factorize and generalize

app.py

@@ -1,153 +1,154 @@
-import time
-import json
+from huggingface_hub import hf_hub_download
+
 import os
-import pickle
+import time
 import gradio as gr
-import numpy as np
 import yt_dlp
 import sys
-import csv
-from huggingface_hub import hf_hub_download
 import uuid
 import traceback
-
+import tensorflow
 
 embed_html1 = '<iframe width="560" height="315" src="https://www.youtube.com/embed/'
 embed_html2 = '" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>'
 # NO GPU
-#os.environ['CUDA_VISIBLE_DEVICES'] = '-1'
-os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
-import tensorflow 
-#tensorflow.config.threading.set_intra_op_parallelism_threads(8)
-#tensorflow.config.threading.set_inter_op_parallelism_threads(8)
-print(tensorflow.config.list_physical_devices('GPU'))
+# os.environ['CUDA_VISIBLE_DEVICES'] = '-1'
+os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
+
+# tensorflow.config.threading.set_intra_op_parallelism_threads(8)
+# tensorflow.config.threading.set_inter_op_parallelism_threads(8)
+print(tensorflow.config.list_physical_devices("GPU"))
 #
-python_path = hf_hub_download(repo_id=os.environ['REPO_ID'], repo_type="space", filename=os.environ['MODEL_FILE'],
-                                use_auth_token=os.environ['TOKEN'])
-#print("PATH : ", python_path)
-sys.path.append(os.environ['PRIVATE_DIR'])
+python_path = hf_hub_download(
+    repo_id=os.environ["REPO_ID"],
+    repo_type="space",
+    filename=os.environ["MODEL_FILE"],
+    use_auth_token=os.environ["TOKEN"],
+)
+# print("PATH : ", python_path)
+sys.path.append(os.environ["PRIVATE_DIR"])
 from models import *
 
+max_results = 50
+
+
 def download_audio_(link):
-  file_name = f"{uuid.uuid4()}.mp3"
-  with yt_dlp.YoutubeDL({'extract_audio': True, 'format': 'bestaudio', 'outtmpl': file_name}) as video:
-    info_dict = video.extract_info(link, download = True)
-    video_title = info_dict['title']
-    video.download(link)
-    return video_title, file_name
+    file_name = f"{uuid.uuid4()}.mp3"
+    with yt_dlp.YoutubeDL(
+        {"extract_audio": True, "format": "bestaudio", "outtmpl": file_name}
+    ) as video:
+        video.download(link)
+        return file_name
+
 
 def download_audio(id_video):
     id = id_video.split("?v=")[-1][:11]
-    audio_title, file_name = download_audio_(id_video)
-    embed_html_all = embed_html1 + id +embed_html2
-    return audio_title, file_name, embed_html_all
+    file_name = download_audio_(id_video)
+    embed_html_all = embed_html1 + id + embed_html2
+    return file_name, embed_html_all
+
 
 def process_url(input_path):
     # setup the client
-    #try :
-    audio_title, file_name, embed_html_all = download_audio(input_path)
-    return process(file_name, audio_title, embed_html_all)
-    #except:
+    # try :
+    file_name, embed_html_all = download_audio(input_path)
+    return process(file_name, embed_html_all)
+    # except:
     #    return "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", ""
 
+
 def process_file(input_path):
-    return process(input_path, '', '')[1:]
+    return process(input_path, "")[1:]
 
-def process(file_name, audio_title, embed_html_all):
-    '''try : '''
-    emb, ts = get_embed(file_name)
+
+def process(file_name, embed_html_all):
+    """try :"""
+    emb, _ = get_embed(file_name)
     try:
         os.remove(file_name)
     except:
         print(traceback.format_exc())
+
     ind = get_index()
     ind_filenames = get_audio_names()
     catalog = get_catalog()
     url_dict = get_durl_myma()
+    dict_catalog = get_dict_catalog()
 
     timestart = time.time()
     _, I = do_search(emb, ind)
-    print("search time :", time.time()-timestart)
-    top1, top2, top3, top4, top5, top6, top7, top8, top9, top10 = get_top(I, ind_filenames, url_dict, catalog)   
-    return embed_html_all, ind_filenames[I[0][0]].split('.')[0], top1, ind_filenames[I[0][1]].split('.')[0], top2, ind_filenames[I[0][2]].split('.')[0], top3, ind_filenames[I[0][3]].split('.')[0], top4, ind_filenames[I[0][4]].split('.')[0], top5, ind_filenames[I[0][5]].split('.')[0], top6, ind_filenames[I[0][6]].split('.')[0], top7, ind_filenames[I[0][7]].split('.')[0], top8, ind_filenames[I[0][8]].split('.')[0], top9, ind_filenames[I[0][9]].split('.')[0], top10
-    '''except:
+    print("search time :", time.time() - timestart)
+
+    tops = get_topN(I, ind_filenames, url_dict, catalog, max_results)
+    formated = []
+    for i in range(max_results):
+        file = os.path.splitext(os.path.basename(ind_filenames[I[0][0]]))[0]
+        formated.append(file)
+        formated.append(tops[i])
+    return (embed_html_all, *formated)
+    """except:
         return embed_html_all, "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", "", "Erreur Input", ""
-    '''        
+    """
+
+
 with gr.Blocks() as demo:
-    
+    tops = []
     with gr.Row():
-
         with gr.Column():
-
             with gr.Row():
-                #gr.HTML(embed_html)
+                # gr.HTML(embed_html)
                 html = gr.HTML()
-            
+
             with gr.Row():
                 with gr.Column():
-                    audio_url_input = gr.Textbox(placeholder='YouTube video URL', label='YouTube video URL')
-                    analyze_url_btn = gr.Button('Search from URL')
+                    audio_url_input = gr.Textbox(
+                        placeholder="YouTube video URL", label="YouTube video URL"
+                    )
+                    analyze_url_btn = gr.Button("Search from URL")
 
