nihyun", "njiwon"],
+        "youth": ["nara","vara", "nminyoung", "nyuna", "vyuna", "vhyeri", "nes_c_hyeri", "ngoeun", "ntiffany", "nnarae", "njangj", "nyejin",
+                  "njiyun", "nsujin", "nes_c_sohyun", "noyj", "neunseo", "nheera", "nyoungmi", "nyeji", "nsabina", "nyounghwa", "nshasha"],
+        "middle_aged": ["mijin", "neunyoung", "vmikyung", "nsunkyung", "nyujin", "nsunhee", "nes_c_mikyung", "nminjeong", "nkyunglee", "napple", ""]
+    },
+    "male": {
+        "child": ["nwoof", "nhajun"],
+        "teenager": ["njonghyun", "njoonyoung", "njaewook"],
+        "youth": ["jinho", "nminsang", "njinho", "njihun", "njihwan", "nseonghoon", "nsiyoon", "ntaejin", "njooahn", "nian", "vian", "vdonghyun",
+                  "ndonghyun", "nsangdo", "neunwoo", "nraewon", "nreview", "nmovie"],
+        "middle_aged": ["nseungpyo", "nkyungtae", "nwontak", "nwoosik", "nyoungil", "nes_c_kihyo", "nkitae", "nkyuwon"]
+    }
+}