             with gr.Row():
                 with gr.Column():
-                    audio_input_file = gr.Audio(type="filepath", label='Audio Input')
-                    analyze_file_btn = gr.Button('Search from file')
-                
+                    audio_input_file = gr.Audio(type="filepath", label="Audio Input")
+                    analyze_file_btn = gr.Button("Search from file")
 
             with gr.Row():
                 with gr.Column():
-                    '''
-                    gr.HTML("<h3>Top 1</h3>")
-                    top1 = gr.Textbox(label="top1", show_label=False)
-                    gr.HTML("<h3>Top 2</h3>")
-                    top2 = gr.Textbox(label="top2", show_label=False)
-                    gr.HTML("<h3>Top 3</h3>")
-                    top3 = gr.Textbox(label="top3", show_label=False)
-                    gr.HTML("<h3>Top 4</h3>")
-                    top4 = gr.Textbox(label="top4", show_label=False)
-                    gr.HTML("<h3>Top 5</h3>")
-                    top5 = gr.Textbox(label="top5", show_label=False)
-                    '''
-                    top1_name = gr.Textbox(label="top1 track name", show_label=True)
-                    top1 = gr.Audio(label="top1", show_label=False)
-                    top2_name = gr.Textbox(label="top2 track name", show_label=True)
-                    top2 = gr.Audio(label="top2", show_label=False)
-                    top3_name = gr.Textbox(label="top3 track name", show_label=True)
-                    top3 = gr.Audio(label="top3", show_label=False)
-                    top4_name = gr.Textbox(label="top4 track name", show_label=True)
-                    top4 = gr.Audio(label="top4", show_label=False)
-                    top5_name = gr.Textbox(label="top5 track name", show_label=True)
-                    top5 = gr.Audio(label="top5", show_label=False)
-                    top6_name = gr.Textbox(label="top6 track name", show_label=True)
-                    top6 = gr.Audio(label="top6", show_label=False)
-                    top7_name = gr.Textbox(label="top7 track name", show_label=True)
-                    top7 = gr.Audio(label="top7", show_label=False)
-                    top8_name = gr.Textbox(label="top8 track name", show_label=True)
-                    top8 = gr.Audio(label="top8", show_label=False)
-                    top9_name = gr.Textbox(label="top9 track name", show_label=True)
-                    top9 = gr.Audio(label="top9", show_label=False)
-                    top10_name = gr.Textbox(label="top10 track name", show_label=True)
-                    top10 = gr.Audio(label="top10", show_label=False)
-                    
-                
-    
-    analyze_url_btn.click(process_url, inputs=[audio_url_input], 
-                      outputs=[html,top1_name, top1, top2_name, top2, top3_name, top3, top4_name, top4, top5_name, top5, top6_name, top6, top7_name, top7, top8_name, top8, top9_name, top9, top10_name, top10])
+                    for i in range(max_results):
+                        tops.append(
+                            gr.Textbox(label=f"top{i} track name", show_label=True)
+                        )
+                        tops.append(gr.Audio(label=f"top{i}", show_label=False))
+
+    analyze_url_btn.click(
+        process_url,
+        inputs=[audio_url_input],
+        outputs=[
+            html,
+            *tops,
+        ],
+    )
     gr.Examples(
-        examples = [
-         "https://www.youtube.com/watch?v=aNzCDt2eidg", 
-         "https://www.youtube.com/watch?v=NBE-uBgtINg", 
-         "https://www.youtube.com/watch?v=5NV6Rdv1a3I",
-         "https://www.youtube.com/watch?v=OiC1rgCPmUQ",#
-         "https://www.youtube.com/watch?v=dRX0wDNK6S4",#
-         "https://www.youtube.com/watch?v=Guzu9aAeDIc"],
-        inputs = [audio_url_input],
-        outputs = [html,top1_name, top1, top2_name, top2, top3_name, top3, top4_name, top4, top5_name, top5, top6_name, top6, top7_name, top7, top8_name, top8, top9_name, top9, top10_name, top10],
-        fn = process_url,
+        examples=[
+            "https://www.youtube.com/watch?v=aNzCDt2eidg",
+            "https://www.youtube.com/watch?v=NBE-uBgtINg",
+            "https://www.youtube.com/watch?v=5NV6Rdv1a3I",
+            "https://www.youtube.com/watch?v=OiC1rgCPmUQ",  #
+            "https://www.youtube.com/watch?v=dRX0wDNK6S4",  #
+            "https://www.youtube.com/watch?v=Guzu9aAeDIc",
+        ],
+        inputs=[audio_url_input],
+        outputs=[
+            html,
+            *tops,
+        ],
+        fn=process_url,
         cache_examples=False,
     )
-    analyze_file_btn.click(process_file, inputs=[audio_input_file], 
-                      outputs=[top1_name, top1, top2_name, top2, top3_name, top3, top4_name, top4, top5_name, top5, top6_name, top6, top7_name, top7, top8_name, top8, top9_name, top9, top10_name, top10])
+    analyze_file_btn.click(
+        process_file,
+        inputs=[audio_input_file],
+        outputs=tops,
+    )
 
 
-demo.launch(debug=False)
+demo.launch(debug=False)